Instructions to use LiquidAI/LFM2-VL-1.6B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LiquidAI/LFM2-VL-1.6B with Transformers:

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

pipe = pipeline("image-text-to-text", model="LiquidAI/LFM2-VL-1.6B")
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("LiquidAI/LFM2-VL-1.6B")
model = AutoModelForImageTextToText.from_pretrained("LiquidAI/LFM2-VL-1.6B")
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 LiquidAI/LFM2-VL-1.6B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LiquidAI/LFM2-VL-1.6B"
# 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/LFM2-VL-1.6B",
		"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/LFM2-VL-1.6B

SGLang

How to use LiquidAI/LFM2-VL-1.6B 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/LFM2-VL-1.6B" \
    --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/LFM2-VL-1.6B",
		"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/LFM2-VL-1.6B" \
        --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/LFM2-VL-1.6B",
		"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/LFM2-VL-1.6B with Docker Model Runner:
```
docker model run hf.co/LiquidAI/LFM2-VL-1.6B
```

Model weights loading issue for 1.6 B model

by Veb-BLK - opened Sep 24, 2025

Discussion

Veb-BLK

Sep 24, 2025

from transformers import AutoModelForImageTextToText, AutoProcessor

model_id = "LiquidAI/LFM2-VL-1.6B"

processor_id = model_id


processor = AutoProcessor.from_pretrained(
    model_id,
    dtype="bfloat16",
    trust_remote_code=True,
    device_map="auto",
)

model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    dtype="bfloat16",
    trust_remote_code=True,
    device_map="auto",
)

Output:

model.safetensors: 100%
 3.17G/3.17G [00:24<00:00, 601MB/s]
Some weights of the model checkpoint at LiquidAI/LFM2-VL-1.6B were not used when initializing Lfm2VlForConditionalGeneration: ['model.vision_tower.vision_model.encoder.layers.25.layer_norm1.bias', 'model.vision_tower.vision_model.encoder.layers.25.layer_norm1.weight', 'model.vision_tower.vision_model.encoder.layers.25.layer_norm2.bias', 'model.vision_tower.vision_model.encoder.layers.25.layer_norm2.weight', 'model.vision_tower.vision_model.encoder.layers.25.mlp.fc1.bias', 'model.vision_tower.vision_model.encoder.layers.25.mlp.fc1.weight', 'model.vision_tower.vision_model.encoder.layers.25.mlp.fc2.bias', 'model.vision_tower.vision_model.encoder.layers.25.mlp.fc2.weight', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.bias', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.weight', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.bias', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.weight', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.bias', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.weight', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.bias', 'model.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.weight']
- This IS expected if you are initializing Lfm2VlForConditionalGeneration from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing Lfm2VlForConditionalGeneration from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).

EdoardoMosca

Liquid AI org Jan 30

what version of transformers do you have installed? I recommend using 4.57.x.

Although not entirely correct, I don't get any issue when running your snippet.
However, here's the correct code to load the model (from the official model card).

from transformers import AutoProcessor, AutoModelForImageTextToText
from transformers.image_utils import load_image

# Load model and processor
model_id = "LiquidAI/LFM2-VL-1.6B"
model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    device_map="auto",
    dtype="bfloat16"
)
processor = AutoProcessor.from_pretrained(model_id)

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