Instructions to use TIGER-Lab/Mantis-llava-7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TIGER-Lab/Mantis-llava-7b with Transformers:

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

pipe = pipeline("image-text-to-text", model="TIGER-Lab/Mantis-llava-7b")

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("TIGER-Lab/Mantis-llava-7b")
model = AutoModelForMultimodalLM.from_pretrained("TIGER-Lab/Mantis-llava-7b")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use TIGER-Lab/Mantis-llava-7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TIGER-Lab/Mantis-llava-7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/Mantis-llava-7b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/TIGER-Lab/Mantis-llava-7b

SGLang

How to use TIGER-Lab/Mantis-llava-7b 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/Mantis-llava-7b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/Mantis-llava-7b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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/Mantis-llava-7b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TIGER-Lab/Mantis-llava-7b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use TIGER-Lab/Mantis-llava-7b with Docker Model Runner:
```
docker model run hf.co/TIGER-Lab/Mantis-llava-7b
```

What do you think of "List Items One by One: A New Data Source and Learning Paradigm for Multimodal LLMs"

by Shure-Dev - opened Apr 29, 2024

Discussion

Shure-Dev

Apr 29, 2024

https://arxiv.org/pdf/2404.16375

I want to know why you do not concat multiple images to make one image and solve with only prompt engineering.

wenhu

TIGER-Lab org May 18, 2024

That's the baseline results we compared against across all the benchmarks. Also, concatenating images make co-reference almost impossible. We don't think that's the way to go.

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