BIT-DA/DUSA

[NeurIPS 2024] Exploring Structured Semantic Priors Underlying Diffusion Score for Test-time Adaptation

30
/ 100
Emerging

This project helps machine learning engineers adapt their image classification or segmentation models to new, unseen data conditions during deployment without needing to retrain the entire model. It takes an existing, pre-trained image model and data from a new environment, then uses a diffusion-based method to improve the model's accuracy on the fly. This is for machine learning practitioners and researchers deploying models in dynamic real-world settings.

No commits in the last 6 months.

Use this if your deployed image classification or segmentation models encounter performance drops on new, real-world data due to changing conditions, and you need to adapt them efficiently at the time of inference.

Not ideal if you are looking for a tool to train entirely new image models from scratch or if you do not have access to diffusion models in your pipeline.

machine-learning-deployment image-segmentation image-classification model-adaptation computer-vision
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 8 / 25

How are scores calculated?

Stars

22

Forks

2

Language

Python

License

Apache-2.0

Last pushed

Mar 15, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/BIT-DA/DUSA"

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