nohup: ignoring input
Loading Qwen2.5-Coder-1.5B...

Loading weights:   0%|          | 0/338 [00:00<?, ?it/s]/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/bitsandbytes/backends/cuda/ops.py:213: FutureWarning: _check_is_size will be removed in a future PyTorch release along with guard_size_oblivious.     Use _check(i >= 0) instead.
  torch._check_is_size(blocksize)

Loading weights:   1%|          | 3/338 [00:00<00:13, 25.64it/s]
Loading weights:  15%|█▌        | 52/338 [00:00<00:01, 276.29it/s]
Loading weights:  30%|██▉       | 101/338 [00:00<00:00, 368.26it/s]
Loading weights:  47%|████▋     | 159/338 [00:00<00:00, 443.26it/s]
Loading weights:  62%|██████▏   | 209/338 [00:00<00:00, 459.99it/s]
Loading weights:  79%|███████▉  | 267/338 [00:00<00:00, 499.35it/s]
Loading weights:  94%|█████████▍| 318/338 [00:00<00:00, 499.07it/s]
Loading weights: 100%|██████████| 338/338 [00:00<00:00, 445.27it/s]
Model loaded.
Training with schema-aware prompts...
[transformers] The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'bos_token_id': None, 'pad_token_id': 151645}.

  0%|          | 0/500 [00:00<?, ?it/s]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  0%|          | 1/500 [00:28<3:55:16, 28.29s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  0%|          | 2/500 [00:56<3:54:44, 28.28s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  1%|          | 3/500 [01:24<3:54:02, 28.25s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  1%|          | 4/500 [01:52<3:53:05, 28.20s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.981 s=0.600 a=1.000 total=0.514
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  1%|          | 5/500 [02:21<3:53:01, 28.25s/it]
                                                 
{'loss': '-2.235e-09', 'grad_norm': '0.04688', 'learning_rate': '4.96e-05', 'num_tokens': '3.48e+04', 'completions/mean_length': '254', 'completions/min_length': '224.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '19.6', 'completions/min_terminated_length': '19.6', 'completions/max_terminated_length': '19.6', 'rewards/dbre_reward/mean': '0.025', 'rewards/dbre_reward/std': '0.1', 'reward': '0.025', 'reward_std': '0.1', 'frac_reward_zero_std': '0.8', 'entropy': '0.9038', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.4', 'epoch': '0.1'}

  1%|          | 5/500 [02:21<3:53:01, 28.25s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.559
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  1%|          | 6/500 [02:49<3:52:40, 28.26s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  1%|▏         | 7/500 [03:17<3:52:35, 28.31s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  2%|▏         | 8/500 [03:46<3:51:59, 28.29s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  2%|▏         | 9/500 [04:15<3:53:03, 28.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.993 s=0.600 a=1.000 total=0.918
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  2%|▏         | 10/500 [04:43<3:53:30, 28.59s/it]
                                                  
{'loss': '-5.96e-09', 'grad_norm': '0.06738', 'learning_rate': '4.91e-05', 'num_tokens': '6.949e+04', 'completions/mean_length': '252.6', 'completions/min_length': '202', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '48.4', 'completions/min_terminated_length': '48.4', 'completions/max_terminated_length': '48.4', 'rewards/dbre_reward/mean': '0.09869', 'rewards/dbre_reward/std': '0.2585', 'reward': '0.09869', 'reward_std': '0.2585', 'frac_reward_zero_std': '0.5', 'entropy': '0.9423', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.68', 'epoch': '0.2'}

  2%|▏         | 10/500 [04:43<3:53:30, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.800 a=1.000 total=0.556

  2%|▏         | 11/500 [05:12<3:53:34, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.998 s=0.500 a=1.000 total=0.899

  2%|▏         | 12/500 [05:41<3:53:06, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.800 a=1.000 total=0.558
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  3%|▎         | 13/500 [06:09<3:52:18, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  3%|▎         | 14/500 [06:38<3:51:02, 28.52s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.400 a=1.000 total=0.480
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  3%|▎         | 15/500 [07:06<3:50:17, 28.49s/it]
                                                  
{'loss': '-0.002071', 'grad_norm': '0.05933', 'learning_rate': '4.86e-05', 'num_tokens': '1.043e+05', 'completions/mean_length': '254.5', 'completions/min_length': '232.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '79', 'completions/min_terminated_length': '79', 'completions/max_terminated_length': '79', 'rewards/dbre_reward/mean': '0.1104', 'rewards/dbre_reward/std': '0.2751', 'reward': '0.1104', 'reward_std': '0.2751', 'frac_reward_zero_std': '0.5', 'entropy': '0.9289', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.68', 'epoch': '0.3'}

  3%|▎         | 15/500 [07:06<3:50:17, 28.49s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  3%|▎         | 16/500 [07:35<3:50:12, 28.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  3%|▎         | 17/500 [08:03<3:49:24, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.600 a=1.000 total=0.516
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  4%|▎         | 18/500 [08:32<3:49:58, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  4%|▍         | 19/500 [09:01<3:49:14, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  4%|▍         | 20/500 [09:29<3:48:32, 28.57s/it]
                                                  
{'loss': '-0.006692', 'grad_norm': '0.052', 'learning_rate': '4.81e-05', 'num_tokens': '1.388e+05', 'completions/mean_length': '250.2', 'completions/min_length': '163.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '61.4', 'completions/min_terminated_length': '61.4', 'completions/max_terminated_length': '61.4', 'rewards/dbre_reward/mean': '0.1119', 'rewards/dbre_reward/std': '0.3216', 'reward': '0.1119', 'reward_std': '0.3216', 'frac_reward_zero_std': '0.3', 'entropy': '0.8891', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.73', 'epoch': '0.4'}

  4%|▍         | 20/500 [09:29<3:48:32, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  4%|▍         | 21/500 [09:58<3:48:28, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.996 s=0.800 a=1.000 total=0.959

  4%|▍         | 22/500 [10:26<3:47:46, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  5%|▍         | 23/500 [10:55<3:46:54, 28.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  5%|▍         | 24/500 [11:24<3:47:14, 28.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  5%|▌         | 25/500 [11:53<3:47:28, 28.73s/it]
                                                  
{'loss': '0.007356', 'grad_norm': '0', 'learning_rate': '4.76e-05', 'num_tokens': '1.737e+05', 'completions/mean_length': '254.5', 'completions/min_length': '231.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '78.2', 'completions/min_terminated_length': '78.2', 'completions/max_terminated_length': '78.2', 'rewards/dbre_reward/mean': '0.0375', 'rewards/dbre_reward/std': '0.15', 'reward': '0.0375', 'reward_std': '0.15', 'frac_reward_zero_std': '0.7', 'entropy': '0.9065', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.84', 'epoch': '0.5'}

  5%|▌         | 25/500 [11:53<3:47:28, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.973 s=0.800 a=1.000 total=0.552
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.982 s=0.800 a=1.000 total=0.554
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  5%|▌         | 26/500 [12:30<4:08:10, 31.41s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.800 a=1.000 total=0.557
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.972 s=0.600 a=1.000 total=0.511
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  5%|▌         | 27/500 [13:09<4:24:25, 33.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  6%|▌         | 28/500 [13:50<4:40:59, 35.72s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  6%|▌         | 29/500 [14:29<4:49:50, 36.92s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.994 s=0.600 a=1.000 total=0.918
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  6%|▌         | 30/500 [15:05<4:46:33, 36.58s/it]
                                                  
{'loss': '-0.005647', 'grad_norm': '0.05981', 'learning_rate': '4.71e-05', 'num_tokens': '2.082e+05', 'completions/mean_length': '251.2', 'completions/min_length': '196.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '56.1', 'completions/min_terminated_length': '43.2', 'completions/max_terminated_length': '69', 'rewards/dbre_reward/mean': '0.0625', 'rewards/dbre_reward/std': '0.1866', 'reward': '0.0625', 'reward_std': '0.1866', 'frac_reward_zero_std': '0.6', 'entropy': '0.8587', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '37.55', 'epoch': '0.6'}

  6%|▌         | 30/500 [15:05<4:46:33, 36.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  6%|▌         | 31/500 [15:41<4:44:33, 36.40s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.600 a=1.000 total=0.516
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.981 s=0.800 a=1.000 total=0.554
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  6%|▋         | 32/500 [16:20<4:48:43, 37.02s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.964 s=0.800 a=1.000 total=0.549
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  7%|▋         | 33/500 [16:55<4:44:14, 36.52s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517

  7%|▋         | 34/500 [17:35<4:51:01, 37.47s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.991 s=0.600 a=1.000 total=0.917
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  7%|▋         | 35/500 [18:09<4:44:00, 36.65s/it]
                                                  
{'loss': '0.001173', 'grad_norm': '0.06787', 'learning_rate': '4.66e-05', 'num_tokens': '2.43e+05', 'completions/mean_length': '253.6', 'completions/min_length': '218', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '115.6', 'completions/min_terminated_length': '115.6', 'completions/max_terminated_length': '115.6', 'rewards/dbre_reward/mean': '0.1103', 'rewards/dbre_reward/std': '0.311', 'reward': '0.1103', 'reward_std': '0.311', 'frac_reward_zero_std': '0.3', 'entropy': '0.8726', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '35.88', 'epoch': '0.7'}

  7%|▋         | 35/500 [18:09<4:44:00, 36.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.983 s=0.800 a=1.000 total=0.555
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  7%|▋         | 36/500 [18:44<4:38:59, 36.08s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  7%|▋         | 37/500 [19:19<4:36:14, 35.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.967 s=1.000 a=1.000 total=0.590
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  8%|▊         | 38/500 [19:54<4:33:27, 35.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.500 a=1.000 total=0.899
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  8%|▊         | 39/500 [20:29<4:30:38, 35.22s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.969 s=0.600 a=1.000 total=0.511
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  8%|▊         | 40/500 [21:04<4:31:15, 35.38s/it]
                                                  
{'loss': '-8.941e-09', 'grad_norm': '0.04688', 'learning_rate': '4.61e-05', 'num_tokens': '2.78e+05', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.09995', 'rewards/dbre_reward/std': '0.3048', 'reward': '0.09995', 'reward_std': '0.3048', 'frac_reward_zero_std': '0.4', 'entropy': '0.909', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '34.06', 'epoch': '0.8'}

  8%|▊         | 40/500 [21:04<4:31:15, 35.38s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  8%|▊         | 41/500 [21:41<4:32:32, 35.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  8%|▊         | 42/500 [22:17<4:33:18, 35.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.981 s=0.600 a=1.000 total=0.514
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.558
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474

  9%|▊         | 43/500 [22:52<4:31:21, 35.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.992 s=0.800 a=1.000 total=0.958
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.500 a=1.000 total=0.498
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  9%|▉         | 44/500 [23:27<4:28:30, 35.33s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  9%|▉         | 45/500 [24:03<4:30:40, 35.69s/it]
                                                  
{'loss': '1.49e-09', 'grad_norm': '0', 'learning_rate': '4.56e-05', 'num_tokens': '3.128e+05', 'completions/mean_length': '254.4', 'completions/min_length': '230.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '26', 'completions/min_terminated_length': '26', 'completions/max_terminated_length': '26', 'rewards/dbre_reward/mean': '0.07415', 'rewards/dbre_reward/std': '0.1971', 'reward': '0.07415', 'reward_std': '0.1971', 'frac_reward_zero_std': '0.6', 'entropy': '0.8978', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '34.82', 'epoch': '0.9'}

  9%|▉         | 45/500 [24:03<4:30:40, 35.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  9%|▉         | 46/500 [24:39<4:29:42, 35.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.988 s=0.600 a=1.000 total=0.916
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.100 a=1.000 total=0.418
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.974 s=0.400 a=1.000 total=0.472
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

  9%|▉         | 47/500 [25:15<4:30:18, 35.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 10%|▉         | 48/500 [25:50<4:29:07, 35.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 10%|▉         | 49/500 [26:27<4:31:23, 36.11s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 10%|█         | 50/500 [27:09<4:42:04, 37.61s/it]
                                                  
{'loss': '-2.235e-09', 'grad_norm': '0', 'learning_rate': '4.51e-05', 'num_tokens': '3.474e+05', 'completions/mean_length': '250.9', 'completions/min_length': '174', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '20.4', 'completions/min_terminated_length': '20.4', 'completions/max_terminated_length': '20.4', 'rewards/dbre_reward/mean': '0.04708', 'rewards/dbre_reward/std': '0.1246', 'reward': '0.04708', 'reward_std': '0.1246', 'frac_reward_zero_std': '0.8', 'entropy': '1.01', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '36.14', 'epoch': '1'}

 10%|█         | 50/500 [27:09<4:42:04, 37.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 10%|█         | 51/500 [27:50<4:50:42, 38.85s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 10%|█         | 52/500 [28:18<4:25:54, 35.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 11%|█         | 53/500 [28:47<4:09:00, 33.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 11%|█         | 54/500 [29:16<3:58:28, 32.08s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 11%|█         | 55/500 [29:44<3:50:26, 31.07s/it]
                                                  
{'loss': '-0.01794', 'grad_norm': '0.06396', 'learning_rate': '4.46e-05', 'num_tokens': '3.814e+05', 'completions/mean_length': '244.6', 'completions/min_length': '161.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9375', 'completions/mean_terminated_length': '18.55', 'completions/min_terminated_length': '7.8', 'completions/max_terminated_length': '41.8', 'rewards/dbre_reward/mean': '0.06066', 'rewards/dbre_reward/std': '0.2125', 'reward': '0.06066', 'reward_std': '0.2125', 'frac_reward_zero_std': '0.6', 'entropy': '0.993', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '28.95', 'epoch': '1.1'}

 11%|█         | 55/500 [29:44<3:50:26, 31.07s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 11%|█         | 56/500 [30:13<3:44:25, 30.33s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 11%|█▏        | 57/500 [30:41<3:39:52, 29.78s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 12%|█▏        | 58/500 [31:10<3:36:47, 29.43s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 12%|█▏        | 59/500 [31:39<3:34:19, 29.16s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.600 a=1.000 total=0.918
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.300 a=1.000 total=0.458
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.996 s=0.800 a=1.000 total=0.959

 12%|█▏        | 60/500 [32:07<3:32:37, 28.99s/it]
                                                  
{'loss': '-4.47e-09', 'grad_norm': '0.06104', 'learning_rate': '4.41e-05', 'num_tokens': '4.162e+05', 'completions/mean_length': '253.4', 'completions/min_length': '214.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '9.6', 'completions/min_terminated_length': '9.6', 'completions/max_terminated_length': '9.6', 'rewards/dbre_reward/mean': '0.1113', 'rewards/dbre_reward/std': '0.2741', 'reward': '0.1113', 'reward_std': '0.2741', 'frac_reward_zero_std': '0.5', 'entropy': '0.8151', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.71', 'epoch': '1.2'}

 12%|█▏        | 60/500 [32:07<3:32:37, 28.99s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959

 12%|█▏        | 61/500 [32:36<3:31:16, 28.88s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 12%|█▏        | 62/500 [33:04<3:29:45, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.800 a=1.000 total=0.958
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.516
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 13%|█▎        | 63/500 [33:33<3:28:58, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 13%|█▎        | 64/500 [34:01<3:28:28, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 13%|█▎        | 65/500 [34:30<3:28:20, 28.74s/it]
                                                  
{'loss': '-8.941e-09', 'grad_norm': '0', 'learning_rate': '4.36e-05', 'num_tokens': '4.508e+05', 'completions/mean_length': '252.2', 'completions/min_length': '194.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '41', 'completions/min_terminated_length': '41', 'completions/max_terminated_length': '41', 'rewards/dbre_reward/mean': '0.0875', 'rewards/dbre_reward/std': '0.2112', 'reward': '0.0875', 'reward_std': '0.2112', 'frac_reward_zero_std': '0.7', 'entropy': '0.931', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.77', 'epoch': '1.3'}

 13%|█▎        | 65/500 [34:30<3:28:20, 28.74s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.994 s=0.800 a=1.000 total=0.958
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 13%|█▎        | 66/500 [34:59<3:27:47, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 13%|█▎        | 67/500 [35:27<3:26:37, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 14%|█▎        | 68/500 [35:56<3:26:36, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 14%|█▍        | 69/500 [36:25<3:25:28, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.300 a=1.000 total=0.458
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 14%|█▍        | 70/500 [36:53<3:24:42, 28.56s/it]
                                                  
{'loss': '-2.235e-09', 'grad_norm': '0.05078', 'learning_rate': '4.31e-05', 'num_tokens': '4.857e+05', 'completions/mean_length': '254.7', 'completions/min_length': '235.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '30.4', 'completions/min_terminated_length': '30.4', 'completions/max_terminated_length': '30.4', 'rewards/dbre_reward/mean': '0.07392', 'rewards/dbre_reward/std': '0.1946', 'reward': '0.07392', 'reward_std': '0.1946', 'frac_reward_zero_std': '0.6', 'entropy': '0.8473', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.7', 'epoch': '1.4'}

 14%|█▍        | 70/500 [36:53<3:24:42, 28.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 14%|█▍        | 71/500 [37:22<3:24:02, 28.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.500 a=1.000 total=0.498

 14%|█▍        | 72/500 [37:50<3:23:38, 28.55s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 15%|█▍        | 73/500 [38:19<3:23:26, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 15%|█▍        | 74/500 [38:48<3:23:06, 28.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.996 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 15%|█▌        | 75/500 [39:16<3:22:38, 28.61s/it]
                                                  
{'loss': '-0.004587', 'grad_norm': '0.04688', 'learning_rate': '4.26e-05', 'num_tokens': '5.204e+05', 'completions/mean_length': '252.8', 'completions/min_length': '204.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '51', 'completions/min_terminated_length': '51', 'completions/max_terminated_length': '51', 'rewards/dbre_reward/mean': '0.09866', 'rewards/dbre_reward/std': '0.2948', 'reward': '0.09866', 'reward_std': '0.2948', 'frac_reward_zero_std': '0.4', 'entropy': '0.8822', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.72', 'epoch': '1.5'}

 15%|█▌        | 75/500 [39:16<3:22:38, 28.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 15%|█▌        | 76/500 [39:45<3:22:03, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 15%|█▌        | 77/500 [40:13<3:21:28, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518

 16%|█▌        | 78/500 [40:42<3:20:57, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 16%|█▌        | 79/500 [41:11<3:20:52, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.559
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 16%|█▌        | 80/500 [41:39<3:20:26, 28.64s/it]
                                                  
{'loss': '-4.47e-09', 'grad_norm': '0.05273', 'learning_rate': '4.21e-05', 'num_tokens': '5.553e+05', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.1363', 'rewards/dbre_reward/std': '0.3454', 'reward': '0.1363', 'reward_std': '0.3454', 'frac_reward_zero_std': '0.4', 'entropy': '0.8492', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.74', 'epoch': '1.6'}

 16%|█▌        | 80/500 [41:39<3:20:26, 28.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 16%|█▌        | 81/500 [42:08<3:20:00, 28.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 16%|█▋        | 82/500 [42:37<3:19:47, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 17%|█▋        | 83/500 [43:05<3:19:02, 28.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 17%|█▋        | 84/500 [43:34<3:18:19, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.559
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 17%|█▋        | 85/500 [44:03<3:18:30, 28.70s/it]
                                                  
{'loss': '-0.003848', 'grad_norm': '0.05347', 'learning_rate': '4.16e-05', 'num_tokens': '5.902e+05', 'completions/mean_length': '254.2', 'completions/min_length': '227.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '23', 'completions/min_terminated_length': '23', 'completions/max_terminated_length': '23', 'rewards/dbre_reward/mean': '0.1358', 'rewards/dbre_reward/std': '0.3496', 'reward': '0.1358', 'reward_std': '0.3496', 'frac_reward_zero_std': '0.4', 'entropy': '0.9009', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.81', 'epoch': '1.7'}

 17%|█▋        | 85/500 [44:03<3:18:30, 28.70s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.800 a=1.000 total=0.556
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 17%|█▋        | 86/500 [44:31<3:18:24, 28.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 17%|█▋        | 87/500 [45:00<3:18:03, 28.77s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 18%|█▊        | 88/500 [45:29<3:17:24, 28.75s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.800 a=1.000 total=0.558
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475

 18%|█▊        | 89/500 [45:58<3:16:54, 28.75s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.000 a=1.000 total=0.400
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 18%|█▊        | 90/500 [46:26<3:16:11, 28.71s/it]
                                                  
{'loss': '-0.002317', 'grad_norm': '0.08447', 'learning_rate': '4.11e-05', 'num_tokens': '6.249e+05', 'completions/mean_length': '253.7', 'completions/min_length': '236.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '33.1', 'completions/min_terminated_length': '31.8', 'completions/max_terminated_length': '34.4', 'rewards/dbre_reward/mean': '0.145', 'rewards/dbre_reward/std': '0.3356', 'reward': '0.145', 'reward_std': '0.3356', 'frac_reward_zero_std': '0.3', 'entropy': '0.975', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.87', 'epoch': '1.8'}

 18%|█▊        | 90/500 [46:26<3:16:11, 28.71s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 18%|█▊        | 91/500 [46:55<3:15:51, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 18%|█▊        | 92/500 [47:24<3:15:04, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 19%|█▊        | 93/500 [47:52<3:14:25, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 19%|█▉        | 94/500 [48:21<3:13:39, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 19%|█▉        | 95/500 [48:49<3:12:48, 28.56s/it]
                                                  
{'loss': '-0.01061', 'grad_norm': '0.05762', 'learning_rate': '4.06e-05', 'num_tokens': '6.595e+05', 'completions/mean_length': '250.8', 'completions/min_length': '199.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '60.6', 'completions/min_terminated_length': '45.6', 'completions/max_terminated_length': '75.6', 'rewards/dbre_reward/mean': '0.1351', 'rewards/dbre_reward/std': '0.3332', 'reward': '0.1351', 'reward_std': '0.3332', 'frac_reward_zero_std': '0.4', 'entropy': '0.9379', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.71', 'epoch': '1.9'}

 19%|█▉        | 95/500 [48:49<3:12:48, 28.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 19%|█▉        | 96/500 [49:18<3:12:15, 28.55s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 19%|█▉        | 97/500 [49:46<3:11:59, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 20%|█▉        | 98/500 [50:15<3:11:37, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.800 a=1.000 total=0.958
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 20%|█▉        | 99/500 [50:44<3:11:45, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.998 s=0.500 a=1.000 total=0.900
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 20%|██        | 100/500 [51:13<3:12:15, 28.84s/it]
                                                   
{'loss': '0.002807', 'grad_norm': '0.06396', 'learning_rate': '4.01e-05', 'num_tokens': '6.943e+05', 'completions/mean_length': '254.6', 'completions/min_length': '233.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '80', 'completions/min_terminated_length': '80', 'completions/max_terminated_length': '80', 'rewards/dbre_reward/mean': '0.1369', 'rewards/dbre_reward/std': '0.3314', 'reward': '0.1369', 'reward_std': '0.3314', 'frac_reward_zero_std': '0.4', 'entropy': '0.8717', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.91', 'epoch': '2'}

 20%|██        | 100/500 [51:13<3:12:15, 28.84s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 20%|██        | 101/500 [51:43<3:13:46, 29.14s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 20%|██        | 102/500 [52:12<3:12:32, 29.03s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518

 21%|██        | 103/500 [52:40<3:11:13, 28.90s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 21%|██        | 104/500 [53:09<3:10:20, 28.84s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.558
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 21%|██        | 105/500 [53:38<3:09:18, 28.76s/it]
                                                   
{'loss': '-0.004697', 'grad_norm': '0.06079', 'learning_rate': '3.96e-05', 'num_tokens': '7.29e+05', 'completions/mean_length': '251.9', 'completions/min_length': '189.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '138.6', 'completions/min_terminated_length': '138.6', 'completions/max_terminated_length': '138.6', 'rewards/dbre_reward/mean': '0.0852', 'rewards/dbre_reward/std': '0.2487', 'reward': '0.0852', 'reward_std': '0.2487', 'frac_reward_zero_std': '0.5', 'entropy': '0.9405', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.78', 'epoch': '2.1'}

 21%|██        | 105/500 [53:38<3:09:18, 28.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 21%|██        | 106/500 [54:06<3:08:06, 28.64s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.983 s=0.600 a=1.000 total=0.515
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518

 21%|██▏       | 107/500 [54:35<3:07:21, 28.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.999 s=0.500 a=1.000 total=0.900
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 22%|██▏       | 108/500 [55:03<3:06:15, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 22%|██▏       | 109/500 [55:31<3:05:42, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.800 a=1.000 total=0.556
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 22%|██▏       | 110/500 [56:00<3:05:15, 28.50s/it]
                                                   
{'loss': '-0.0075', 'grad_norm': '0.1348', 'learning_rate': '3.91e-05', 'num_tokens': '7.633e+05', 'completions/mean_length': '248.4', 'completions/min_length': '171.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '19.9', 'completions/min_terminated_length': '18', 'completions/max_terminated_length': '21.8', 'rewards/dbre_reward/mean': '0.1362', 'rewards/dbre_reward/std': '0.3244', 'reward': '0.1362', 'reward_std': '0.3244', 'frac_reward_zero_std': '0.3', 'entropy': '1.044', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.56', 'epoch': '2.2'}

 22%|██▏       | 110/500 [56:00<3:05:15, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 22%|██▏       | 111/500 [56:28<3:04:21, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.996 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 22%|██▏       | 112/500 [56:57<3:03:53, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.500 a=1.000 total=0.500
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 23%|██▎       | 113/500 [57:25<3:03:48, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 23%|██▎       | 114/500 [57:54<3:03:09, 28.47s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.996 s=0.800 a=1.000 total=0.959
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.800 a=1.000 total=0.556
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517

 23%|██▎       | 115/500 [58:22<3:02:45, 28.48s/it]
                                                   
{'loss': '-0.003744', 'grad_norm': '0.08301', 'learning_rate': '3.86e-05', 'num_tokens': '7.981e+05', 'completions/mean_length': '253.4', 'completions/min_length': '214.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '10', 'completions/min_terminated_length': '10', 'completions/max_terminated_length': '10', 'rewards/dbre_reward/mean': '0.172', 'rewards/dbre_reward/std': '0.3799', 'reward': '0.172', 'reward_std': '0.3799', 'frac_reward_zero_std': '0.2', 'entropy': '0.9914', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.58', 'epoch': '2.3'}

 23%|██▎       | 115/500 [58:22<3:02:45, 28.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=1.000 a=1.000 total=0.596
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.800 a=1.000 total=0.557
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 23%|██▎       | 116/500 [58:51<3:02:16, 28.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 23%|██▎       | 117/500 [59:19<3:01:56, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 24%|██▎       | 118/500 [59:48<3:01:27, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 24%|██▍       | 119/500 [1:00:16<3:00:35, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519

 24%|██▍       | 120/500 [1:00:44<3:00:14, 28.46s/it]
                                                     
{'loss': '-0.004933', 'grad_norm': '0.09277', 'learning_rate': '3.81e-05', 'num_tokens': '8.327e+05', 'completions/mean_length': '252.5', 'completions/min_length': '209', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '117.6', 'completions/min_terminated_length': '106.6', 'completions/max_terminated_length': '128.6', 'rewards/dbre_reward/mean': '0.1836', 'rewards/dbre_reward/std': '0.3651', 'reward': '0.1836', 'reward_std': '0.3651', 'frac_reward_zero_std': '0.2', 'entropy': '1.025', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.59', 'epoch': '2.4'}

 24%|██▍       | 120/500 [1:00:44<3:00:14, 28.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 24%|██▍       | 121/500 [1:01:13<2:59:37, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.993 s=0.300 a=1.000 total=0.858
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.995 s=0.700 a=1.000 total=0.538
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478

 24%|██▍       | 122/500 [1:01:41<2:59:21, 28.47s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 25%|██▍       | 123/500 [1:02:10<2:58:49, 28.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 25%|██▍       | 124/500 [1:02:38<2:58:19, 28.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 25%|██▌       | 125/500 [1:03:07<2:58:22, 28.54s/it]
                                                     
{'loss': '-0.004206', 'grad_norm': '0.08398', 'learning_rate': '3.76e-05', 'num_tokens': '8.675e+05', 'completions/mean_length': '253.1', 'completions/min_length': '210', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '5.2', 'completions/min_terminated_length': '5.2', 'completions/max_terminated_length': '5.2', 'rewards/dbre_reward/mean': '0.222', 'rewards/dbre_reward/std': '0.3662', 'reward': '0.222', 'reward_std': '0.3662', 'frac_reward_zero_std': '0.3', 'entropy': '0.9524', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.64', 'epoch': '2.5'}

 25%|██▌       | 125/500 [1:03:07<2:58:22, 28.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.995 s=1.000 a=1.000 total=0.999
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 25%|██▌       | 126/500 [1:03:36<2:57:57, 28.55s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 25%|██▌       | 127/500 [1:04:04<2:57:02, 28.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.700 a=1.000 total=0.539
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 26%|██▌       | 128/500 [1:04:32<2:56:48, 28.52s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.082 e=0.991 s=0.600 a=1.000 total=0.550
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 26%|██▌       | 129/500 [1:05:01<2:56:15, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 26%|██▌       | 130/500 [1:05:30<2:56:16, 28.58s/it]
                                                     
{'loss': '-0.002261', 'grad_norm': '0.06055', 'learning_rate': '3.71e-05', 'num_tokens': '9.023e+05', 'completions/mean_length': '254.9', 'completions/min_length': '239.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '34.4', 'completions/min_terminated_length': '34.4', 'completions/max_terminated_length': '34.4', 'rewards/dbre_reward/mean': '0.1958', 'rewards/dbre_reward/std': '0.3975', 'reward': '0.1958', 'reward_std': '0.3975', 'frac_reward_zero_std': '0.4', 'entropy': '0.9269', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.68', 'epoch': '2.6'}

 26%|██▌       | 130/500 [1:05:30<2:56:16, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 26%|██▌       | 131/500 [1:05:58<2:55:15, 28.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 26%|██▋       | 132/500 [1:06:27<2:55:05, 28.55s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 27%|██▋       | 133/500 [1:06:55<2:54:44, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 27%|██▋       | 134/500 [1:07:24<2:55:20, 28.74s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 27%|██▋       | 135/500 [1:07:53<2:54:43, 28.72s/it]
                                                     
{'loss': '-0.004485', 'grad_norm': '0.1025', 'learning_rate': '3.66e-05', 'num_tokens': '9.372e+05', 'completions/mean_length': '254.5', 'completions/min_length': '231.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '26.6', 'completions/min_terminated_length': '26.6', 'completions/max_terminated_length': '26.6', 'rewards/dbre_reward/mean': '0.1942', 'rewards/dbre_reward/std': '0.3968', 'reward': '0.1942', 'reward_std': '0.3968', 'frac_reward_zero_std': '0.1', 'entropy': '0.9508', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.8', 'epoch': '2.7'}

 27%|██▋       | 135/500 [1:07:53<2:54:43, 28.72s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477

 27%|██▋       | 136/500 [1:08:22<2:54:22, 28.74s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 27%|██▋       | 137/500 [1:08:51<2:53:43, 28.71s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.996 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 28%|██▊       | 138/500 [1:09:19<2:53:23, 28.74s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 28%|██▊       | 139/500 [1:09:48<2:53:00, 28.75s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518

 28%|██▊       | 140/500 [1:10:17<2:52:41, 28.78s/it]
                                                     
{'loss': '2.98e-09', 'grad_norm': '0.07812', 'learning_rate': '3.61e-05', 'num_tokens': '9.721e+05', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.1721', 'rewards/dbre_reward/std': '0.381', 'reward': '0.1721', 'reward_std': '0.381', 'frac_reward_zero_std': '0.1', 'entropy': '0.9166', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.89', 'epoch': '2.8'}

 28%|██▊       | 140/500 [1:10:17<2:52:41, 28.78s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 28%|██▊       | 141/500 [1:10:46<2:52:15, 28.79s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.999 s=0.500 a=1.000 total=0.500
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 28%|██▊       | 142/500 [1:11:14<2:51:34, 28.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.998 s=0.500 a=1.000 total=0.900
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.959 s=0.800 a=1.000 total=0.548
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 29%|██▊       | 143/500 [1:11:44<2:52:27, 28.98s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 29%|██▉       | 144/500 [1:12:13<2:51:11, 28.85s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 29%|██▉       | 145/500 [1:12:41<2:50:22, 28.80s/it]
                                                     
{'loss': '0.003438', 'grad_norm': '0.0874', 'learning_rate': '3.56e-05', 'num_tokens': '1.007e+06', 'completions/mean_length': '252.3', 'completions/min_length': '201.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '108.2', 'completions/min_terminated_length': '99.2', 'completions/max_terminated_length': '117.2', 'rewards/dbre_reward/mean': '0.2209', 'rewards/dbre_reward/std': '0.4018', 'reward': '0.2209', 'reward_std': '0.4018', 'frac_reward_zero_std': '0.2', 'entropy': '1.039', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.96', 'epoch': '2.9'}

 29%|██▉       | 145/500 [1:12:41<2:50:22, 28.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478

 29%|██▉       | 146/500 [1:13:10<2:49:56, 28.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.994 s=0.600 a=1.000 total=0.918
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.800 a=1.000 total=0.558
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.800 a=1.000 total=0.557
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.500 a=1.000 total=0.498
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.969 s=0.400 a=1.000 total=0.471

 29%|██▉       | 147/500 [1:13:47<3:03:29, 31.19s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.800 a=1.000 total=0.556
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 30%|██▉       | 148/500 [1:14:22<3:10:06, 32.41s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 30%|██▉       | 149/500 [1:14:58<3:15:05, 33.35s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.996 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.973 s=0.400 a=1.000 total=0.472
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 30%|███       | 150/500 [1:15:35<3:21:55, 34.62s/it]
                                                     
{'loss': '-0.006568', 'grad_norm': '0.09668', 'learning_rate': '3.51e-05', 'num_tokens': '1.042e+06', 'completions/mean_length': '253.1', 'completions/min_length': '220', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '27.7', 'completions/min_terminated_length': '15.2', 'completions/max_terminated_length': '40.2', 'rewards/dbre_reward/mean': '0.2542', 'rewards/dbre_reward/std': '0.432', 'reward': '0.2542', 'reward_std': '0.432', 'frac_reward_zero_std': '0.1', 'entropy': '0.9524', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.83', 'epoch': '3'}

 30%|███       | 150/500 [1:15:35<3:21:55, 34.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.994 s=0.600 a=1.000 total=0.918
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 30%|███       | 151/500 [1:16:16<3:33:00, 36.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.100 a=1.000 total=0.419
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 30%|███       | 152/500 [1:16:53<3:31:51, 36.53s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.971 s=0.400 a=1.000 total=0.471
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.971 s=0.800 a=1.000 total=0.551

 31%|███       | 153/500 [1:17:29<3:31:16, 36.53s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.500 a=1.000 total=0.497
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.500 a=1.000 total=0.498
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 31%|███       | 154/500 [1:18:05<3:29:01, 36.25s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.516
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 31%|███       | 155/500 [1:18:40<3:26:29, 35.91s/it]
                                                     
{'loss': '-0.003655', 'grad_norm': '0.07031', 'learning_rate': '3.46e-05', 'num_tokens': '1.076e+06', 'completions/mean_length': '254.5', 'completions/min_length': '236.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '134.9', 'completions/min_terminated_length': '134', 'completions/max_terminated_length': '135.8', 'rewards/dbre_reward/mean': '0.2434', 'rewards/dbre_reward/std': '0.4292', 'reward': '0.2434', 'reward_std': '0.4292', 'frac_reward_zero_std': '0.2', 'entropy': '0.9468', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '35.74', 'epoch': '3.1'}

 31%|███       | 155/500 [1:18:40<3:26:29, 35.91s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.971 s=0.400 a=1.000 total=0.471
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517

 31%|███       | 156/500 [1:19:14<3:21:55, 35.22s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.981 s=0.600 a=1.000 total=0.514
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.983 s=0.600 a=1.000 total=0.515
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 31%|███▏      | 157/500 [1:19:47<3:18:51, 34.79s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.983 s=0.600 a=1.000 total=0.515
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.800 a=1.000 total=0.556

 32%|███▏      | 158/500 [1:20:22<3:17:35, 34.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 32%|███▏      | 159/500 [1:20:56<3:15:23, 34.38s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.100 a=1.000 total=0.419
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.962 s=0.600 a=1.000 total=0.509
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 32%|███▏      | 160/500 [1:21:30<3:15:37, 34.52s/it]
                                                     
{'loss': '-0.005657', 'grad_norm': '0.06934', 'learning_rate': '3.41e-05', 'num_tokens': '1.111e+06', 'completions/mean_length': '253.4', 'completions/min_length': '215.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '10.4', 'completions/min_terminated_length': '10.4', 'completions/max_terminated_length': '10.4', 'rewards/dbre_reward/mean': '0.2539', 'rewards/dbre_reward/std': '0.4269', 'reward': '0.2539', 'reward_std': '0.4269', 'frac_reward_zero_std': '0.3', 'entropy': '0.9374', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.11', 'epoch': '3.2'}

 32%|███▏      | 160/500 [1:21:30<3:15:37, 34.52s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.993 s=0.800 a=1.000 total=0.958
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 32%|███▏      | 161/500 [1:22:05<3:14:27, 34.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.981 s=0.600 a=1.000 total=0.514
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 32%|███▏      | 162/500 [1:22:39<3:14:06, 34.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.992 s=0.800 a=1.000 total=0.958
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 33%|███▎      | 163/500 [1:23:14<3:14:17, 34.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.975 s=0.400 a=1.000 total=0.473

 33%|███▎      | 164/500 [1:23:50<3:16:20, 35.06s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.974 s=0.600 a=1.000 total=0.512
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.976 s=0.400 a=1.000 total=0.473
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.600 a=1.000 total=0.516

 33%|███▎      | 165/500 [1:24:24<3:13:47, 34.71s/it]
                                                     
{'loss': '0.001345', 'grad_norm': '0.09717', 'learning_rate': '3.36e-05', 'num_tokens': '1.145e+06', 'completions/mean_length': '244.7', 'completions/min_length': '93.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.925', 'completions/mean_terminated_length': '95.4', 'completions/min_terminated_length': '93.6', 'completions/max_terminated_length': '97.2', 'rewards/dbre_reward/mean': '0.182', 'rewards/dbre_reward/std': '0.378', 'reward': '0.182', 'reward_std': '0.378', 'frac_reward_zero_std': '0.1', 'entropy': '1.076', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.76', 'epoch': '3.3'}

 33%|███▎      | 165/500 [1:24:24<3:13:47, 34.71s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 33%|███▎      | 166/500 [1:24:59<3:12:59, 34.67s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.500 a=1.000 total=0.497

 33%|███▎      | 167/500 [1:25:32<3:10:51, 34.39s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.991 s=0.400 a=1.000 total=0.877
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 34%|███▎      | 168/500 [1:26:06<3:09:24, 34.23s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 34%|███▍      | 169/500 [1:26:43<3:13:38, 35.10s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.992 s=0.600 a=1.000 total=0.918
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.515
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 34%|███▍      | 170/500 [1:27:25<3:24:31, 37.19s/it]
                                                     
{'loss': '-0.001752', 'grad_norm': '0.09277', 'learning_rate': '3.31e-05', 'num_tokens': '1.18e+06', 'completions/mean_length': '250.3', 'completions/min_length': '168.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.925', 'completions/mean_terminated_length': '168.7', 'completions/min_terminated_length': '168.2', 'completions/max_terminated_length': '169.2', 'rewards/dbre_reward/mean': '0.2185', 'rewards/dbre_reward/std': '0.415', 'reward': '0.2185', 'reward_std': '0.415', 'frac_reward_zero_std': '0.1', 'entropy': '0.9454', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '35.3', 'epoch': '3.4'}

 34%|███▍      | 170/500 [1:27:25<3:24:31, 37.19s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.600 a=1.000 total=0.516
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 34%|███▍      | 171/500 [1:27:59<3:18:31, 36.20s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.300 a=1.000 total=0.458
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476

 34%|███▍      | 172/500 [1:28:33<3:13:08, 35.33s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.000 a=1.000 total=0.399

 35%|███▍      | 173/500 [1:29:06<3:10:10, 34.90s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.973 s=0.400 a=1.000 total=0.472
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 35%|███▍      | 174/500 [1:29:40<3:06:55, 34.40s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.992 s=0.600 a=1.000 total=0.918
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.978 s=0.400 a=1.000 total=0.473
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.800 a=1.000 total=0.556
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.979 s=0.300 a=1.000 total=0.454
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 35%|███▌      | 175/500 [1:30:13<3:04:55, 34.14s/it]
                                                     
{'loss': '-0.01071', 'grad_norm': '0.0918', 'learning_rate': '3.26e-05', 'num_tokens': '1.214e+06', 'completions/mean_length': '251.6', 'completions/min_length': '185', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '82.6', 'completions/min_terminated_length': '82.6', 'completions/max_terminated_length': '82.6', 'rewards/dbre_reward/mean': '0.2743', 'rewards/dbre_reward/std': '0.4277', 'reward': '0.2743', 'reward_std': '0.4277', 'frac_reward_zero_std': '0', 'entropy': '0.9627', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '32.61', 'epoch': '3.5'}

 35%|███▌      | 175/500 [1:30:13<3:04:55, 34.14s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.982 s=0.600 a=1.000 total=0.515
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 35%|███▌      | 176/500 [1:30:47<3:04:31, 34.17s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 35%|███▌      | 177/500 [1:31:22<3:04:18, 34.24s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.300 a=1.000 total=0.458
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 36%|███▌      | 178/500 [1:31:56<3:03:50, 34.26s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.300 a=1.000 total=0.459
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 36%|███▌      | 179/500 [1:32:29<3:01:42, 33.97s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.983 s=0.600 a=1.000 total=0.515

 36%|███▌      | 180/500 [1:33:03<3:00:49, 33.91s/it]
                                                     
{'loss': '-0.005979', 'grad_norm': '0.09912', 'learning_rate': '3.21e-05', 'num_tokens': '1.249e+06', 'completions/mean_length': '253.3', 'completions/min_length': '213', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '8.2', 'completions/min_terminated_length': '8.2', 'completions/max_terminated_length': '8.2', 'rewards/dbre_reward/mean': '0.2385', 'rewards/dbre_reward/std': '0.4241', 'reward': '0.2385', 'reward_std': '0.4241', 'frac_reward_zero_std': '0', 'entropy': '1.029', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.03', 'epoch': '3.6'}

 36%|███▌      | 180/500 [1:33:03<3:00:49, 33.91s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475

 36%|███▌      | 181/500 [1:33:37<2:59:37, 33.78s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.963 s=0.500 a=1.000 total=0.489
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.977 s=0.600 a=1.000 total=0.513
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516

 36%|███▋      | 182/500 [1:34:12<3:01:14, 34.20s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.994 s=0.600 a=1.000 total=0.918
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.978 s=0.500 a=1.000 total=0.493
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 37%|███▋      | 183/500 [1:34:48<3:03:16, 34.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.982 s=0.600 a=1.000 total=0.514
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.600 a=1.000 total=0.515
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475

 37%|███▋      | 184/500 [1:35:25<3:06:52, 35.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.600 a=1.000 total=0.516
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 37%|███▋      | 185/500 [1:36:02<3:09:00, 36.00s/it]
                                                     
{'loss': '-0.009967', 'grad_norm': '0.09619', 'learning_rate': '3.16e-05', 'num_tokens': '1.283e+06', 'completions/mean_length': '246.2', 'completions/min_length': '131', 'completions/max_length': '256', 'completions/clipped_ratio': '0.925', 'completions/mean_terminated_length': '131.9', 'completions/min_terminated_length': '131', 'completions/max_terminated_length': '132.8', 'rewards/dbre_reward/mean': '0.3372', 'rewards/dbre_reward/std': '0.4626', 'reward': '0.3372', 'reward_std': '0.4626', 'frac_reward_zero_std': '0', 'entropy': '0.9208', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '34.83', 'epoch': '3.7'}

 37%|███▋      | 185/500 [1:36:02<3:09:00, 36.00s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.979 s=0.600 a=1.000 total=0.514
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 37%|███▋      | 186/500 [1:36:38<3:08:10, 35.96s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.990 s=0.600 a=1.000 total=0.917
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.984 s=0.600 a=1.000 total=0.515
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476

 37%|███▋      | 187/500 [1:37:14<3:07:41, 35.98s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.100 a=1.000 total=0.418
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 38%|███▊      | 188/500 [1:37:51<3:08:53, 36.32s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476

 38%|███▊      | 189/500 [1:38:30<3:11:41, 36.98s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 38%|███▊      | 190/500 [1:39:08<3:13:33, 37.46s/it]
                                                     
{'loss': '0.006586', 'grad_norm': '0.09766', 'learning_rate': '3.11e-05', 'num_tokens': '1.318e+06', 'completions/mean_length': '248.9', 'completions/min_length': '170.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '45.4', 'completions/min_terminated_length': '16.6', 'completions/max_terminated_length': '74.2', 'rewards/dbre_reward/mean': '0.2771', 'rewards/dbre_reward/std': '0.4383', 'reward': '0.2771', 'reward_std': '0.4383', 'frac_reward_zero_std': '0.2', 'entropy': '1.085', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '36.25', 'epoch': '3.8'}

 38%|███▊      | 190/500 [1:39:08<3:13:33, 37.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.515
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476

 38%|███▊      | 191/500 [1:39:43<3:08:46, 36.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 38%|███▊      | 192/500 [1:40:13<2:57:05, 34.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499

 39%|███▊      | 193/500 [1:40:41<2:47:16, 32.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.500 a=1.000 total=0.500
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 39%|███▉      | 194/500 [1:41:10<2:40:15, 31.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 39%|███▉      | 195/500 [1:41:38<2:35:13, 30.54s/it]
                                                     
{'loss': '-0.01146', 'grad_norm': '0.09961', 'learning_rate': '3.06e-05', 'num_tokens': '1.352e+06', 'completions/mean_length': '249.7', 'completions/min_length': '154.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '103.4', 'completions/min_terminated_length': '103.4', 'completions/max_terminated_length': '103.4', 'rewards/dbre_reward/mean': '0.3028', 'rewards/dbre_reward/std': '0.4453', 'reward': '0.3028', 'reward_std': '0.4453', 'frac_reward_zero_std': '0', 'entropy': '0.9323', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '29.03', 'epoch': '3.9'}

 39%|███▉      | 195/500 [1:41:38<2:35:13, 30.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.515
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 39%|███▉      | 196/500 [1:42:07<2:31:35, 29.92s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518

 39%|███▉      | 197/500 [1:42:35<2:28:45, 29.46s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 40%|███▉      | 198/500 [1:43:03<2:26:36, 29.13s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517

 40%|███▉      | 199/500 [1:43:32<2:25:02, 28.91s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 40%|████      | 200/500 [1:44:00<2:23:48, 28.76s/it]
                                                     
{'loss': '-0.01771', 'grad_norm': '0.1001', 'learning_rate': '3.01e-05', 'num_tokens': '1.387e+06', 'completions/mean_length': '251', 'completions/min_length': '175.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '21.8', 'completions/min_terminated_length': '21.8', 'completions/max_terminated_length': '21.8', 'rewards/dbre_reward/mean': '0.2792', 'rewards/dbre_reward/std': '0.4371', 'reward': '0.2792', 'reward_std': '0.4371', 'frac_reward_zero_std': '0', 'entropy': '1.012', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.53', 'epoch': '4'}

 40%|████      | 200/500 [1:44:00<2:23:48, 28.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.100 a=1.000 total=0.419
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 40%|████      | 201/500 [1:44:29<2:24:10, 28.93s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.100 a=1.000 total=0.419
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 40%|████      | 202/500 [1:44:58<2:23:01, 28.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.200 a=1.000 total=0.439
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 41%|████      | 203/500 [1:45:26<2:22:12, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 41%|████      | 204/500 [1:45:55<2:21:22, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476

 41%|████      | 205/500 [1:46:24<2:20:50, 28.65s/it]
                                                     
{'loss': '0.006697', 'grad_norm': '0.09082', 'learning_rate': '2.96e-05', 'num_tokens': '1.421e+06', 'completions/mean_length': '254.3', 'completions/min_length': '229.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '24.4', 'completions/min_terminated_length': '24.4', 'completions/max_terminated_length': '24.4', 'rewards/dbre_reward/mean': '0.3339', 'rewards/dbre_reward/std': '0.4616', 'reward': '0.3339', 'reward_std': '0.4616', 'frac_reward_zero_std': '0', 'entropy': '1.001', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.63', 'epoch': '4.1'}

 41%|████      | 205/500 [1:46:24<2:20:50, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 41%|████      | 206/500 [1:46:52<2:20:17, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479

 41%|████▏     | 207/500 [1:47:21<2:19:42, 28.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 42%|████▏     | 208/500 [1:47:49<2:19:06, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 42%|████▏     | 209/500 [1:48:18<2:18:31, 28.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.999 s=0.000 a=1.000 total=0.400
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.300 a=1.000 total=0.456
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.558

 42%|████▏     | 210/500 [1:48:46<2:18:05, 28.57s/it]
                                                     
{'loss': '-5.96e-09', 'grad_norm': '0.1104', 'learning_rate': '2.91e-05', 'num_tokens': '1.456e+06', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.3235', 'rewards/dbre_reward/std': '0.4596', 'reward': '0.3235', 'reward_std': '0.4596', 'frac_reward_zero_std': '0', 'entropy': '1.023', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.68', 'epoch': '4.2'}

 42%|████▏     | 210/500 [1:48:46<2:18:05, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 42%|████▏     | 211/500 [1:49:15<2:17:17, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.999 s=0.600 a=1.000 total=0.920
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.500 a=1.000 total=0.498
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 42%|████▏     | 212/500 [1:49:43<2:16:53, 28.52s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.999 s=0.300 a=1.000 total=0.460
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 43%|████▎     | 213/500 [1:50:12<2:16:44, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 43%|████▎     | 214/500 [1:50:41<2:16:26, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 43%|████▎     | 215/500 [1:51:09<2:15:50, 28.60s/it]
                                                     
{'loss': '-0.004874', 'grad_norm': '0.07666', 'learning_rate': '2.86e-05', 'num_tokens': '1.49e+06', 'completions/mean_length': '244.6', 'completions/min_length': '125.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9', 'completions/mean_terminated_length': '113.6', 'completions/min_terminated_length': '74', 'completions/max_terminated_length': '153.2', 'rewards/dbre_reward/mean': '0.2298', 'rewards/dbre_reward/std': '0.4038', 'reward': '0.2298', 'reward_std': '0.4038', 'frac_reward_zero_std': '0.3', 'entropy': '1.047', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.7', 'epoch': '4.3'}

 43%|████▎     | 215/500 [1:51:09<2:15:50, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.800 a=1.000 total=0.558
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 43%|████▎     | 216/500 [1:51:38<2:15:20, 28.59s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478

 43%|████▎     | 217/500 [1:52:06<2:14:46, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.100 a=1.000 total=0.416
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 44%|████▎     | 218/500 [1:52:35<2:14:19, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.600 a=1.000 total=0.520
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 44%|████▍     | 219/500 [1:53:04<2:13:51, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 44%|████▍     | 220/500 [1:53:32<2:13:21, 28.58s/it]
                                                     
{'loss': '0.001157', 'grad_norm': '0.105', 'learning_rate': '2.81e-05', 'num_tokens': '1.525e+06', 'completions/mean_length': '255.7', 'completions/min_length': '251.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '46.4', 'completions/min_terminated_length': '46.4', 'completions/max_terminated_length': '46.4', 'rewards/dbre_reward/mean': '0.3366', 'rewards/dbre_reward/std': '0.4732', 'reward': '0.3366', 'reward_std': '0.4732', 'frac_reward_zero_std': '0', 'entropy': '1.035', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.69', 'epoch': '4.4'}

 44%|████▍     | 220/500 [1:53:32<2:13:21, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.988 s=0.800 a=1.000 total=0.556
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 44%|████▍     | 221/500 [1:54:01<2:12:54, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919

 44%|████▍     | 222/500 [1:54:29<2:12:21, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 45%|████▍     | 223/500 [1:54:58<2:11:37, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 45%|████▍     | 224/500 [1:55:26<2:11:08, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 45%|████▌     | 225/500 [1:55:55<2:10:46, 28.53s/it]
                                                     
{'loss': '-0.004409', 'grad_norm': '0.1064', 'learning_rate': '2.76e-05', 'num_tokens': '1.56e+06', 'completions/mean_length': '254.5', 'completions/min_length': '231.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '27', 'completions/min_terminated_length': '27', 'completions/max_terminated_length': '27', 'rewards/dbre_reward/mean': '0.3252', 'rewards/dbre_reward/std': '0.4448', 'reward': '0.3252', 'reward_std': '0.4448', 'frac_reward_zero_std': '0.1', 'entropy': '1.002', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.64', 'epoch': '4.5'}

 45%|████▌     | 225/500 [1:55:55<2:10:46, 28.53s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 45%|████▌     | 226/500 [1:56:23<2:10:19, 28.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477

 45%|████▌     | 227/500 [1:56:52<2:09:53, 28.55s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479

 46%|████▌     | 228/500 [1:57:20<2:09:33, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 46%|████▌     | 229/500 [1:57:49<2:09:09, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 46%|████▌     | 230/500 [1:58:18<2:09:43, 28.83s/it]
                                                     
{'loss': '-0.00292', 'grad_norm': '0.09375', 'learning_rate': '2.71e-05', 'num_tokens': '1.595e+06', 'completions/mean_length': '255.2', 'completions/min_length': '243.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '38.6', 'completions/min_terminated_length': '38.6', 'completions/max_terminated_length': '38.6', 'rewards/dbre_reward/mean': '0.2889', 'rewards/dbre_reward/std': '0.4446', 'reward': '0.2889', 'reward_std': '0.4446', 'frac_reward_zero_std': '0.1', 'entropy': '0.9833', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.88', 'epoch': '4.6'}

 46%|████▌     | 230/500 [1:58:18<2:09:43, 28.83s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 46%|████▌     | 231/500 [1:58:49<2:11:01, 29.22s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 46%|████▋     | 232/500 [1:59:18<2:10:58, 29.32s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519

 47%|████▋     | 233/500 [1:59:47<2:10:20, 29.29s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.994 s=0.600 a=1.000 total=0.918
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 47%|████▋     | 234/500 [2:00:16<2:09:04, 29.11s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 47%|████▋     | 235/500 [2:00:45<2:08:26, 29.08s/it]
                                                     
{'loss': '-0.01724', 'grad_norm': '0.1055', 'learning_rate': '2.66e-05', 'num_tokens': '1.629e+06', 'completions/mean_length': '242.7', 'completions/min_length': '123.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.875', 'completions/mean_terminated_length': '157.2', 'completions/min_terminated_length': '123.4', 'completions/max_terminated_length': '181.2', 'rewards/dbre_reward/mean': '0.3132', 'rewards/dbre_reward/std': '0.4637', 'reward': '0.3132', 'reward_std': '0.4637', 'frac_reward_zero_std': '0.1', 'entropy': '1.049', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '28.43', 'epoch': '4.7'}

 47%|████▋     | 235/500 [2:00:45<2:08:26, 29.08s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.500 a=1.000 total=0.498
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 47%|████▋     | 236/500 [2:01:14<2:08:04, 29.11s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499

 47%|████▋     | 237/500 [2:01:43<2:06:52, 28.94s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.100 a=1.000 total=0.418
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478

 48%|████▊     | 238/500 [2:02:11<2:05:59, 28.85s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.300 a=1.000 total=0.458

 48%|████▊     | 239/500 [2:02:40<2:05:32, 28.86s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 48%|████▊     | 240/500 [2:03:09<2:04:50, 28.81s/it]
                                                     
{'loss': '0.0005952', 'grad_norm': '0.1108', 'learning_rate': '2.61e-05', 'num_tokens': '1.664e+06', 'completions/mean_length': '252.7', 'completions/min_length': '203.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '49.6', 'completions/min_terminated_length': '49.6', 'completions/max_terminated_length': '49.6', 'rewards/dbre_reward/mean': '0.4192', 'rewards/dbre_reward/std': '0.4799', 'reward': '0.4192', 'reward_std': '0.4799', 'frac_reward_zero_std': '0', 'entropy': '0.9967', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.9', 'epoch': '4.8'}

 48%|████▊     | 240/500 [2:03:09<2:04:50, 28.81s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 48%|████▊     | 241/500 [2:03:38<2:05:04, 28.97s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477

 48%|████▊     | 242/500 [2:04:08<2:04:48, 29.03s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.997 s=0.800 a=1.000 total=0.959

 49%|████▊     | 243/500 [2:04:36<2:03:50, 28.91s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 49%|████▉     | 244/500 [2:05:05<2:03:00, 28.83s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 49%|████▉     | 245/500 [2:05:33<2:02:11, 28.75s/it]
                                                     
{'loss': '-0.004135', 'grad_norm': '0.08398', 'learning_rate': '2.56e-05', 'num_tokens': '1.698e+06', 'completions/mean_length': '250.9', 'completions/min_length': '186.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '38.3', 'completions/min_terminated_length': '32.8', 'completions/max_terminated_length': '43.8', 'rewards/dbre_reward/mean': '0.3234', 'rewards/dbre_reward/std': '0.4586', 'reward': '0.3234', 'reward_std': '0.4586', 'frac_reward_zero_std': '0', 'entropy': '0.978', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.98', 'epoch': '4.9'}

 49%|████▉     | 245/500 [2:05:33<2:02:11, 28.75s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.600 a=1.000 total=0.520
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 49%|████▉     | 246/500 [2:06:02<2:02:06, 28.84s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.100 a=1.000 total=0.420
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478

 49%|████▉     | 247/500 [2:06:31<2:01:24, 28.79s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.800 a=1.000 total=0.555
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 50%|████▉     | 248/500 [2:07:00<2:00:38, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 50%|████▉     | 249/500 [2:07:28<1:59:47, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 50%|█████     | 250/500 [2:07:57<2:00:11, 28.85s/it]
                                                     
{'loss': '0.002627', 'grad_norm': '0.105', 'learning_rate': '2.51e-05', 'num_tokens': '1.733e+06', 'completions/mean_length': '251.9', 'completions/min_length': '191.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '47', 'completions/min_terminated_length': '38.2', 'completions/max_terminated_length': '55.8', 'rewards/dbre_reward/mean': '0.3704', 'rewards/dbre_reward/std': '0.4648', 'reward': '0.3704', 'reward_std': '0.4648', 'frac_reward_zero_std': '0.1', 'entropy': '0.9665', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.93', 'epoch': '5'}

 50%|█████     | 250/500 [2:07:57<2:00:11, 28.85s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 50%|█████     | 251/500 [2:08:27<2:01:04, 29.18s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.800 a=1.000 total=0.559
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 50%|█████     | 252/500 [2:08:56<2:00:02, 29.04s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 51%|█████     | 253/500 [2:09:25<1:59:27, 29.02s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.300 a=1.000 total=0.457
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.100 a=1.000 total=0.418

 51%|█████     | 254/500 [2:09:54<1:59:01, 29.03s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 51%|█████     | 255/500 [2:10:23<1:58:11, 28.94s/it]
                                                     
{'loss': '0.02185', 'grad_norm': '0.1167', 'learning_rate': '2.46e-05', 'num_tokens': '1.768e+06', 'completions/mean_length': '251.6', 'completions/min_length': '193.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '49.5', 'completions/min_terminated_length': '39.8', 'completions/max_terminated_length': '59.2', 'rewards/dbre_reward/mean': '0.3832', 'rewards/dbre_reward/std': '0.4825', 'reward': '0.3832', 'reward_std': '0.4825', 'frac_reward_zero_std': '0.1', 'entropy': '0.9532', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '28.02', 'epoch': '5.1'}

 51%|█████     | 255/500 [2:10:23<1:58:11, 28.94s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 51%|█████     | 256/500 [2:10:51<1:57:09, 28.81s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.520
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 51%|█████▏    | 257/500 [2:11:20<1:56:24, 28.74s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 52%|█████▏    | 258/500 [2:11:49<1:55:42, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 52%|█████▏    | 259/500 [2:12:17<1:55:23, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 52%|█████▏    | 260/500 [2:12:46<1:54:46, 28.69s/it]
                                                     
{'loss': '-0.003285', 'grad_norm': '0.1016', 'learning_rate': '2.41e-05', 'num_tokens': '1.802e+06', 'completions/mean_length': '255.3', 'completions/min_length': '245', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '40.2', 'completions/min_terminated_length': '40.2', 'completions/max_terminated_length': '40.2', 'rewards/dbre_reward/mean': '0.4074', 'rewards/dbre_reward/std': '0.4871', 'reward': '0.4074', 'reward_std': '0.4871', 'frac_reward_zero_std': '0', 'entropy': '0.9447', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.73', 'epoch': '5.2'}

 52%|█████▏    | 260/500 [2:12:46<1:54:46, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 52%|█████▏    | 261/500 [2:13:15<1:54:27, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.100 a=1.000 total=0.419

 52%|█████▏    | 262/500 [2:13:43<1:53:55, 28.72s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 53%|█████▎    | 263/500 [2:14:12<1:53:04, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 53%|█████▎    | 264/500 [2:14:41<1:52:47, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.999 s=0.300 a=1.000 total=0.460
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 53%|█████▎    | 265/500 [2:15:09<1:52:19, 28.68s/it]
                                                     
{'loss': '0.01128', 'grad_norm': '0.09863', 'learning_rate': '2.36e-05', 'num_tokens': '1.837e+06', 'completions/mean_length': '252.5', 'completions/min_length': '199.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '97', 'completions/min_terminated_length': '97', 'completions/max_terminated_length': '97', 'rewards/dbre_reward/mean': '0.3811', 'rewards/dbre_reward/std': '0.4779', 'reward': '0.3811', 'reward_std': '0.4779', 'frac_reward_zero_std': '0.1', 'entropy': '1.006', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.8', 'epoch': '5.3'}

 53%|█████▎    | 265/500 [2:15:09<1:52:19, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 53%|█████▎    | 266/500 [2:15:38<1:51:50, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479

 53%|█████▎    | 267/500 [2:16:07<1:51:16, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 54%|█████▎    | 268/500 [2:16:35<1:50:52, 28.67s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 54%|█████▍    | 269/500 [2:17:04<1:50:28, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.998 s=0.100 a=1.000 total=0.419
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 54%|█████▍    | 270/500 [2:17:33<1:49:59, 28.69s/it]
                                                     
{'loss': '0.006578', 'grad_norm': '0.09717', 'learning_rate': '2.31e-05', 'num_tokens': '1.872e+06', 'completions/mean_length': '254.3', 'completions/min_length': '228.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '24', 'completions/min_terminated_length': '24', 'completions/max_terminated_length': '24', 'rewards/dbre_reward/mean': '0.4185', 'rewards/dbre_reward/std': '0.484', 'reward': '0.4185', 'reward_std': '0.484', 'frac_reward_zero_std': '0', 'entropy': '0.9434', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.8', 'epoch': '5.4'}

 54%|█████▍    | 270/500 [2:17:33<1:49:59, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 54%|█████▍    | 271/500 [2:18:02<1:49:39, 28.73s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.980 s=0.500 a=1.000 total=0.494
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 54%|█████▍    | 272/500 [2:18:30<1:49:04, 28.71s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 55%|█████▍    | 273/500 [2:18:59<1:48:55, 28.79s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 55%|█████▍    | 274/500 [2:19:28<1:48:26, 28.79s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.500 a=1.000 total=0.498
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919

 55%|█████▌    | 275/500 [2:19:57<1:47:54, 28.77s/it]
                                                     
{'loss': '-5.026e-05', 'grad_norm': '0.1064', 'learning_rate': '2.26e-05', 'num_tokens': '1.907e+06', 'completions/mean_length': '252.8', 'completions/min_length': '211.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.925', 'completions/mean_terminated_length': '117.6', 'completions/min_terminated_length': '109.4', 'completions/max_terminated_length': '125.6', 'rewards/dbre_reward/mean': '0.374', 'rewards/dbre_reward/std': '0.4772', 'reward': '0.374', 'reward_std': '0.4772', 'frac_reward_zero_std': '0', 'entropy': '1.008', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.9', 'epoch': '5.5'}

 55%|█████▌    | 275/500 [2:19:57<1:47:54, 28.77s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.998 s=0.500 a=1.000 total=0.900
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 55%|█████▌    | 276/500 [2:20:25<1:47:13, 28.72s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 55%|█████▌    | 277/500 [2:20:54<1:46:35, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.995 s=0.600 a=1.000 total=0.918
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.700 a=1.000 total=0.539
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 56%|█████▌    | 278/500 [2:21:23<1:46:01, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 56%|█████▌    | 279/500 [2:21:51<1:45:41, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 56%|█████▌    | 280/500 [2:22:20<1:45:02, 28.65s/it]
                                                     
{'loss': '-1.788e-08', 'grad_norm': '0.1025', 'learning_rate': '2.21e-05', 'num_tokens': '1.942e+06', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.421', 'rewards/dbre_reward/std': '0.4867', 'reward': '0.421', 'reward_std': '0.4867', 'frac_reward_zero_std': '0', 'entropy': '0.9705', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.74', 'epoch': '5.6'}

 56%|█████▌    | 280/500 [2:22:20<1:45:02, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 56%|█████▌    | 281/500 [2:22:49<1:44:42, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 56%|█████▋    | 282/500 [2:23:17<1:44:06, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 57%|█████▋    | 283/500 [2:23:46<1:43:50, 28.71s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 57%|█████▋    | 284/500 [2:24:15<1:43:07, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 57%|█████▋    | 285/500 [2:24:43<1:42:35, 28.63s/it]
                                                     
{'loss': '-0.005169', 'grad_norm': '0.1016', 'learning_rate': '2.16e-05', 'num_tokens': '1.976e+06', 'completions/mean_length': '253.9', 'completions/min_length': '222.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '68.6', 'completions/min_terminated_length': '68.6', 'completions/max_terminated_length': '68.6', 'rewards/dbre_reward/mean': '0.3837', 'rewards/dbre_reward/std': '0.4813', 'reward': '0.3837', 'reward_std': '0.4813', 'frac_reward_zero_std': '0', 'entropy': '0.924', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.77', 'epoch': '5.7'}

 57%|█████▋    | 285/500 [2:24:43<1:42:35, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=1.000 e=0.998 s=0.500 a=1.000 total=0.900

 57%|█████▋    | 286/500 [2:25:12<1:42:10, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.480
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 57%|█████▋    | 287/500 [2:25:41<1:42:08, 28.77s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 58%|█████▊    | 288/500 [2:26:09<1:41:25, 28.70s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 58%|█████▊    | 289/500 [2:26:38<1:40:41, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 58%|█████▊    | 290/500 [2:27:06<1:40:00, 28.58s/it]
                                                     
{'loss': '-0.003315', 'grad_norm': '0.1064', 'learning_rate': '2.11e-05', 'num_tokens': '2.011e+06', 'completions/mean_length': '254.5', 'completions/min_length': '231.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '26.8', 'completions/min_terminated_length': '26.8', 'completions/max_terminated_length': '26.8', 'rewards/dbre_reward/mean': '0.3967', 'rewards/dbre_reward/std': '0.4796', 'reward': '0.3967', 'reward_std': '0.4796', 'frac_reward_zero_std': '0', 'entropy': '1.049', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.76', 'epoch': '5.8'}

 58%|█████▊    | 290/500 [2:27:06<1:40:00, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 58%|█████▊    | 291/500 [2:27:35<1:39:29, 28.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 58%|█████▊    | 292/500 [2:28:03<1:38:50, 28.51s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 59%|█████▊    | 293/500 [2:28:32<1:38:16, 28.49s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.100 a=1.000 total=0.418
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477

 59%|█████▉    | 294/500 [2:29:00<1:37:49, 28.49s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 59%|█████▉    | 295/500 [2:29:29<1:37:16, 28.47s/it]
                                                     
{'loss': '-2.98e-09', 'grad_norm': '0.1001', 'learning_rate': '2.06e-05', 'num_tokens': '2.046e+06', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.4173', 'rewards/dbre_reward/std': '0.4842', 'reward': '0.4173', 'reward_std': '0.4842', 'frac_reward_zero_std': '0', 'entropy': '0.9523', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.58', 'epoch': '5.9'}

 59%|█████▉    | 295/500 [2:29:29<1:37:16, 28.47s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.800 a=1.000 total=0.559
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 59%|█████▉    | 296/500 [2:29:57<1:36:40, 28.43s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 59%|█████▉    | 297/500 [2:30:25<1:36:09, 28.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.100 a=1.000 total=0.418
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.300 a=1.000 total=0.458
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518

 60%|█████▉    | 298/500 [2:30:54<1:35:45, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 60%|█████▉    | 299/500 [2:31:22<1:35:16, 28.44s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 60%|██████    | 300/500 [2:31:51<1:34:44, 28.42s/it]
                                                     
{'loss': '0.004358', 'grad_norm': '0.1016', 'learning_rate': '2.01e-05', 'num_tokens': '2.081e+06', 'completions/mean_length': '253.9', 'completions/min_length': '222', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '68.4', 'completions/min_terminated_length': '68.4', 'completions/max_terminated_length': '68.4', 'rewards/dbre_reward/mean': '0.4049', 'rewards/dbre_reward/std': '0.4861', 'reward': '0.4049', 'reward_std': '0.4861', 'frac_reward_zero_std': '0', 'entropy': '1.028', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.53', 'epoch': '6'}

 60%|██████    | 300/500 [2:31:51<1:34:44, 28.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477

 60%|██████    | 301/500 [2:32:20<1:35:09, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479

 60%|██████    | 302/500 [2:32:49<1:34:32, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 61%|██████    | 303/500 [2:33:17<1:34:04, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.700 a=1.000 total=0.539
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 61%|██████    | 304/500 [2:33:46<1:33:28, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 61%|██████    | 305/500 [2:34:14<1:32:58, 28.61s/it]
                                                     
{'loss': '-0.02022', 'grad_norm': '0.09473', 'learning_rate': '1.96e-05', 'num_tokens': '2.116e+06', 'completions/mean_length': '250.9', 'completions/min_length': '178.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '89', 'completions/min_terminated_length': '76.4', 'completions/max_terminated_length': '101.6', 'rewards/dbre_reward/mean': '0.4432', 'rewards/dbre_reward/std': '0.4895', 'reward': '0.4432', 'reward_std': '0.4895', 'frac_reward_zero_std': '0', 'entropy': '0.9466', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.66', 'epoch': '6.1'}

 61%|██████    | 305/500 [2:34:14<1:32:58, 28.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 61%|██████    | 306/500 [2:34:43<1:32:22, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.976 s=0.400 a=1.000 total=0.473
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 61%|██████▏   | 307/500 [2:35:12<1:32:12, 28.67s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 62%|██████▏   | 308/500 [2:35:40<1:31:43, 28.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 62%|██████▏   | 309/500 [2:36:09<1:31:07, 28.63s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=1.000 e=0.998 s=0.600 a=1.000 total=0.919
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 62%|██████▏   | 310/500 [2:36:37<1:30:29, 28.58s/it]
                                                     
{'loss': '-0.001285', 'grad_norm': '0.1001', 'learning_rate': '1.91e-05', 'num_tokens': '2.15e+06', 'completions/mean_length': '255.6', 'completions/min_length': '248.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '44', 'completions/min_terminated_length': '44', 'completions/max_terminated_length': '44', 'rewards/dbre_reward/mean': '0.3469', 'rewards/dbre_reward/std': '0.4704', 'reward': '0.3469', 'reward_std': '0.4704', 'frac_reward_zero_std': '0', 'entropy': '1.027', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.73', 'epoch': '6.2'}

 62%|██████▏   | 310/500 [2:36:37<1:30:29, 28.58s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 62%|██████▏   | 311/500 [2:37:06<1:29:59, 28.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.100 a=1.000 total=0.419
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 62%|██████▏   | 312/500 [2:37:34<1:29:29, 28.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 63%|██████▎   | 313/500 [2:38:03<1:29:07, 28.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.800 a=1.000 total=0.557
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 63%|██████▎   | 314/500 [2:38:32<1:28:43, 28.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.999 s=0.500 a=1.000 total=0.500

 63%|██████▎   | 315/500 [2:39:01<1:28:20, 28.65s/it]
                                                     
{'loss': '0.001368', 'grad_norm': '0.1045', 'learning_rate': '1.86e-05', 'num_tokens': '2.185e+06', 'completions/mean_length': '250.4', 'completions/min_length': '167', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '115.8', 'completions/min_terminated_length': '115.8', 'completions/max_terminated_length': '115.8', 'rewards/dbre_reward/mean': '0.4655', 'rewards/dbre_reward/std': '0.4876', 'reward': '0.4655', 'reward_std': '0.4876', 'frac_reward_zero_std': '0', 'entropy': '0.9481', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.75', 'epoch': '6.3'}

 63%|██████▎   | 315/500 [2:39:01<1:28:20, 28.65s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 63%|██████▎   | 316/500 [2:39:29<1:27:56, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477

 63%|██████▎   | 317/500 [2:39:58<1:27:31, 28.70s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.516
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 64%|██████▎   | 318/500 [2:40:27<1:26:59, 28.68s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.300 a=1.000 total=0.459
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 64%|██████▍   | 319/500 [2:40:55<1:26:32, 28.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.460
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 64%|██████▍   | 320/500 [2:41:25<1:26:57, 28.98s/it]
                                                     
{'loss': '-0.006759', 'grad_norm': '0.1055', 'learning_rate': '1.81e-05', 'num_tokens': '2.219e+06', 'completions/mean_length': '249.6', 'completions/min_length': '153.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '51.2', 'completions/min_terminated_length': '51.2', 'completions/max_terminated_length': '51.2', 'rewards/dbre_reward/mean': '0.3813', 'rewards/dbre_reward/std': '0.4784', 'reward': '0.3813', 'reward_std': '0.4784', 'frac_reward_zero_std': '0', 'entropy': '0.947', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.99', 'epoch': '6.4'}

 64%|██████▍   | 320/500 [2:41:25<1:26:57, 28.98s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 64%|██████▍   | 321/500 [2:42:02<1:33:57, 31.49s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.978 s=0.400 a=1.000 total=0.473
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476

 64%|██████▍   | 322/500 [2:42:38<1:37:18, 32.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 65%|██████▍   | 323/500 [2:43:14<1:39:42, 33.80s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476

 65%|██████▍   | 324/500 [2:43:49<1:39:45, 34.01s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474

 65%|██████▌   | 325/500 [2:44:28<1:44:00, 35.66s/it]
                                                     
{'loss': '2.98e-09', 'grad_norm': '0.09619', 'learning_rate': '1.76e-05', 'num_tokens': '2.254e+06', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.441', 'rewards/dbre_reward/std': '0.4835', 'reward': '0.441', 'reward_std': '0.4835', 'frac_reward_zero_std': '0', 'entropy': '1.006', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '35.7', 'epoch': '6.5'}

 65%|██████▌   | 325/500 [2:44:28<1:44:00, 35.66s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.000 a=1.000 total=0.398
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.985 s=0.600 a=1.000 total=0.515
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 65%|██████▌   | 326/500 [2:45:05<1:43:54, 35.83s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 65%|██████▌   | 327/500 [2:45:39<1:41:48, 35.31s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.600 a=1.000 total=0.515
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474

 66%|██████▌   | 328/500 [2:46:13<1:40:11, 34.95s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.100 a=1.000 total=0.417
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477

 66%|██████▌   | 329/500 [2:46:48<1:39:26, 34.89s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.975 s=0.400 a=1.000 total=0.472
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 66%|██████▌   | 330/500 [2:47:25<1:40:41, 35.54s/it]
                                                     
{'loss': '8.941e-09', 'grad_norm': '0.1172', 'learning_rate': '1.71e-05', 'num_tokens': '2.289e+06', 'completions/mean_length': '256', 'completions/min_length': '256', 'completions/max_length': '256', 'completions/clipped_ratio': '1', 'completions/mean_terminated_length': '0', 'completions/min_terminated_length': '0', 'completions/max_terminated_length': '0', 'rewards/dbre_reward/mean': '0.3665', 'rewards/dbre_reward/std': '0.4638', 'reward': '0.3665', 'reward_std': '0.4638', 'frac_reward_zero_std': '0', 'entropy': '1.004', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '34.28', 'epoch': '6.6'}

 66%|██████▌   | 330/500 [2:47:25<1:40:41, 35.54s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476

 66%|██████▌   | 331/500 [2:48:02<1:42:00, 36.21s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.100 a=1.000 total=0.418
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476

 66%|██████▋   | 332/500 [2:48:36<1:39:28, 35.53s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.984 s=0.600 a=1.000 total=0.515
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477

 67%|██████▋   | 333/500 [2:49:10<1:37:22, 34.98s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 67%|██████▋   | 334/500 [2:49:44<1:35:45, 34.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.978 s=0.400 a=1.000 total=0.473
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475

 67%|██████▋   | 335/500 [2:50:20<1:36:43, 35.17s/it]
                                                     
{'loss': '-0.001411', 'grad_norm': '0.104', 'learning_rate': '1.66e-05', 'num_tokens': '2.324e+06', 'completions/mean_length': '254.1', 'completions/min_length': '226', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '72.4', 'completions/min_terminated_length': '72.4', 'completions/max_terminated_length': '72.4', 'rewards/dbre_reward/mean': '0.4385', 'rewards/dbre_reward/std': '0.4876', 'reward': '0.4385', 'reward_std': '0.4876', 'frac_reward_zero_std': '0', 'entropy': '1.06', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '34.16', 'epoch': '6.7'}

 67%|██████▋   | 335/500 [2:50:20<1:36:43, 35.17s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.995 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.200 a=1.000 total=0.438
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.082 e=0.983 s=0.400 a=1.000 total=0.507
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 67%|██████▋   | 336/500 [2:50:57<1:37:17, 35.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.500 a=1.000 total=0.498
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.964 s=0.400 a=1.000 total=0.469
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.968 s=0.400 a=1.000 total=0.470
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 67%|██████▋   | 337/500 [2:51:33<1:37:24, 35.86s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 68%|██████▊   | 338/500 [2:52:09<1:36:32, 35.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 68%|██████▊   | 339/500 [2:52:42<1:34:11, 35.10s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.978 s=0.400 a=1.000 total=0.473
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 68%|██████▊   | 340/500 [2:53:21<1:36:03, 36.02s/it]
                                                     
{'loss': '-0.003443', 'grad_norm': '0.1035', 'learning_rate': '1.61e-05', 'num_tokens': '2.359e+06', 'completions/mean_length': '250.4', 'completions/min_length': '181.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '85.8', 'completions/min_terminated_length': '79.2', 'completions/max_terminated_length': '92.4', 'rewards/dbre_reward/mean': '0.4145', 'rewards/dbre_reward/std': '0.4793', 'reward': '0.4145', 'reward_std': '0.4793', 'frac_reward_zero_std': '0.1', 'entropy': '1.05', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '35.09', 'epoch': '6.8'}

 68%|██████▊   | 340/500 [2:53:21<1:36:03, 36.02s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.500 a=1.000 total=0.497
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477

 68%|██████▊   | 341/500 [2:53:55<1:34:23, 35.62s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.978 s=0.400 a=1.000 total=0.473
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 68%|██████▊   | 342/500 [2:54:30<1:32:46, 35.23s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 69%|██████▊   | 343/500 [2:55:04<1:31:30, 34.97s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476

 69%|██████▉   | 344/500 [2:55:38<1:30:24, 34.77s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.995 s=0.500 a=1.000 total=0.899
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 69%|██████▉   | 345/500 [2:56:13<1:29:33, 34.67s/it]
                                                     
{'loss': '-0.008482', 'grad_norm': '0.105', 'learning_rate': '1.56e-05', 'num_tokens': '2.393e+06', 'completions/mean_length': '251.8', 'completions/min_length': '199.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9625', 'completions/mean_terminated_length': '55.7', 'completions/min_terminated_length': '46', 'completions/max_terminated_length': '65.4', 'rewards/dbre_reward/mean': '0.4771', 'rewards/dbre_reward/std': '0.4884', 'reward': '0.4771', 'reward_std': '0.4884', 'frac_reward_zero_std': '0', 'entropy': '0.9596', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.45', 'epoch': '6.9'}

 69%|██████▉   | 345/500 [2:56:13<1:29:33, 34.67s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 69%|██████▉   | 346/500 [2:56:47<1:28:48, 34.60s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 69%|██████▉   | 347/500 [2:57:22<1:28:09, 34.57s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.983 s=0.100 a=1.000 total=0.415
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 70%|██████▉   | 348/500 [2:57:56<1:27:19, 34.47s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 70%|██████▉   | 349/500 [2:58:30<1:26:49, 34.50s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.995 s=0.300 a=1.000 total=0.459

 70%|███████   | 350/500 [2:59:04<1:25:52, 34.35s/it]
                                                     
{'loss': '-0.002433', 'grad_norm': '0.105', 'learning_rate': '1.51e-05', 'num_tokens': '2.428e+06', 'completions/mean_length': '254.2', 'completions/min_length': '227.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '36.8', 'completions/min_terminated_length': '22.6', 'completions/max_terminated_length': '51', 'rewards/dbre_reward/mean': '0.3437', 'rewards/dbre_reward/std': '0.4671', 'reward': '0.3437', 'reward_std': '0.4671', 'frac_reward_zero_std': '0', 'entropy': '1.091', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.38', 'epoch': '7'}

 70%|███████   | 350/500 [2:59:04<1:25:52, 34.35s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 70%|███████   | 351/500 [2:59:42<1:27:26, 35.21s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.300 a=1.000 total=0.459
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 70%|███████   | 352/500 [3:00:16<1:26:15, 34.97s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 71%|███████   | 353/500 [3:00:51<1:25:21, 34.84s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 71%|███████   | 354/500 [3:01:25<1:24:21, 34.67s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.500 a=1.000 total=0.497
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.100 a=1.000 total=0.419
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 71%|███████   | 355/500 [3:01:59<1:23:31, 34.56s/it]
                                                     
{'loss': '-0.005638', 'grad_norm': '0.105', 'learning_rate': '1.46e-05', 'num_tokens': '2.463e+06', 'completions/mean_length': '253.2', 'completions/min_length': '234.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.95', 'completions/mean_terminated_length': '83.4', 'completions/min_terminated_length': '80.6', 'completions/max_terminated_length': '86.4', 'rewards/dbre_reward/mean': '0.3677', 'rewards/dbre_reward/std': '0.4754', 'reward': '0.3677', 'reward_std': '0.4754', 'frac_reward_zero_std': '0', 'entropy': '1.032', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.67', 'epoch': '7.1'}

 71%|███████   | 355/500 [3:01:59<1:23:31, 34.56s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477

 71%|███████   | 356/500 [3:02:33<1:22:44, 34.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.977 s=0.400 a=1.000 total=0.473
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.300 a=1.000 total=0.458
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476

 71%|███████▏  | 357/500 [3:03:08<1:21:54, 34.37s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.979 s=0.400 a=1.000 total=0.474
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 72%|███████▏  | 358/500 [3:03:42<1:21:19, 34.36s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.500 a=1.000 total=0.498
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.475
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.975 s=0.400 a=1.000 total=0.473
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475

 72%|███████▏  | 359/500 [3:04:16<1:20:48, 34.39s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.981 s=0.400 a=1.000 total=0.474
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 72%|███████▏  | 360/500 [3:04:51<1:20:19, 34.42s/it]
                                                     
{'loss': '0.002459', 'grad_norm': '0.1021', 'learning_rate': '1.41e-05', 'num_tokens': '2.498e+06', 'completions/mean_length': '254.3', 'completions/min_length': '228.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '74.8', 'completions/min_terminated_length': '74.8', 'completions/max_terminated_length': '74.8', 'rewards/dbre_reward/mean': '0.4031', 'rewards/dbre_reward/std': '0.4822', 'reward': '0.4031', 'reward_std': '0.4822', 'frac_reward_zero_std': '0', 'entropy': '0.982', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.37', 'epoch': '7.2'}

 72%|███████▏  | 360/500 [3:04:51<1:20:19, 34.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 72%|███████▏  | 361/500 [3:05:25<1:19:46, 34.43s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.983 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476

 72%|███████▏  | 362/500 [3:06:00<1:19:06, 34.39s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.600 a=1.000 total=0.516
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476

 73%|███████▎  | 363/500 [3:06:34<1:18:29, 34.37s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 73%|███████▎  | 364/500 [3:07:08<1:17:54, 34.37s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 73%|███████▎  | 365/500 [3:07:43<1:17:22, 34.39s/it]
                                                     
{'loss': '-0.005127', 'grad_norm': '0.1035', 'learning_rate': '1.36e-05', 'num_tokens': '2.532e+06', 'completions/mean_length': '254.2', 'completions/min_length': '228', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '23.2', 'completions/min_terminated_length': '23.2', 'completions/max_terminated_length': '23.2', 'rewards/dbre_reward/mean': '0.4164', 'rewards/dbre_reward/std': '0.486', 'reward': '0.4164', 'reward_std': '0.486', 'frac_reward_zero_std': '0', 'entropy': '0.9915', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.41', 'epoch': '7.3'}

 73%|███████▎  | 365/500 [3:07:43<1:17:22, 34.39s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.500 a=1.000 total=0.499
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 73%|███████▎  | 366/500 [3:08:17<1:16:43, 34.35s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.982 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 73%|███████▎  | 367/500 [3:08:52<1:16:15, 34.41s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 74%|███████▎  | 368/500 [3:09:26<1:15:42, 34.42s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.100 a=1.000 total=0.419
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.300 a=1.000 total=0.458
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.300 a=1.000 total=0.456
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 74%|███████▍  | 369/500 [3:10:00<1:15:02, 34.37s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 74%|███████▍  | 370/500 [3:10:35<1:14:25, 34.35s/it]
                                                     
{'loss': '-0.006533', 'grad_norm': '0.1167', 'learning_rate': '1.31e-05', 'num_tokens': '2.567e+06', 'completions/mean_length': '253.8', 'completions/min_length': '221.4', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '67.8', 'completions/min_terminated_length': '67.8', 'completions/max_terminated_length': '67.8', 'rewards/dbre_reward/mean': '0.4027', 'rewards/dbre_reward/std': '0.4834', 'reward': '0.4027', 'reward_std': '0.4834', 'frac_reward_zero_std': '0', 'entropy': '1.042', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.39', 'epoch': '7.4'}

 74%|███████▍  | 370/500 [3:10:35<1:14:25, 34.35s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.976 s=0.400 a=1.000 total=0.473
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476

 74%|███████▍  | 371/500 [3:11:09<1:13:41, 34.28s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 74%|███████▍  | 372/500 [3:11:43<1:13:04, 34.26s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 75%|███████▍  | 373/500 [3:12:17<1:12:32, 34.27s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.600 a=1.000 total=0.517
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 75%|███████▍  | 374/500 [3:12:51<1:11:56, 34.25s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.988 s=0.600 a=1.000 total=0.516

 75%|███████▌  | 375/500 [3:13:26<1:11:26, 34.30s/it]
                                                     
{'loss': '-0.01167', 'grad_norm': '0.1011', 'learning_rate': '1.26e-05', 'num_tokens': '2.602e+06', 'completions/mean_length': '252.8', 'completions/min_length': '205.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '52', 'completions/min_terminated_length': '52', 'completions/max_terminated_length': '52', 'rewards/dbre_reward/mean': '0.3453', 'rewards/dbre_reward/std': '0.4609', 'reward': '0.3453', 'reward_std': '0.4609', 'frac_reward_zero_std': '0', 'entropy': '0.9213', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '33.28', 'epoch': '7.5'}

 75%|███████▌  | 375/500 [3:13:26<1:11:26, 34.30s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=1.000 e=0.997 s=0.600 a=1.000 total=0.919
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 75%|███████▌  | 376/500 [3:14:00<1:11:04, 34.39s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 75%|███████▌  | 377/500 [3:14:30<1:07:38, 32.99s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.600 a=1.000 total=0.517
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477

 76%|███████▌  | 378/500 [3:14:59<1:04:41, 31.81s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 76%|███████▌  | 379/500 [3:15:29<1:03:00, 31.24s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.800 a=1.000 total=0.557
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.995 s=0.600 a=1.000 total=0.519
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 76%|███████▌  | 380/500 [3:15:59<1:01:46, 30.88s/it]
                                                     
{'loss': '0.003405', 'grad_norm': '0.1021', 'learning_rate': '1.21e-05', 'num_tokens': '2.637e+06', 'completions/mean_length': '254.4', 'completions/min_length': '230', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '25.2', 'completions/min_terminated_length': '25.2', 'completions/max_terminated_length': '25.2', 'rewards/dbre_reward/mean': '0.4568', 'rewards/dbre_reward/std': '0.4889', 'reward': '0.4568', 'reward_std': '0.4889', 'frac_reward_zero_std': '0', 'entropy': '1.017', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '29.77', 'epoch': '7.6'}

 76%|███████▌  | 380/500 [3:15:59<1:01:46, 30.88s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.998 s=0.300 a=1.000 total=0.459
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 76%|███████▌  | 381/500 [3:16:28<59:46, 30.14s/it]  Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 76%|███████▋  | 382/500 [3:16:56<58:22, 29.69s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.980 s=0.400 a=1.000 total=0.474
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 77%|███████▋  | 383/500 [3:17:31<1:00:50, 31.20s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.997 s=0.600 a=1.000 total=0.519
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 77%|███████▋  | 384/500 [3:18:00<59:00, 30.52s/it]  Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 77%|███████▋  | 385/500 [3:18:31<58:58, 30.77s/it]
                                                   
{'loss': '0.005464', 'grad_norm': '0.1045', 'learning_rate': '1.16e-05', 'num_tokens': '2.672e+06', 'completions/mean_length': '253.5', 'completions/min_length': '216.2', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '11.4', 'completions/min_terminated_length': '11.4', 'completions/max_terminated_length': '11.4', 'rewards/dbre_reward/mean': '0.3811', 'rewards/dbre_reward/std': '0.4744', 'reward': '0.3811', 'reward_std': '0.4744', 'frac_reward_zero_std': '0.1', 'entropy': '1.111', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '29.51', 'epoch': '7.7'}

 77%|███████▋  | 385/500 [3:18:31<58:58, 30.77s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.986 s=0.400 a=1.000 total=0.476
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478

 77%|███████▋  | 386/500 [3:19:04<59:22, 31.25s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.987 s=0.400 a=1.000 total=0.476
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.984 s=0.600 a=1.000 total=0.515
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.985 s=0.400 a=1.000 total=0.475
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.988 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477

 77%|███████▋  | 387/500 [3:19:37<1:00:05, 31.91s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479

 78%|███████▊  | 388/500 [3:20:05<57:31, 30.82s/it]  Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.989 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.300 a=1.000 total=0.459
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 78%|███████▊  | 389/500 [3:20:36<56:54, 30.76s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477

 78%|███████▊  | 390/500 [3:21:06<55:53, 30.48s/it]
                                                   
{'loss': '0.0019', 'grad_norm': '0.09424', 'learning_rate': '1.11e-05', 'num_tokens': '2.706e+06', 'completions/mean_length': '255.6', 'completions/min_length': '249.6', 'completions/max_length': '256', 'completions/clipped_ratio': '0.9875', 'completions/mean_terminated_length': '44.8', 'completions/min_terminated_length': '44.8', 'completions/max_terminated_length': '44.8', 'rewards/dbre_reward/mean': '0.4406', 'rewards/dbre_reward/std': '0.4896', 'reward': '0.4406', 'reward_std': '0.4896', 'frac_reward_zero_std': '0', 'entropy': '1.013', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '30.01', 'epoch': '7.8'}

 78%|███████▊  | 390/500 [3:21:06<55:53, 30.48s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.998 s=0.600 a=1.000 total=0.520
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 78%|███████▊  | 391/500 [3:21:35<54:39, 30.09s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 78%|███████▊  | 392/500 [3:22:03<53:17, 29.61s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.998 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478

 79%|███████▊  | 393/500 [3:22:32<52:11, 29.27s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.994 s=0.600 a=1.000 total=0.518
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.992 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 79%|███████▉  | 394/500 [3:23:00<51:19, 29.05s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.984 s=0.400 a=1.000 total=0.475
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.999 s=0.300 a=1.000 total=0.460
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478

 79%|███████▉  | 395/500 [3:23:29<50:32, 28.88s/it]
                                                   
{'loss': '0.01185', 'grad_norm': '0.1064', 'learning_rate': '1.06e-05', 'num_tokens': '2.741e+06', 'completions/mean_length': '254', 'completions/min_length': '223.8', 'completions/max_length': '256', 'completions/clipped_ratio': '0.975', 'completions/mean_terminated_length': '70.2', 'completions/min_terminated_length': '70.2', 'completions/max_terminated_length': '70.2', 'rewards/dbre_reward/mean': '0.4067', 'rewards/dbre_reward/std': '0.4831', 'reward': '0.4067', 'reward_std': '0.4831', 'frac_reward_zero_std': '0', 'entropy': '1.036', 'clip_ratio/low_mean': '0', 'clip_ratio/low_min': '0', 'clip_ratio/high_mean': '0', 'clip_ratio/high_max': '0', 'clip_ratio/region_mean': '0', 'step_time': '27.75', 'epoch': '7.9'}

 79%|███████▉  | 395/500 [3:23:29<50:32, 28.88s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN city FROM customers
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP INDEX idx_customers_email
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 79%|███████▉  | 396/500 [3:23:59<50:38, 29.22s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.992 s=0.600 a=1.000 total=0.518
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.990 s=0.400 a=1.000 total=0.477
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.997 s=0.400 a=1.000 total=0.479
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_2291 TO products
[REWARD] c=0.000 e=0.996 s=0.400 a=1.000 total=0.479
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id

 79%|███████▉  | 397/500 [3:24:27<49:48, 29.02s/it]Applied drift: ADD COLUMN drift_col_5506 TO customers
[REWARD] c=0.000 e=0.994 s=0.400 a=1.000 total=0.478
Applied drift: DROP COLUMN city FROM customers
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
Applied drift: CREATE INDEX idx_customers_email ON customers(email)
[REWARD] c=0.000 e=0.995 s=0.400 a=1.000 total=0.478
Applied drift: ADD COLUMN drift_col_4811 TO customers
Applied drift: DROP INDEX idx_customers_email
[REWARD] c=0.000 e=0.991 s=0.400 a=1.000 total=0.477
Applied drift: DROP COLUMN status FROM orders
Applied drift: CREATE INDEX idx_reviews_customer_id ON reviews(customer_id)
Applied drift: ADD COLUMN drift_col_7924 TO customers
Applied drift: RENAME COLUMN customers.city TO city_v320
Applied drift: ADD COLUMN drift_col_7224 TO customers
Applied drift: RENAME COLUMN products.category TO category_v718
Applied drift: ADD COLUMN drift_col_8527 TO orders
Applied drift: ADD COLUMN drift_col_2291 TO products
Applied drift: RENAME COLUMN orders.status TO status_v733
Applied drift: DROP INDEX idx_reviews_customer_id
[REWARD] c=0.000 e=0.993 s=0.400 a=1.000 total=0.478
Traceback (most recent call last):
  File "/home/zeroij/round2/train.py", line 81, in <module>
    trainer.train()
    ~~~~~~~~~~~~~^^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/transformers/trainer.py", line 1427, in train
    return inner_training_loop(
        args=args,
    ...<2 lines>...
        ignore_keys_for_eval=ignore_keys_for_eval,
    )
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/transformers/trainer.py", line 1509, in _inner_training_loop
    self._run_epoch(
    ~~~~~~~~~~~~~~~^
        model=model,
        ^^^^^^^^^^^^
    ...<9 lines>...
        steps_trained_in_current_epoch=steps_trained_in_current_epoch,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/transformers/trainer.py", line 1737, in _run_epoch
    tr_loss_step = self.training_step(model, inputs, num_items_in_batch)
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/trl/trainer/grpo_trainer.py", line 1113, in training_step
    output = super().training_step(model, inputs, num_items_in_batch)
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/transformers/trainer.py", line 1937, in training_step
    self.accelerator.backward(loss, **kwargs)
    ~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/accelerate/accelerator.py", line 2838, in backward
    loss.backward(**kwargs)
    ~~~~~~~~~~~~~^^^^^^^^^^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/torch/_tensor.py", line 631, in backward
    torch.autograd.backward(
    ~~~~~~~~~~~~~~~~~~~~~~~^
        self, gradient, retain_graph, create_graph, inputs=inputs
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/torch/autograd/__init__.py", line 381, in backward
    _engine_run_backward(
    ~~~~~~~~~~~~~~~~~~~~^
        tensors,
        ^^^^^^^^
    ...<5 lines>...
        accumulate_grad=True,
        ^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/home/zeroij/.local/share/mise/installs/python/3.14.4/lib/python3.14/site-packages/torch/autograd/graph.py", line 869, in _engine_run_backward
    return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        t_outputs, *args, **kwargs
        ^^^^^^^^^^^^^^^^^^^^^^^^^^
    )  # Calls into the C++ engine to run the backward pass
    ^
KeyboardInterrupt