nazim1021/OOD-detection-using-OECC

Outlier Exposure with Confidence Control for Out-of-Distribution Detection

30
/ 100
Emerging

This project helps machine learning engineers build more robust image and text classification systems. It takes a pre-trained deep neural network and a dataset of 'in-distribution' samples, then trains the network to reliably identify data points that are significantly different from what it was originally trained on. The output is a more confident and accurate classification model that can flag unusual or unexpected inputs.

No commits in the last 6 months.

Use this if you need your image or text classification models to reliably detect data points that fall outside of their expected input distribution, without needing separate examples of 'outlier' data.

Not ideal if your problem does not involve deep neural networks or if you already have a large, diverse dataset of known outlier samples.

image-classification text-classification anomaly-detection model-robustness machine-learning-engineering
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 8 / 25
Community 13 / 25

How are scores calculated?

Stars

71

Forks

8

Language

Jupyter Notebook

License

Last pushed

Mar 13, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/nazim1021/OOD-detection-using-OECC"

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