Skip to content

Code for the paper "The Implicit Bias of Depth: How Incremental Learning Drives Generalization"

License

Notifications You must be signed in to change notification settings

dsgissin/Incremental-Learning

Repository files navigation

Incremental-Learning

This repository holds the scripts that can be used to recreate the experiments detailed in the paper "The Implicit Bias of Depth: How Incremental Learning Drives Generalization". The scripts are self-contained and should be relatively self explanatory - each one requires arguments such as the learning rate, dimensionality of the data, initialization scale and so on.

Dependencies

About

Code for the paper "The Implicit Bias of Depth: How Incremental Learning Drives Generalization"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages