dlmacedo/robust-deep-learning

A project to train your model from scratch or fine-tune a pretrained model using the losses provided in this library to improve out-of-distribution detection and uncertainty estimation performances. Calibrate your model to produce enhanced uncertainty estimations. Detect out-of-distribution data using the defined score type and threshold.

44
/ 100
Emerging

This tool helps machine learning engineers and researchers build more reliable deep learning models. It takes your existing deep learning model and training data to produce a model that can better identify data points that are different from what it was trained on, and provide more accurate assessments of its own confidence. The output is a robustly trained or fine-tuned model ready for deployment in real-world scenarios.

No commits in the last 6 months. Available on PyPI.

Use this if you need your deep learning models to accurately detect unusual or unexpected data and provide trustworthy confidence scores for their predictions, without needing extra data or complex hyperparameter tuning.

Not ideal if you are working with simple models or datasets where out-of-distribution detection and uncertainty estimation are not critical concerns.

machine-learning-robustness out-of-distribution-detection uncertainty-quantification model-calibration deep-learning-deployment
Stale 6m
Maintenance 0 / 25
Adoption 6 / 25
Maturity 25 / 25
Community 13 / 25

How are scores calculated?

Stars

17

Forks

3

Language

Python

License

Apache-2.0

Last pushed

Nov 26, 2022

Commits (30d)

0

Dependencies

12

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/dlmacedo/robust-deep-learning"

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