Skip to content

Net-Trim: Convex Pruning of Deep Neural Networks with Performance Guarantee

Notifications You must be signed in to change notification settings

DNNToolBox/Net-Trim-v1

Folders and files

NameName
Last commit message
Last commit date

Latest commit

1ecdf62 · Jun 19, 2018

History

13 Commits
Jun 12, 2018
Oct 31, 2017
Oct 31, 2017
Jun 19, 2018
Oct 31, 2017

Repository files navigation

Net-Trim (first version)

This page contains the first version of Net-Trim, which addresses the regularized form of the Net-Trim convex program discussed in the NIPS (2017) paper.

See main.py for an example of network training followed by a pruning. The Additional Material folder contains the NIPS paper, presentation and supplementary files.

New Version with Manual Available

Starting June 20th, 2018, a new version of Net-Trim addressing the original constrained form of the convex program is available. The code, along with a comprehensive manual and the new version of the paper are available at the following links:

About

Net-Trim: Convex Pruning of Deep Neural Networks with Performance Guarantee

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages