JinXins/Adversarial-AutoMixup

Official PyTorch(MMCV) implementation of “Adversarial AutoMixup” (ICLR 2024 spotlight)

34
/ 100
Emerging

When training an image classification model, it's crucial to ensure it can accurately categorize images even if they are slightly unusual or 'hard to classify'. This project, AdAutoMix, improves model robustness by automatically generating complex, blended images and using them to challenge and train your image classifier. It takes your existing image datasets and classification models as input and produces a more resilient, better-performing classifier ready for real-world use.

No commits in the last 6 months.

Use this if you are a machine learning engineer or researcher focused on computer vision and need to build highly robust image classification models that perform well on diverse and challenging real-world data.

Not ideal if you are looking for a plug-and-play solution for general machine learning tasks beyond image classification, or if you do not have experience with deep learning frameworks like PyTorch and MMClassification.

image-classification computer-vision model-robustness deep-learning-training adversarial-learning
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 9 / 25

How are scores calculated?

Stars

71

Forks

5

Language

Python

License

Apache-2.0

Last pushed

Nov 02, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/JinXins/Adversarial-AutoMixup"

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