Instructions to use LiquidAI/LFM3-VL-2B-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LiquidAI/LFM3-VL-2B-NVFP4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="LiquidAI/LFM3-VL-2B-NVFP4", trust_remote_code=True)
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 AutoModelForImageTextToText
model = AutoModelForImageTextToText.from_pretrained("LiquidAI/LFM3-VL-2B-NVFP4", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use LiquidAI/LFM3-VL-2B-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LiquidAI/LFM3-VL-2B-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": "LiquidAI/LFM3-VL-2B-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/LiquidAI/LFM3-VL-2B-NVFP4

SGLang

How to use LiquidAI/LFM3-VL-2B-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 "LiquidAI/LFM3-VL-2B-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": "LiquidAI/LFM3-VL-2B-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 "LiquidAI/LFM3-VL-2B-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": "LiquidAI/LFM3-VL-2B-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 LiquidAI/LFM3-VL-2B-NVFP4 with Docker Model Runner:
```
docker model run hf.co/LiquidAI/LFM3-VL-2B-NVFP4
```

LFM3-VL-2B-NVFP4 / hf_quant_config.json

samuelstevens

s-jse

tugot17

Initial release

b6b889b 5 days ago

Raw History Blame Contribute Delete

2.78 kB

{
  "producer": {
    "name": "modelopt",
    "version": "0.0.1.dev1+g4c1f81382"
  },
  "quantization": {
    "quant_algo": "NVFP4",
    "kv_cache_quant_algo": null,
    "group_size": 16,
    "exclude_modules": [
      "lm_head",
      "model.language_model.embed_tokens",
      "model.language_model.layers.0.conv*",
      "model.language_model.layers.1.conv*",
      "model.language_model.layers.10.self_attn*",
      "model.language_model.layers.11.conv*",
      "model.language_model.layers.12.self_attn*",
      "model.language_model.layers.13.conv*",
      "model.language_model.layers.14.self_attn*",
      "model.language_model.layers.15.conv*",
      "model.language_model.layers.16.self_attn*",
      "model.language_model.layers.17.conv*",
      "model.language_model.layers.18.self_attn*",
      "model.language_model.layers.19.conv*",
      "model.language_model.layers.2.self_attn*",
      "model.language_model.layers.20.self_attn*",
      "model.language_model.layers.21.conv*",
      "model.language_model.layers.22.self_attn*",
      "model.language_model.layers.23.conv*",
      "model.language_model.layers.3.conv*",
      "model.language_model.layers.4.self_attn*",
      "model.language_model.layers.5.conv*",
      "model.language_model.layers.6.self_attn*",
      "model.language_model.layers.7.conv*",
      "model.language_model.layers.8.self_attn*",
      "model.language_model.layers.9.conv*",
      "model.multi_modal_projector*",
      "model.vision_tower*",
      "language_model.lm_head",
      "language_model.model.embed_tokens",
      "language_model.model.layers.0.conv*",
      "language_model.model.layers.1.conv*",
      "language_model.model.layers.10.self_attn*",
      "language_model.model.layers.11.conv*",
      "language_model.model.layers.12.self_attn*",
      "language_model.model.layers.13.conv*",
      "language_model.model.layers.14.self_attn*",
      "language_model.model.layers.15.conv*",
      "language_model.model.layers.16.self_attn*",
      "language_model.model.layers.17.conv*",
      "language_model.model.layers.18.self_attn*",
      "language_model.model.layers.19.conv*",
      "language_model.model.layers.2.self_attn*",
      "language_model.model.layers.20.self_attn*",
      "language_model.model.layers.21.conv*",
      "language_model.model.layers.22.self_attn*",
      "language_model.model.layers.23.conv*",
      "language_model.model.layers.3.conv*",
      "language_model.model.layers.4.self_attn*",
      "language_model.model.layers.5.conv*",
      "language_model.model.layers.6.self_attn*",
      "language_model.model.layers.7.conv*",
      "language_model.model.layers.8.self_attn*",
      "language_model.model.layers.9.conv*",
      "vision_projector*",
      "vision_encoder*"
    ]
  }
}

	{
	"producer": {
	"name": "modelopt",
	"version": "0.0.1.dev1+g4c1f81382"
	},
	"quantization": {
	"quant_algo": "NVFP4",
	"kv_cache_quant_algo": null,
	"group_size": 16,
	"exclude_modules": [
	"lm_head",
	"model.language_model.embed_tokens",
	"model.language_model.layers.0.conv*",
	"model.language_model.layers.1.conv*",
	"model.language_model.layers.10.self_attn*",
	"model.language_model.layers.11.conv*",
	"model.language_model.layers.12.self_attn*",
	"model.language_model.layers.13.conv*",
	"model.language_model.layers.14.self_attn*",
	"model.language_model.layers.15.conv*",
	"model.language_model.layers.16.self_attn*",
	"model.language_model.layers.17.conv*",
	"model.language_model.layers.18.self_attn*",
	"model.language_model.layers.19.conv*",
	"model.language_model.layers.2.self_attn*",
	"model.language_model.layers.20.self_attn*",
	"model.language_model.layers.21.conv*",
	"model.language_model.layers.22.self_attn*",
	"model.language_model.layers.23.conv*",
	"model.language_model.layers.3.conv*",
	"model.language_model.layers.4.self_attn*",
	"model.language_model.layers.5.conv*",
	"model.language_model.layers.6.self_attn*",
	"model.language_model.layers.7.conv*",
	"model.language_model.layers.8.self_attn*",
	"model.language_model.layers.9.conv*",
	"model.multi_modal_projector*",
	"model.vision_tower*",
	"language_model.lm_head",
	"language_model.model.embed_tokens",
	"language_model.model.layers.0.conv*",
	"language_model.model.layers.1.conv*",
	"language_model.model.layers.10.self_attn*",
	"language_model.model.layers.11.conv*",
	"language_model.model.layers.12.self_attn*",
	"language_model.model.layers.13.conv*",
	"language_model.model.layers.14.self_attn*",
	"language_model.model.layers.15.conv*",
	"language_model.model.layers.16.self_attn*",
	"language_model.model.layers.17.conv*",
	"language_model.model.layers.18.self_attn*",
	"language_model.model.layers.19.conv*",
	"language_model.model.layers.2.self_attn*",
	"language_model.model.layers.20.self_attn*",
	"language_model.model.layers.21.conv*",
	"language_model.model.layers.22.self_attn*",
	"language_model.model.layers.23.conv*",
	"language_model.model.layers.3.conv*",
	"language_model.model.layers.4.self_attn*",
	"language_model.model.layers.5.conv*",
	"language_model.model.layers.6.self_attn*",
	"language_model.model.layers.7.conv*",
	"language_model.model.layers.8.self_attn*",
	"language_model.model.layers.9.conv*",
	"vision_projector*",
	"vision_encoder*"
	]
	}
	}