Instructions to use JierunChen/SFT-RL-SynergyDilemma-Model_Merging with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use JierunChen/SFT-RL-SynergyDilemma-Model_Merging with Transformers:

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

pipe = pipeline("image-text-to-text", model="JierunChen/SFT-RL-SynergyDilemma-Model_Merging")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("JierunChen/SFT-RL-SynergyDilemma-Model_Merging")
model = AutoModelForImageTextToText.from_pretrained("JierunChen/SFT-RL-SynergyDilemma-Model_Merging")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use JierunChen/SFT-RL-SynergyDilemma-Model_Merging with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "JierunChen/SFT-RL-SynergyDilemma-Model_Merging"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JierunChen/SFT-RL-SynergyDilemma-Model_Merging",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/JierunChen/SFT-RL-SynergyDilemma-Model_Merging

SGLang

How to use JierunChen/SFT-RL-SynergyDilemma-Model_Merging 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 "JierunChen/SFT-RL-SynergyDilemma-Model_Merging" \
    --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": "JierunChen/SFT-RL-SynergyDilemma-Model_Merging",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "JierunChen/SFT-RL-SynergyDilemma-Model_Merging" \
        --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": "JierunChen/SFT-RL-SynergyDilemma-Model_Merging",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use JierunChen/SFT-RL-SynergyDilemma-Model_Merging with Docker Model Runner:
```
docker model run hf.co/JierunChen/SFT-RL-SynergyDilemma-Model_Merging
```

This repository contains the fine-tuned Qwen2.5-VL-7B-Instruct model using our model merging approach, as detailed in our paper The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMs.

Citation

If you find this project useful in your research, please consider citing this BibTex:

@misc{chen2025synergydilemmalongcotsft,
      title={The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMs}, 
      author={Jierun Chen and Tiezheng Yu and Haoli Bai and Lewei Yao and Jiannan Wu and Kaican Li and Fei Mi and Chaofan Tao and Lei Zhu and Manyi Zhang and Xiaohui Li and Lu Hou and Lifeng Shang and Qun Liu},
      year={2025},
      eprint={2507.07562},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2507.07562}, 
}

Downloads last month: 5

Safetensors

Model size

8B params

Tensor type

F16

Model tree for JierunChen/SFT-RL-SynergyDilemma-Model_Merging

Base model

Qwen/Qwen2.5-VL-7B-Instruct

Finetuned

(1067)

this model

Paper for JierunChen/SFT-RL-SynergyDilemma-Model_Merging

The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMs

Paper • 2507.07562 • Published Jul 10, 2025 • 1