Instructions to use graph-based-captions/GBC10M-PromptGen-200M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use graph-based-captions/GBC10M-PromptGen-200M with Transformers:

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

pipe = pipeline("text-generation", model="graph-based-captions/GBC10M-PromptGen-200M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("graph-based-captions/GBC10M-PromptGen-200M")
model = AutoModelForMultimodalLM.from_pretrained("graph-based-captions/GBC10M-PromptGen-200M")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use graph-based-captions/GBC10M-PromptGen-200M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "graph-based-captions/GBC10M-PromptGen-200M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "graph-based-captions/GBC10M-PromptGen-200M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/graph-based-captions/GBC10M-PromptGen-200M

SGLang

How to use graph-based-captions/GBC10M-PromptGen-200M 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 "graph-based-captions/GBC10M-PromptGen-200M" \
    --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": "graph-based-captions/GBC10M-PromptGen-200M",
		"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 "graph-based-captions/GBC10M-PromptGen-200M" \
        --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": "graph-based-captions/GBC10M-PromptGen-200M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use graph-based-captions/GBC10M-PromptGen-200M with Docker Model Runner:
```
docker model run hf.co/graph-based-captions/GBC10M-PromptGen-200M
```

GBC10M-PromptGen-200M / README.md

yhsieh

Add library_name, pipeline_tag and project page (#1)

379b2fa verified over 1 year ago

preview code

Raw

History Blame Contribute Delete

1.46 kB

metadata

datasets:
  - graph-based-captions/GBC10M
language:
  - en
license: apple-ascl
library_name: transformers
pipeline_tag: text-generation

Graph-based captioning (GBC) is a new image annotation paradigm that combines the strengths of long captions, region captions, and scene graphs

GBC interconnects region captions to create a unified description akin to a long caption, while also providing structural information similar to scene graphs.

Text-to-Image with GBC as Middleware

We propose to use GBC as middleware for text-to-image generation. This repository provides a model for generating GBC annotation from a simple text prompt.

For futher detail on how to use the model please refer to the accompanying code repository.

License

For license please checkout the LICENSE file.

Citation

@article{GBC2024,
  title={Graph-Based Captioning: Enhancing Visual Descriptions by Interconnecting Region Captions},
  author={Yu-Guan Hsieh and Cheng-Yu Hsieh and Shih-Ying Yeh and Louis Béthune and Hadi Pouransari and Pavan Kumar Anasosalu Vasu and Chun-Liang Li and Ranjay Krishna and Oncel Tuzel and Marco Cuturi},
  journal={arXiv preprint arXiv:2407.06723},
  year={2024}
}

Project page