ssrllib is a pytorch-based package that implements self supervised learning techniques for image classification.
The training can be run with virtually no code. The library has been built on top of PytorchLightning, and makes use of LightningCLI to remove the typical boilerplate that comes with Deep Learning projects and experiments.
The file run_examples.sh
contains an example of commands used to launch training. While the configs
folder contains a couple of examples of configuration files.