Exploring loss function topology with cyclical learning rates

February 14, 2017 · Entered Twilight · 🏛 arXiv.org

"Last commit was 8.0 years ago (≥5 year threshold)"

Evidence collected by the PWNC Scanner

Repo contents: Instructions for adding CLR to Caffe.pdf, LICENSE, LRrange-solver.prototxt, README.md, architectures, clrsolver.prototxt, interpolation, solver.prototxt, train.sh

Authors Leslie N. Smith, Nicholay Topin arXiv ID 1702.04283 Category cs.LG: Machine Learning Cross-listed cs.NE Citations 27 Venue arXiv.org Repository https://github.com/lnsmith54/exploring-loss ⭐ 22 Last Checked 2 months ago

Abstract

We present observations and discussion of previously unreported phenomena discovered while training residual networks. The goal of this work is to better understand the nature of neural networks through the examination of these new empirical results. These behaviors were identified through the application of Cyclical Learning Rates (CLR) and linear network interpolation. Among these behaviors are counterintuitive increases and decreases in training loss and instances of rapid training. For example, we demonstrate how CLR can produce greater testing accuracy than traditional training despite using large learning rates. Files to replicate these results are available at https://github.com/lnsmith54/exploring-loss