KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #dsh #dsh-plugin #ai-agent #codex #ai-agents #deepseek-harness #claude-code #cli #react #mcp

SnapshotEnsemble

★ 190 Open GitHub ↗

Snapshot Ensembles in Torch (Snapshot Ensembles: Train 1, Get M for Free)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Snapshot Ensembles: Train 1, Get M for Free
This repository contains the Torch code for the paper [Snapshot Ensembles: Train 1, Get M for Free](https://arxiv.org/abs/1704.00109).

The code is based on [fb.resnet.torch](https://github.com/facebook/fb.resnet.torch) by [Facebook](https://github.com/facebook) .

There is also a nice [Keras implementation](https://github.com/titu1994/Snapshot-Ensembles) by [titu1994](https://github.com/titu1994).

## Table of Contents
0. [Introduction](#intro)
0. [Usage](#usage)
0. [Contact](#contact)

## Introduction
Snapshot Ensemble is a method to obtain ensembles of multiple neural network at no additional training cost. This is achieved by letting a single neural network converge into several local minima along its optimization path and save the model parameters. The repeated rapid convergence is realized using multiple learning rate annealing cycles.

Figure 1: Left: Illustration of SGD optimization with a typical learning rate schedule. The model converges
to a minimum at the end of training. Right: Illustration of Snapshot Ensembling optimization. The model
undergoes several learning rate annealing cycles, converging to and escaping from multiple local minima. We
take a snapshot at each minimum for test time ensembling.

## Usage 
0. Install Torch ResNet (https://github.com/facebook/fb.resnet.torch);
1. Clone the files to the ```fb.resnet.torch/``` directory. Note that you need to replace ```train.lua``` with the one from this repository;
2. An example command to train a Snapshot Ensemble with ResNet-110 (B = 200 epochs, M = 5 cycles, Initial learning rate alpha = 0.2) on CIFAR-100:

```th main.lua -netType resnet -depth 110 -dataset cifar100 -batchSize 64 -nEpochs 200 -lrShape cosine -nCycles 5 -LR 0.2 -save checkpoints/```

## Contact
[gh349, yl2363]  at cornell.edu   
Any discussions, suggestions and questions are welcome!

SnapshotEnsemble

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README