Instructions to use webbrain-one/webbrain-vl-2-450M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use webbrain-one/webbrain-vl-2-450M with Transformers:

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

pipe = pipeline("image-text-to-text", model="webbrain-one/webbrain-vl-2-450M")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("webbrain-one/webbrain-vl-2-450M")
model = AutoModelForMultimodalLM.from_pretrained("webbrain-one/webbrain-vl-2-450M", device_map="auto")
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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use webbrain-one/webbrain-vl-2-450M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "webbrain-one/webbrain-vl-2-450M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "webbrain-one/webbrain-vl-2-450M",
		"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/webbrain-one/webbrain-vl-2-450M

SGLang

How to use webbrain-one/webbrain-vl-2-450M 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 "webbrain-one/webbrain-vl-2-450M" \
    --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": "webbrain-one/webbrain-vl-2-450M",
		"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 "webbrain-one/webbrain-vl-2-450M" \
        --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": "webbrain-one/webbrain-vl-2-450M",
		"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 webbrain-one/webbrain-vl-2-450M with Docker Model Runner:
```
docker model run hf.co/webbrain-one/webbrain-vl-2-450M
```

webbrain-one commited on Aug 22

Commit

e6e5d73

verified ·

1 Parent(s): ce71fe7

Move benchmark charts near the top of the card

Browse files

Files changed (2) hide show

README.md +31 -31
SHA256SUMS +1 -1

README.md CHANGED Viewed

@@ -21,6 +21,37 @@ It turns a browser viewport screenshot into the six-section observation format
 consumed by WebBrain's planning agent. This repository preserves both the
 merged Transformers checkpoint and the matching LoRA adapter.
 ## Intended use
 - Browser screenshot description for a downstream automation planner.
@@ -108,37 +139,6 @@ six-section prompt exactly once after checkpoint selection.
 Machine-readable results are in `eval/`. Latency is implementation- and
 hardware-specific, so it is not presented as a portable speed claim.
-## Deployed ONNX/WebGPU comparison
-The separately published
-[`webbrain-one/webbrain-vl-2-450M-onnx`](https://huggingface.co/webbrain-one/webbrain-vl-2-450M-onnx)
-package reached **44/100 strict passes** and a **76.24% mean rubric score** in its
-Q4-decoder/FP16-vision deployment. The WebBrain point in both charts below is
-that deployed ONNX/WebGPU row, not the 36/100 merged-PyTorch row above.
-![Strict-pass comparison across compact browser-vision models](./assets/browser-vision-benchmark-strict-passes.png)
-![Strict passes plotted against advertised or effective parameter class](./assets/browser-vision-performance-vs-size.png)
-| Model and tested runtime | Parameter class | Strict passes | Mean rubric |
-| --- | ---: | ---: | ---: |
-| Qwen 3.5 4B, MLX | 4B | 55/100 | 83.4% |
-| **WebBrain VL 2, deployed ONNX Q4/FP16** | **450M** | **44/100** | 76.2% |
-| Gemma 4 E4B IT, MLX | E4B | 41/100 | 81.3% |
-| Qwen 3.5 2B, MLX | 2B | 39/100 | 79.1% |
-| Qwen 3.5 0.8B | 0.8B | 20/100 | 41.6% |
-| LFM2.5-VL-1.6B, MLX | 1.6B | 12/100 | 44.4% |
-All rows used the same 100 screenshots and production six-section prompt. The
-[benchmark images, rubrics, runner, and committed result files](https://github.com/esokullu/webbrain/tree/main/test/vision)
-are public. Gemma E4B is plotted at its effective 4B class. Latency is omitted
-because the rows used different machines, quantizations, and inference stacks.
-Further reading:
-- [Tiny Vision Models Compared II](https://webbrain.one/blog/tiny-vision-models-compared-ii/)
-- [From 16K to 50K Browser Screenshots](https://webbrain.one/blog/fine-tuning-lfm25-vl-450m-for-browser-tasks/)
 ## Limitations
 - Small, low-contrast, occluded, or non-Latin text can still be missed.

 consumed by WebBrain's planning agent. This repository preserves both the
 merged Transformers checkpoint and the matching LoRA adapter.
+## Deployed ONNX/WebGPU benchmark
+The separately published
+[`webbrain-one/webbrain-vl-2-450M-onnx`](https://huggingface.co/webbrain-one/webbrain-vl-2-450M-onnx)
+package reached **44/100 strict passes** and a **76.24% mean rubric score** in its
+Q4-decoder/FP16-vision deployment. The WebBrain point in both charts below is
+that deployed ONNX/WebGPU row, not the 36/100 merged-PyTorch row reported later.
+![Strict-pass comparison across compact browser-vision models](./assets/browser-vision-benchmark-strict-passes.png)
+![Strict passes plotted against advertised or effective parameter class](./assets/browser-vision-performance-vs-size.png)
+| Model and tested runtime | Parameter class | Strict passes | Mean rubric |
+| --- | ---: | ---: | ---: |
+| Qwen 3.5 4B, MLX | 4B | 55/100 | 83.4% |
+| **WebBrain VL 2, deployed ONNX Q4/FP16** | **450M** | **44/100** | 76.2% |
+| Gemma 4 E4B IT, MLX | E4B | 41/100 | 81.3% |
+| Qwen 3.5 2B, MLX | 2B | 39/100 | 79.1% |
+| Qwen 3.5 0.8B | 0.8B | 20/100 | 41.6% |
+| LFM2.5-VL-1.6B, MLX | 1.6B | 12/100 | 44.4% |
+All rows used the same 100 screenshots and production six-section prompt. The
+[benchmark images, rubrics, runner, and committed result files](https://github.com/esokullu/webbrain/tree/main/test/vision)
+are public. Gemma E4B is plotted at its effective 4B class. Latency is omitted
+because the rows used different machines, quantizations, and inference stacks.
+Further reading:
+- [Tiny Vision Models Compared II](https://webbrain.one/blog/tiny-vision-models-compared-ii/)
+- [From 16K to 50K Browser Screenshots](https://webbrain.one/blog/fine-tuning-lfm25-vl-450m-for-browser-tasks/)
 ## Intended use
 - Browser screenshot description for a downstream automation planner.
 Machine-readable results are in `eval/`. Latency is implementation- and
 hardware-specific, so it is not presented as a portable speed claim.
 ## Limitations
 - Small, low-contrast, occluded, or non-Latin text can still be missed.

SHA256SUMS CHANGED Viewed

@@ -1,5 +1,5 @@
 4d28ca14dedc0b3d0fcc2b3339f0e79931faa33874f3d24f522183a8fc70068c  LICENSE
-f7c417aaa7c1b91ae43ab920c9a6ac3f2e3d86e545e94c4abefb1ce41805601d  README.md
 2af924baa6fa3c8bbe6242393b7b689594802312e27a8d834f5df978d15eb2a2  assets/browser-vision-benchmark-strict-passes.png
 fb47d96d673c04d1ea747de2dca4e8d0ad81f8708508dc508b79236199c8b4d9  assets/browser-vision-performance-vs-size.png
 117e4eff25a5b8c8bc06044de386033edc69dc1b51f56852bd649286e6f7facd  adapter/README.md

 4d28ca14dedc0b3d0fcc2b3339f0e79931faa33874f3d24f522183a8fc70068c  LICENSE
+94ba0cd8a83eb01ad3d7e7c27f979ae61f17af8001a731a8af2567784d590398  README.md
 2af924baa6fa3c8bbe6242393b7b689594802312e27a8d834f5df978d15eb2a2  assets/browser-vision-benchmark-strict-passes.png
 fb47d96d673c04d1ea747de2dca4e8d0ad81f8708508dc508b79236199c8b4d9  assets/browser-vision-performance-vs-size.png
 117e4eff25a5b8c8bc06044de386033edc69dc1b51f56852bd649286e6f7facd  adapter/README.md