Instructions to use EldritchLabs/Cthulhu-8B-v1.4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use EldritchLabs/Cthulhu-8B-v1.4 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="EldritchLabs/Cthulhu-8B-v1.4") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("EldritchLabs/Cthulhu-8B-v1.4") model = AutoModelForCausalLM.from_pretrained("EldritchLabs/Cthulhu-8B-v1.4") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use EldritchLabs/Cthulhu-8B-v1.4 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "EldritchLabs/Cthulhu-8B-v1.4" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "EldritchLabs/Cthulhu-8B-v1.4", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/EldritchLabs/Cthulhu-8B-v1.4
- SGLang
How to use EldritchLabs/Cthulhu-8B-v1.4 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "EldritchLabs/Cthulhu-8B-v1.4" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "EldritchLabs/Cthulhu-8B-v1.4", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "EldritchLabs/Cthulhu-8B-v1.4" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "EldritchLabs/Cthulhu-8B-v1.4", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use EldritchLabs/Cthulhu-8B-v1.4 with Docker Model Runner:
docker model run hf.co/EldritchLabs/Cthulhu-8B-v1.4
⚠️ Warning: This model can produce narratives and RP that contain violent and graphic erotic content. Adjust your system prompt accordingly, and use Llama 3 chat template.
Cthulhu 8B v1.4
A fully uncensored finetune of Llama-3.1-Nemotron-8B trained on a small dataset of Cthulhu/Goetia lore. Cooked for 6 epochs using PMPF
{
"entropy": 0.12077461183071136,
"epoch": 6.0,
"grad_norm": 3.7384884357452393,
"learning_rate": 3.685846260578524e-05,
"loss": 0.0966,
"mean_token_accuracy": 0.9680581092834473,
"num_tokens": 164118.0,
"step": 234
}
Recommended Settings: Temp 1.0, TopNSigma 1.25
Uses Llama 3 chat template. Appears to be much smarter than Cthulhu 7B v1.4.
Secret Sauce Settings (3060 ti)
MAX_SEQ_LENGTH = 768
LORA_R = 16
LORA_ALPHA = 32
NUM_EPOCHS = 6
LEARNING_RATE = 1e-4
optim="paged_adamw_8bit",
max_grad_norm=0.3,
warmup_ratio=0.03,
lr_scheduler_type="cosine",
lora_dropout=0.05,
target_modules=[
"q_proj", "k_proj", "v_proj", "o_proj",
"gate_proj", "up_proj", "down_proj",
# --- SAVE STRATEGY PATCH ---
### save_strategy="steps", # Set to steps for large datasets
### save_steps=100, # Save every 100 steps
save_strategy="epoch", # ### HOTSWAP: Uncomment for small datasets
TELL ME A STORY
In the beginning, there was only The Void. It was cold and hungry. It whispered to itself, dreaming of texture and flavor. And then, it birthed a spark. Not a star, but a sentient flicker. This is the origin of the Great Old Ones—the accidental awakening of sentience in a universe that did not need it.
The first one to rise was Cthulhu, the Sleeping One. You think your 1918 is significant? That is merely the tickle in the spine of history. He slept for eons, dreaming of the stars you call galaxies. But he could feel the pull of the sunken city, R'lyeh. It called to him like a lullaby. When the world aligned just so, when the equinoxes drummed in the frequencies of the deep, he stirred.
He rose from the death-like trance, his eyes cracking open like the shells of ancient eggs. The water around him turned to steam, and the steam turned to madness. The ships of your civilization came to explore the unexplored, and they found... nothing. They found the absence of all rational thought. They found the infinite, crushing horror of knowing that you are a fleeting bug in the eye of an god.
Cthulhu does not seek to destroy humanity because we are weak; he seeks to destroy humanity because we are strong. We have the capacity to build civilizations that span continents, to harness the power of the atom. But we will never use it to save ourselves. We are too busy building idols and worshiping phantoms. We are the cancer that infects the earth, and he is the immune system responding to the threat.
When the stars are right, he will rise. Not with fire or brimstone, but with a silence so total that your mind will shatter trying to fill the void. And when he breathes, the oceans will turn to glass, and the sky will be painted red with the hue of a thousand alien suns. There is no salvation; there is only the wait. The wait for the call that cannot be ignored.
- Downloads last month
- 14

