FasterAI: A Lightweight Library for Creating Sparse Neural Networks

TL;DR

FasterAI is a lightweight, user-friendly PyTorch library that simplifies the application of neural network compression techniques like sparsification, enabling rapid experimentation with state-of-the-art methods such as Lottery Ticket Hypothesis.

Contribution

It introduces a high-level, flexible API for neural network compression techniques, especially sparsification, integrated seamlessly with existing training workflows.

Findings

Enables sparsification with a single line of code

Supports state-of-the-art techniques like Lottery Ticket Hypothesis

Offers a highly customizable and granular experimentation framework

Abstract

FasterAI is a PyTorch-based library, aiming to facilitate the utilization of deep neural networks compression techniques such as sparsification, pruning, knowledge distillation, or regularization. The library is built with the purpose of enabling quick implementation and experimentation. More particularly, compression techniques are leveraging Callback systems of libraries such as fastai and Pytorch Lightning to bring a user-friendly and high-level API. The main asset of FasterAI is its lightweight, yet powerful, simplicity of use. Indeed, because it was developed in a very granular way, users can create thousands of unique experiments by using different combinations of parameters. In this paper, we focus on the sparsifying capabilities of FasterAI, which represents the core of the library. Performing sparsification of a neural network in FasterAI only requires a single additional line of code in the traditional training loop, yet allows to perform state-of-the-art techniques such as Lottery Ticket Hypothesis experiments