train_copa_1754648732

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the copa dataset. It achieves the following results on the evaluation set:

• Loss: 1.3193
• Num Input Tokens Seen: 281856

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 123
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.2698	0.5	45	0.2507	14016
0.2395	1.0	90	0.3042	28096
0.1339	1.5	135	0.1602	42144
0.0821	2.0	180	0.0774	56128
0.0255	2.5	225	0.1003	70272
0.0231	3.0	270	0.0559	84352
0.0061	3.5	315	0.1132	98464
0.0033	4.0	360	0.0911	112576
0.0059	4.5	405	0.0957	126624
0.0002	5.0	450	0.1163	140832
0.0	5.5	495	0.1131	154976
0.0	6.0	540	0.1138	169056
0.0001	6.5	585	0.1122	183200
0.0	7.0	630	0.1181	197344
0.0	7.5	675	0.1132	211392
0.0	8.0	720	0.1171	225536
0.0	8.5	765	0.1130	239680
0.0	9.0	810	0.1171	253696
0.0	9.5	855	0.1125	267840
0.0	10.0	900	0.1186	281856

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.1