Instructions to use digitalartdynamics/facadelab-vision-assist with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use digitalartdynamics/facadelab-vision-assist with Transformers:

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

pipe = pipeline("image-text-to-text", model="digitalartdynamics/facadelab-vision-assist")
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("digitalartdynamics/facadelab-vision-assist")
model = AutoModelForMultimodalLM.from_pretrained("digitalartdynamics/facadelab-vision-assist", 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 digitalartdynamics/facadelab-vision-assist with vLLM:

Install from pip and serve model

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

SGLang

How to use digitalartdynamics/facadelab-vision-assist 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 "digitalartdynamics/facadelab-vision-assist" \
    --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": "digitalartdynamics/facadelab-vision-assist",
		"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 "digitalartdynamics/facadelab-vision-assist" \
        --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": "digitalartdynamics/facadelab-vision-assist",
		"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 digitalartdynamics/facadelab-vision-assist with Docker Model Runner:
```
docker model run hf.co/digitalartdynamics/facadelab-vision-assist
```

FacadeLab Vision Assist picture reader (Qwen3.5-4B, 4-bit NF4)

This repository holds the picture reader used by FacadeLab's Vision Assist. It is a 4-bit bitsandbytes NF4 copy of Qwen/Qwen3.5-4B at revision 851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a.

What was changed

This is a modified (quantized) version of the original Qwen3.5-4B weights:

The language model's Linear layers are quantized with bitsandbytes NF4, double quantization, bf16 compute dtype.
The vision tower (model.visual) and lm_head are left unchanged in bf16.

The copy was made by FacadeLab's deterministic builder (fixed recipe; transformers 5.2.0, bitsandbytes 0.49.2, torch 2.9.1, built on CUDA). facadelab_build.json records the source repository and revision, the recipe, the software stack, and the size and sha256 of every other file in this repository.

Files

File	Size (bytes)	sha256
`LICENSE`	11,544	`bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a`
`chat_template.jinja`	7,909	`a92e1dd97cb1cb175c9b70c0828e146bea4371c2643319b661b777e89811972e`
`config.json`	3,487	`6740d8956c5aad665bbce09b00f7e9c17f5e0bdb6d18d3d20312f5fc3a6f2433`
`facadelab_build.json`	3,489	`b183f2a21a51b08e2fdd197298c64df05c05a58f75e82586c9d8b97dc1255a17`
`generation_config.json`	121	`86f2223f5c89e1852550f70224eda5a32c5742ce8e42aa078f664b78536aee9b`
`model.safetensors`	3,782,105,770	`0bbc485afd65fc4386ce3246af5070e66cf00137b1f5e37b10cff7f8819544b8`
`processor_config.json`	1,363	`bcd178c4fd1f8679d38aa195b7f2e2a907aa2f2f4462dfd45dd2d69b6423c0fc`
`tokenizer.json`	19,989,343	`87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4`
`tokenizer_config.json`	1,171	`cab31eeb28e731839c7b97370197ae2913cb3b744890f86407c993b4d2531178`

Licence and attribution

Qwen3.5-4B is made by the Qwen team at Alibaba Cloud and is released under the Apache License 2.0 (Copyright 2026 Alibaba Cloud). This modified copy is distributed under the same licence. LICENSE is the upstream licence file from Qwen/Qwen3.5-4B, unchanged.

Modified by Digital Art Dynamics for FacadeLab.

Downloads last month: 42

Safetensors

Model size

5B params

Tensor type

BF16

Model tree for digitalartdynamics/facadelab-vision-assist

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Quantized

(528)

this model