Instructions to use TIGER-Lab/VLM2Vec-LoRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TIGER-Lab/VLM2Vec-LoRA with Transformers:

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

pipe = pipeline("text-generation", model="TIGER-Lab/VLM2Vec-LoRA", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("TIGER-Lab/VLM2Vec-LoRA", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use TIGER-Lab/VLM2Vec-LoRA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TIGER-Lab/VLM2Vec-LoRA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/VLM2Vec-LoRA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TIGER-Lab/VLM2Vec-LoRA

SGLang

How to use TIGER-Lab/VLM2Vec-LoRA 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 "TIGER-Lab/VLM2Vec-LoRA" \
    --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": "TIGER-Lab/VLM2Vec-LoRA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "TIGER-Lab/VLM2Vec-LoRA" \
        --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": "TIGER-Lab/VLM2Vec-LoRA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use TIGER-Lab/VLM2Vec-LoRA with Docker Model Runner:
```
docker model run hf.co/TIGER-Lab/VLM2Vec-LoRA
```

ziyjiang commited on Jul 13, 2025

Commit

795dcd0

verified ·

1 Parent(s): 3e21148

Update README.md

Browse files

Files changed (1) hide show

README.md +1 -2

README.md CHANGED Viewed

@@ -17,7 +17,6 @@ tags:
 This repo contains the model checkpoint for [VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks](https://arxiv.org/abs/2410.05160). In this paper, we aimed at building a unified multimodal embedding model for any tasks. Our model is based on converting an existing well-trained VLM (Phi-3.5-V) into an embedding model. The basic idea is to add an [EOS] token in the end of the sequence, which will be used as the representation of the multimodal inputs.
-<img width="1432" alt="abs" src="https://raw.githubusercontent.com/TIGER-AI-Lab/VLM2Vec/refs/heads/main/figures//train_vlm.png">
 ## Release
 Our model is being trained on MMEB-train and evaluated on MMEB-eval with contrastive learning. We only use in-batch negatives for training. Our best results were based on Lora training with batch size of 1024. We also have checkpoint with full training with batch size of 2048. Our results on 36 evaluation datasets are:
@@ -34,7 +33,7 @@ Our model is being trained on MMEB-train and evaluated on MMEB-eval with contras
 ### Experimental Results
 Our model can outperform the existing baselines by a huge margin.
-<img width="900" alt="abs" src="https://raw.githubusercontent.com/TIGER-AI-Lab/VLM2Vec/refs/heads/main/figures//vlm2vec_results.png">
 ## How to use VLM2Vec

 This repo contains the model checkpoint for [VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks](https://arxiv.org/abs/2410.05160). In this paper, we aimed at building a unified multimodal embedding model for any tasks. Our model is based on converting an existing well-trained VLM (Phi-3.5-V) into an embedding model. The basic idea is to add an [EOS] token in the end of the sequence, which will be used as the representation of the multimodal inputs.
 ## Release
 Our model is being trained on MMEB-train and evaluated on MMEB-eval with contrastive learning. We only use in-batch negatives for training. Our best results were based on Lora training with batch size of 1024. We also have checkpoint with full training with batch size of 2048. Our results on 36 evaluation datasets are:
 ### Experimental Results
 Our model can outperform the existing baselines by a huge margin.
+<img width="900" alt="abs" src="vlm2vec_v1_result.png">
 ## How to use VLM2Vec