self-supervised-playground

based on https://github.com/HealthML/self-supervised-3d-tasks.git

Usage instructions

In this codebase we provide configurations for training/evaluation of our models.

Preparing data

Our implementations of the algorithms require the data to be 128x128x128x4 dimensions.

Clone the repository and install dependencies

git clone https://github.com/Zeev1988/self-supervised-playground.git
cd self-supervised-playground
conda env create -f env_all_platforms.yml
conda activate conda-env
pip install -e .

Running the experiments

To train any of the self-supervised tasks with a specific task, run:
train.py configs/train/base_3d_brats.json

In base_3d_brats.json choose which task you want to train by updating the 'task' field with the relevant task name.
to run the combination of the tasks set it to 'all'.

To run the downstream task and initialize the weights from a pretrained checkpoint( for now only support cpc), run:
finetune.py configs/finetune/cpc_3d_brats.json