/test-time-training

Replication of Code for paper Test-Time Training with Self-Supervision for Generalization under Distribution Shifts.

Primary LanguageJupyter Notebook

Test-time Training

Replication of Code for paper Test-Time Training with Self-Supervision for Generalization under Distribution Shifts.

This code is meant to do more experiments and play around with various components of the original paper.

Setup and Installation

Please follow instructions here to setup dependencies.

Datasets

This repository currently only has support to setup CIFAR-10 dataset (and variants). More will follow soon. For required CIFAR datasets, please follow the instructions:

Setup a common folder for all datasets, for example,

DATASET_DIR=/scratch/users/piyushb/test-time-training/datasets
mkdir -p $DATASET_DIR

CIFAR-10-C

To download and setup this, modify datasets/download/cifar-10-c.sh and add your own DATASET_DIR. Then, run

bash datasets/download/cifar-10-c.sh

CIFAR-10.1

To download and setup this, modify datasets/download/cifar-10.1.sh and add your own DATASET_DIR. Then, run

bash datasets/download/cifar-10.1.sh