j-cb/NINCO
Code and data for the paper "In or Out? Fixing ImageNet Out-of-Distribution Detection Evaluation"
This project provides a robust dataset and tools for evaluating how well an AI model can identify images that are truly 'out-of-distribution' (OOD) when the model was trained on ImageNet-1K. It takes in an image classification model and outputs its OOD detection performance scores using carefully curated OOD datasets. Machine learning researchers and practitioners who develop or deploy image classification models will find this useful for rigorous model testing.
No commits in the last 6 months.
Use this if you need to accurately assess how well your image classification model, especially one trained on ImageNet, can recognize images that are completely unrelated to its training data.
Not ideal if you are looking for a dataset or tool to improve your model's classification accuracy on in-distribution data, rather than its ability to detect novel, out-of-scope images.
Stars
27
Forks
—
Language
Jupyter Notebook
License
MIT
Category
Last pushed
Aug 22, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/j-cb/NINCO"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
yzhao062/pyod
A Python Library for Outlier and Anomaly Detection, Integrating Classical and Deep Learning Techniques
unit8co/darts
A python library for user-friendly forecasting and anomaly detection on time series.
elki-project/elki
ELKI Data Mining Toolkit
raphaelvallat/antropy
AntroPy: entropy and complexity of (EEG) time-series in Python
Minqi824/ADBench
Official Implement of "ADBench: Anomaly Detection Benchmark", NeurIPS 2022.