Instructions to use OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B with Transformers:

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

pipe = pipeline("image-text-to-text", model="OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B

SGLang

How to use OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B 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 "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B" \
    --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": "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B",
		"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 "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B" \
        --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": "OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B with Docker Model Runner:
```
docker model run hf.co/OpenGVLab/PIIP-LLaVA_CLIP-BL_512-448_13B
```

This repository contains the PIIP-LLaVA_CLIP-BL_512-448_13B model, based on vicuna-13b-v1.5.

Please refer to our paper and GitHub repository for introduction and usage.

Citation

If you find this project useful in your research, please consider citing:

@article{piip,
  title={Parameter-Inverted Image Pyramid Networks},
  author={Zhu, Xizhou and Yang, Xue and Wang, Zhaokai and Li, Hao and Dou, Wenhan and Ge, Junqi and Lu, Lewei and Qiao, Yu and Dai, Jifeng},
  journal={arXiv preprint arXiv:2406.04330},
  year={2024}
}

@article{piip_v2,
  title={Parameter-Inverted Image Pyramid Networks for Visual Perception and Multimodal Understanding},
  author={Wang, Zhaokai and Zhu, Xizhou and Yang, Xue and Luo, Gen and Li, Hao and Tian, Changyao and Dou, Wenhan and Ge, Junqi and Lu, Lewei and Qiao, Yu and Dai, Jifeng},
  journal={arXiv preprint arXiv:2501.07783},
  year={2025}
}