PeterWang512/GenDataAttribution
Evaluating Data Attribution for Text-to-Image Models: a visual data attribution benchmark for evaluating and learning training image influences.
When you generate images with large AI models, it's often unclear which original training images influenced the output. This project helps you understand that connection by taking a generated image and identifying the most influential source images from the training data. This is useful for researchers or practitioners who need to analyze or ensure the provenance of AI-generated visuals.
No commits in the last 6 months.
Use this if you need to understand which specific training images contributed most to a particular image generated by a text-to-image AI model.
Not ideal if you're looking to simply generate new images or evaluate the aesthetic quality of AI-generated content.
Stars
79
Forks
3
Language
Python
License
—
Category
Last pushed
Jun 25, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/generative-ai/PeterWang512/GenDataAttribution"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
ZiYang-xie/WorldGen
🌍 WorldGen - Generate Any 3D Scene in Seconds
aioz-ai/AIOZ-GDANCE
AIOZ-GDANCE: a large-scale dataset & baseline for music-driven group dance generation. (CVPR 2023)
worldbench/WorldLens
[CVPR 2026] WorldLens: Full-Spectrum Evaluations of Driving World Models in Real World
Kobaayyy/Awesome-CVPR2026-CVPR2025-ICCV2025-CVPR2024-ECCV2024-AIGC
A Collection of Papers and Codes for CVPR2026/CVPR2025/ICCV2025/CVPR2024/ECCV2024 AIGC
nv-tlabs/XCube
[CVPR 2024 Highlight] XCube: Large-Scale 3D Generative Modeling using Sparse Voxel Hierarchies