Instructions to use daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100 with Transformers:

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

pipe = pipeline("text-generation", model="daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100")
model = AutoModelForMultimodalLM.from_pretrained("daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100")
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 daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100

SGLang

How to use daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100 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 "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100" \
    --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": "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100",
		"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 "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100" \
        --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": "daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100 with Docker Model Runner:
```
docker model run hf.co/daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100
```

Model Card for PrAg-PO DeepSeek-R1-Distill-Qwen-1.5B step1100

If you are using the model, a star to our github repo would be really appreciated! 😊

This is the step 1100 checkpoint when training DeepSeek-R1-Distill-Qwen-1.5B on MATH Level-3-to-5 Dataset using PrAg-PO. The training procedure is outlined in the paper PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning.

Model Sources

Repository 🤖: https://github.com/wenquanlu/PrAg-PO
Paper 📝: PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning

Uses

This model is intended for mathematical reasoning tasks. It leverages prompt augmentation to generate reasoning traces under diverse templates, increasing rollout diversity and stability during RL training.

Results

Citation

@misc{lu2026pragpopromptaugmentedpolicy,
      title={PrAg-PO: Prompt Augmented Policy Optimization for Robust and Diverse Mathematical Reasoning}, 
      author={Wenquan Lu and Hai Huang and Enqi Liu and Randall Balestriero},
      journal={arXiv preprint arXiv:2602.03190},
      url={https://arxiv.org/abs/2602.03190},
      year={2026},
}

Downloads last month: 44

Safetensors

Model size

2B params

Tensor type

F32

Model tree for daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100

Base model

deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B

Finetuned

(635)

this model

Paper for daviddavidlu/PrAg-PO-DeepSeek-R1-Distill-Qwen-1.5B-step1100

Prompt Augmentation Scales up GRPO Training on Mathematical Reasoning

Paper • 2602.03190 • Published Feb 3 • 1