opendilab/PRG

[ICCV 2025] Pretrained Reversible Generation as Unsupervised Visual Representation Learning

32
/ 100
Emerging

This project helps machine learning researchers extract meaningful features from images without needing vast amounts of labeled data. It takes raw image datasets and transforms them into versatile, multi-level visual representations. This is ideal for deep learning scientists or computer vision engineers working on advanced image understanding tasks.

Use this if you need to perform unsupervised representation learning on visual data to improve the performance of downstream tasks like classification or detection with less labeled data.

Not ideal if you are looking for an out-of-the-box solution for specific computer vision applications without expertise in deep learning model training.

unsupervised-learning computer-vision image-analysis representation-learning deep-learning-research
No Package No Dependents
Maintenance 6 / 25
Adoption 7 / 25
Maturity 15 / 25
Community 4 / 25

How are scores calculated?

Stars

28

Forks

1

Language

Python

License

Apache-2.0

Last pushed

Nov 05, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/opendilab/PRG"

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