giovp / due

Code for "On Feature Collapse and Deep Kernel Learning for Single Forward Pass Uncertainty".

Deterministic Uncertainty Estimation (DUE)

This repo contains the code for On Feature Collapse and Deep Kernel Learning for Single Forward Pass Uncertainty.

Run Example

Make sure the dependencies listed in environment.yml are available and run:

python train_due.py

which will automatically download the dataset (CIFAR10 by default), and start training. There are several command line flags available for changing the hyper-parameters.

A model trained using the defaults is available from here.

A regression example is implemented in toy_regression.ipynb.

Library

If you want to use DUE in your own project, you can opt to install it using pip:

pip install --upgrade git+https://github.com/y0ast/DUE.git

or clone the repo and run python setup.py.

Acknowledgements: thanks to John Ryan for testing.