1jsingh/Divide-Evaluate-and-Refine

Repo for our NeurIPS 2023 paper on: Divide, Evaluate, and Refine: Evaluating and Improving Text-to-Image Alignment with Iterative VQA Feedback

27
/ 100
Experimental

This project helps creators and researchers evaluate and improve AI-generated images. When you input a detailed text prompt and an image, it outputs a score indicating how well the image aligns with each part of your prompt. This feedback is then used to refine the image iteratively. This tool is for anyone working with text-to-image AI who needs to ensure the generated visuals accurately reflect complex descriptions.

No commits in the last 6 months.

Use this if you need to objectively measure and enhance the accuracy of AI-generated images against complex, multi-faceted text descriptions.

Not ideal if you are looking for a simple pass/fail image quality check or if your text prompts are very basic and don't require detailed semantic alignment.

AI-art-generation image-evaluation prompt-engineering computer-vision content-creation
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

27

Forks

1

Language

Jupyter Notebook

License

MIT

Last pushed

Nov 11, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/1jsingh/Divide-Evaluate-and-Refine"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.