SHI-Labs/Diffusion-Driven-Test-Time-Adaptation-via-Synthetic-Domain-Alignment

Everything to the Synthetic: Diffusion-driven Test-time Adaptation via Synthetic-Domain Alignment, arXiv 2024 / CVPR 2025

21
/ 100
Experimental

This project helps machine learning practitioners improve the accuracy of existing image classification models when deployed in new, unpredictable environments. It takes an existing image classification model and new, corrupted or unusual image data, then processes them to produce more accurate predictions for the new data. Researchers and engineers working with computer vision models would use this to ensure reliable performance in real-world scenarios.

No commits in the last 6 months.

Use this if your pre-trained image classification models perform poorly on new, diverse, or 'corrupted' real-world image data, and you need to boost their accuracy without re-training from scratch.

Not ideal if you are developing new deep learning models from the ground up, or if your image data is consistently clean and perfectly matches your training data.

computer-vision image-classification model-adaptation robust-ai machine-learning-deployment
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 8 / 25
Community 6 / 25

How are scores calculated?

Stars

40

Forks

2

Language

Python

License

Last pushed

Mar 01, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/SHI-Labs/Diffusion-Driven-Test-Time-Adaptation-via-Synthetic-Domain-Alignment"

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