Instructions to use RLHFlow/Llama3-SFT-v2.0-epoch3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RLHFlow/Llama3-SFT-v2.0-epoch3 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="RLHFlow/Llama3-SFT-v2.0-epoch3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RLHFlow/Llama3-SFT-v2.0-epoch3")
model = AutoModelForCausalLM.from_pretrained("RLHFlow/Llama3-SFT-v2.0-epoch3", device_map="auto")
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 Settings

vLLM

How to use RLHFlow/Llama3-SFT-v2.0-epoch3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RLHFlow/Llama3-SFT-v2.0-epoch3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RLHFlow/Llama3-SFT-v2.0-epoch3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RLHFlow/Llama3-SFT-v2.0-epoch3

SGLang

How to use RLHFlow/Llama3-SFT-v2.0-epoch3 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 "RLHFlow/Llama3-SFT-v2.0-epoch3" \
    --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": "RLHFlow/Llama3-SFT-v2.0-epoch3",
		"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 "RLHFlow/Llama3-SFT-v2.0-epoch3" \
        --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": "RLHFlow/Llama3-SFT-v2.0-epoch3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RLHFlow/Llama3-SFT-v2.0-epoch3 with Docker Model Runner:
```
docker model run hf.co/RLHFlow/Llama3-SFT-v2.0-epoch3
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This is the SFT checkpoint used for the project RLHFlow/Online-RLHF

Paper: RLHF Workflow: From Reward Modeling to Online RLHF (Published in TMLR, 2024)
Authors: Hanze Dong*, Wei Xiong*, Bo Pang*, Haoxiang Wang*, Han Zhao, Yingbo Zhou, Nan Jiang, Doyen Sahoo, Caiming Xiong, Tong Zhang
Code: https://github.com/RLHFlow/Online-RLHF

The model is trained from meta-llama/Meta-Llama-3-8B on RLHFlow/RLHFlow-SFT-Dataset-ver2 for 3 epochs. We use a global batch size of 128 and a learning rate of 2e-5, where we pack the samples and split them into chunks of 8192 token. See more training details at https://github.com/RLHFlow/Online-RLHF/blob/main/sft/llama3-8b-it.yaml .

Citation

Please cite our techical report if you find our model is useful for your research or product.

@misc{dong2024rlhf,
      title={RLHF Workflow: From Reward Modeling to Online RLHF}, 
      author={Hanze Dong and Wei Xiong and Bo Pang and Haoxiang Wang and Han Zhao and Yingbo Zhou and Nan Jiang and Doyen Sahoo and Caiming Xiong and Tong Zhang},
      year={2024},
      eprint={2405.07863},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Downloads last month: 69

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for RLHFlow/Llama3-SFT-v2.0-epoch3

Quantizations

2 models

Collection including RLHFlow/Llama3-SFT-v2.0-epoch3

SFT Models

Collection

We train a series of SFT models on the high-quality SFT dataset of RLHFlow for research purpose. • 6 items • Updated Nov 3, 2024 • 2

Paper for RLHFlow/Llama3-SFT-v2.0-epoch3

RLHF Workflow: From Reward Modeling to Online RLHF

Paper • 2405.07863 • Published May 13, 2024 • 71