continual-learning-utils 0.1.0

A Python package for continual learning algorithms and utilities.

Continual Learning

This repository contains code for the demonstration of various concepts in continual learning.

Installation

To install the required dependencies, you can use uv:

uv sync

Usage

To run the demonstrations locally, there are two options:

Running with uv

uv run <demo_name>/demo.py

Replace <demo_name> with the folder name of the specific demonstration you want to run.

In Jupyter Notebook

You can also run the demonstrations in a Jupyter Notebook using locally running Jupyter server.

1. Install jupyter kernel for uv. This will allow the locally running Jupyter server to recognize the uv environment and use it as a kernel for running the notebooks:

   uv sync --dev
   uv run python -m ipykernel install --user --name=continual-learning --display-name "Continual Learning"
   

2. Start the JupyterLab server:

   uv run jupyter lab
   

3. Open the desired notebook from the JupyterLab interface and select the "Continual Learning" kernel to run the notebook.

Demos

The repository includes the following demonstrations:

• Catastrophic Forgetting: This demo illustrates the phenomenon of catastrophic forgetting in neural networks when trained sequentially on multiple tasks.

Contributing

If you would like to contribute to this repository, please feel free to submit a pull request. We welcome contributions that improve the code, add new demonstrations, or enhance the documentation.