Instructions to use Elliott/LUFFY-Qwen-Instruct-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Elliott/LUFFY-Qwen-Instruct-7B with Transformers:

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

pipe = pipeline("text-generation", model="Elliott/LUFFY-Qwen-Instruct-7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Elliott/LUFFY-Qwen-Instruct-7B")
model = AutoModelForCausalLM.from_pretrained("Elliott/LUFFY-Qwen-Instruct-7B")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Elliott/LUFFY-Qwen-Instruct-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Elliott/LUFFY-Qwen-Instruct-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Elliott/LUFFY-Qwen-Instruct-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Elliott/LUFFY-Qwen-Instruct-7B

SGLang

How to use Elliott/LUFFY-Qwen-Instruct-7B 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 "Elliott/LUFFY-Qwen-Instruct-7B" \
    --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": "Elliott/LUFFY-Qwen-Instruct-7B",
		"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 "Elliott/LUFFY-Qwen-Instruct-7B" \
        --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": "Elliott/LUFFY-Qwen-Instruct-7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Elliott/LUFFY-Qwen-Instruct-7B with Docker Model Runner:
```
docker model run hf.co/Elliott/LUFFY-Qwen-Instruct-7B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

📖Introduction

LUFFY is a reinforcement learning framework that bridges the gap between zero-RL and imitation learning by incorporating off-policy reasoning traces into the training process. Built upon GRPO, LUFFY combines on-policy rollouts with off-policy demonstrations during advantage estimation and introduces policy shaping via regularized importance sampling to emphasize low-probability yet crucial actions.

Key Highlights:

Off-Policy Guidance: Seamlessly integrates external reasoning traces to bootstrap learning from stronger models.
Dynamic Balance: Learns when to imitate and when to explore, adapting over the course of training.
Policy Shaping: Emphasizes important actions often ignored in standard policy gradients, enabling better generalization.

Inference

Here’s an example of using LUFFY for inference:

from transformers import AutoTokenizer
from vllm import LLM, SamplingParams

model_path="Elliott/LUFFY-Qwen-Math-7B-Zero"

question = "which number is larger? 9.11 or 9.9?"

tokenizer = AutoTokenizer.from_pretrained(model_path)
messages = [{"role": "user", "content": question}]
chat = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

llm = LLM(model=model_path)
params = SamplingParams(temperature=0.6, max_tokens=8192)
outputs = llm.generate([chat], params)
print(outputs[0].outputs[0].text)

📃Evaluation

Model	AIME 2024	AIME 2025	AMC	MATH-500	Minerva	Olympiad	Avg.
Qwen2.5-7B-Instruct	11.9	7.6	44.1	74.6	30.5	39.7	34.7
LUFFY-Qwen-Instruct-7B	16.6	15.7	52.2	81.4	36.8	48.7	41.9

🌻Acknowledgement

LUFFY builds upon veRL and deepscaler, and utilizes vLLM for inference. We utilize Math-Verify for math reasoning evaluation. We thank the open-source community for datasets and backbones, including NuminaMath, OpenR1-Math-220k, Qwen2.5-Math, and DeepSeek-R1 model.

Code: https://github.com/ElliottYan/LUFFY

Citation

If you find our model, data, or evaluation code useful, please kindly cite our paper:

@misc{luffy,
      title={Learning to Reason under Off-Policy Guidance}, 
      author={Jianhao Yan and Yafu Li and Zican Hu and Zhi Wang and Ganqu Cui and Xiaoye Qu and Yu Cheng and Yue Zhang},
      year={2025},
      eprint={2504.14945},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2504.14945}, 
}