metadata

library_name: transformers
tags:
  - orpo
  - llama 3
license: other
datasets:
  - mlabonne/orpo-dpo-mix-40k
language:
  - en

OrpoLlama-3-8B

This is a quick fine-tune of meta-llama/Meta-Llama-3-8B on 1k samples of mlabonne/orpo-dpo-mix-40k created for this article.

It's not very good at the moment (it's the sassiest model ever), but I'm currently training a version on the entire dataset.

Try the demo: https://huggingface.co/spaces/mlabonne/OrpoLlama-3-8B

🏆 Evaluation

Nous

Evaluation performed using LLM AutoEval, see the entire leaderboard here.

Model	Average	AGIEval	GPT4All	TruthfulQA	Bigbench
teknium/OpenHermes-2.5-Mistral-7B 📄	52.42	42.75	72.99	52.99	40.94
meta-llama/Meta-Llama-3-8B-Instruct 📄	51.34	41.22	69.86	51.65	42.64
mistralai/Mistral-7B-Instruct-v0.1 📄	49.15	33.36	67.87	55.89	39.48
mlabonne/OrpoLlama-3-8B 📄	46.76	31.56	70.19	48.11	37.17
meta-llama/Meta-Llama-3-8B 📄	45.42	31.1	69.95	43.91	36.7

📈 Training curves

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mlabonne/OrpoLlama-3-8B"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])