Continual Learning utilities in PyTorch.
โ ๏ธ This repository is not maintained anymore. Check out Avalanche: an End-to-End Continual Learning library for a real Pytorch-based framework for CL ๐
Feel free to take a look around if you want to check CL strategies implementations or other stuff.
But beware ๐ this repo is highly tailored to my research activity and workflow. Hence, it is not meant for general use as CL framework. If you find it useful in some way, then all the better!
Reach out at any time for any discussion / clarifications ๐
audio
: preprocessing and managing audio signals
datasets
: loading datasets adapted for CL
experiments
: managing experiments and training
extras
: additional stuff
metrics
: popular performance metrics
models
: implementation of models (MLP, RNNs...)
monitors
: monitoring main metrics and experiment logs
strategies
: CL strategies
video
: preprocessing and managing video
pip install [-e] git:github.com/AndreaCossu/clutils
Use -e
to avoid reinstall in case of code improvements.
Otherwise, clone the repository and add the folder /path/to/clutils/repo
to PYTHONPATH
.