abonte/protopdebug
Implementation of Concept-level Debugging of Part-Prototype Networks
This tool helps researchers in fields like medical imaging or biology understand why an AI model makes certain predictions, especially when dealing with image classification. You provide the AI model's training data and the model itself, and it helps you pinpoint specific visual 'concepts' or patterns that the AI is using to decide, for example, if an X-ray shows a disease or if a bird belongs to a certain species. This allows domain experts to identify if the AI is relying on misleading clues or irrelevant features.
No commits in the last 6 months.
Use this if you need to debug or gain confidence in an image classification AI model's decision-making process by understanding the visual features it considers important.
Not ideal if you are not working with image-based AI models, or if you only need to know how well a model performs without needing to understand its internal reasoning.
Stars
12
Forks
—
Language
Python
License
MIT
Category
Last pushed
May 09, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/abonte/protopdebug"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
jacobgil/pytorch-grad-cam
Advanced AI Explainability for computer vision. Support for CNNs, Vision Transformers,...
frgfm/torch-cam
Class activation maps for your PyTorch models (CAM, Grad-CAM, Grad-CAM++, Smooth Grad-CAM++,...
jacobgil/keras-grad-cam
An implementation of Grad-CAM with keras
ramprs/grad-cam
[ICCV 2017] Torch code for Grad-CAM
innat/HybridModel-GradCAM
A Keras implementation of hybrid efficientnet swin transformer model.