Instructions to use ig1/Qwen2.5-VL-7B-Instruct-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ig1/Qwen2.5-VL-7B-Instruct-NVFP4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="ig1/Qwen2.5-VL-7B-Instruct-NVFP4")
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("ig1/Qwen2.5-VL-7B-Instruct-NVFP4")
model = AutoModelForImageTextToText.from_pretrained("ig1/Qwen2.5-VL-7B-Instruct-NVFP4")
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 ig1/Qwen2.5-VL-7B-Instruct-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ig1/Qwen2.5-VL-7B-Instruct-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ig1/Qwen2.5-VL-7B-Instruct-NVFP4",
		"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/ig1/Qwen2.5-VL-7B-Instruct-NVFP4

SGLang

How to use ig1/Qwen2.5-VL-7B-Instruct-NVFP4 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 "ig1/Qwen2.5-VL-7B-Instruct-NVFP4" \
    --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": "ig1/Qwen2.5-VL-7B-Instruct-NVFP4",
		"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 "ig1/Qwen2.5-VL-7B-Instruct-NVFP4" \
        --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": "ig1/Qwen2.5-VL-7B-Instruct-NVFP4",
		"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 ig1/Qwen2.5-VL-7B-Instruct-NVFP4 with Docker Model Runner:
```
docker model run hf.co/ig1/Qwen2.5-VL-7B-Instruct-NVFP4
```

Quantization script:

# Create a dedicated python env
python3 -m venv llmcompressor
source llmcompressor/bin/activate

# Install llm-compressor and additionnal needed libs
pip install llmcompressor qwen_vl_utils torchvision

# Download model in HF cache
hf download Qwen/Qwen2.5-VL-7B-Instruct

# Prepare quantization script
## Download the GPTQ (INT4) as it is the closest to what we need to acheive (it includes the calibration phase)
wget https://github.com/vllm-project/llm-compressor/raw/refs/tags/0.8.1/examples/multimodal_vision/qwen_2_5_vl_example.py -O qwen_2_5_vl_gptq.py
## Create the patch file for NVFP4
cat << EOF > nvfp4.patch
--- qwen_2_5_vl_gptq.py 2025-10-20 13:34:15.446886854 +0200
+++ qwen_2_5_vl_fp4.py  2025-10-19 17:44:04.932080648 +0200
@@ -7,7 +7,7 @@
 from transformers import AutoProcessor, Qwen2_5_VLForConditionalGeneration

 from llmcompressor import oneshot
-from llmcompressor.modifiers.quantization import GPTQModifier
+from llmcompressor.modifiers.quantization import QuantizationModifier
 from llmcompressor.utils import dispatch_for_generation

 # Load model.
@@ -69,13 +69,11 @@


 # Recipe
-recipe = [
-    GPTQModifier(
-        targets="Linear",
-        scheme="W4A16",
-        ignore=["lm_head", "re:visual.*", "re:model.visual.*"],
-    ),
-]
+recipe = QuantizationModifier(
+    targets="Linear",
+    scheme="NVFP4",
+    ignore=["lm_head", "re:visual.*", "re:model.visual.*"],
+)

 # Perform oneshot
 oneshot(
@@ -122,6 +120,6 @@


 # Save to disk compressed.
-SAVE_DIR = model_id.rstrip("/").split("/")[-1] + "-W4A16-G128"
+SAVE_DIR = model_id.rstrip("/").split("/")[-1] + "-NVFP4"
 model.save_pretrained(SAVE_DIR, save_compressed=True)
 processor.save_pretrained(SAVE_DIR)
EOF
## Apply the patch
patch qwen_2_5_vl_gptq.py -i nvfp4.patch -o qwen_2_5_vl_nvfp4.py

# Start the quantization
python3 qwen_2_5_vl_nvfp4.py