Instructions to use omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321 with Transformers:

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

pipe = pipeline("image-text-to-text", model="omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321")
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("omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321")
model = AutoModelForImageTextToText.from_pretrained("omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321")
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 omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321",
		"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/omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321

SGLang

How to use omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321 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 "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321" \
    --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": "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321",
		"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 "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321" \
        --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": "omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321",
		"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 omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321 with Docker Model Runner:
```
docker model run hf.co/omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321
```

Thinking trace reproducibility

by stumbledparams - opened Mar 1

Discussion

stumbledparams

Mar 1

•

edited Mar 1

Hi,

I was trying to reproduce some of the thinking trace using inference with this model `VLM-R1-Qwen2.5VL-3B-OVD-0321` for OVD on the D3 dataset -- but the thinking traces are not extensive as reported in the paper. Is there a sample code to follow?
I am using this prompt:

def build_ovd_prompt(labels):
#VLM-R1
lbl = "\n- " + "\n- ".join(labels) # paper uses a list format for targets
q = (
f"Please carefully check the image and detect the following objects: {lbl}. "
"Output each detected target's bbox coordinates in JSON format."
"The format of the bbox coordinates is:\n"
"`json\n" '[{"bbox_2d": [x1, y1, x2, y2], "label": "target name"},\n' ' {"bbox_2d": [x1, y1, x2, y2], "label": "target name"}]\n' "`\n"
"If there are no such targets in the image, simply respond with None."
"Output the thinking process in and final answer in tags."
)
return q

Thanks,

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment

Thinking trace reproducibility

I was trying to reproduce some of the thinking trace using inference with this model VLM-R1-Qwen2.5VL-3B-OVD-0321 for OVD on the D3 dataset -- but the thinking traces are not extensive as reported in the paper. Is there a sample code to follow?I am using this prompt:

I was trying to reproduce some of the thinking trace using inference with this model `VLM-R1-Qwen2.5VL-3B-OVD-0321` for OVD on the D3 dataset -- but the thinking traces are not extensive as reported in the paper. Is there a sample code to follow?
I am using this prompt: