Instructions to use Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO with Transformers:

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

pipe = pipeline("text-generation", model="Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO")
model = AutoModelForCausalLM.from_pretrained("Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO", 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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO

SGLang

How to use Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO 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 "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO" \
    --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": "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO",
		"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 "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO" \
        --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": "Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO with Docker Model Runner:
```
docker model run hf.co/Jinhe/ReflectRL-Qwen2.5-Math-7B-DAPO
```

ReflectRL-Qwen2.5-Math-7B-GRPO-ReflectRL

This model checkpoint is part of the work presented in ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct Reasoning.

ReflectRL is a lightweight framework for learning from Golden Negative Trajectories (GNTs) during on-policy post-training. Instead of imitating failed expert trajectories directly, ReflectRL uses them as reflective context during training and gradually transitions the policy back to direct reasoning for inference.

Paper: ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct Reasoning
Repository: GitHub Repository

Citation

@article{reflectrl2027,
  title={ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct Reasoning},
  author={Jinhe Bi and Chennan Zhou and Zengjie Jin and Aniri and Shuo Lu and Wenke Huang and Hu Cao and Xun Xiao and Zhihong Zhu and Volker Tresp and Fei Shen and Yunpu Ma and Tat-Seng Chua},
  year={2026}
}