# TD

**Repository Path**: ibsyl/TD

## Basic Information

- **Project Name**: TD
- **Description**: Complementary code for the Targeted Dropout paper
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-08-09
- **Last Updated**: 2021-08-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Targeted Dropout

Aidan N. Gomez, Ivan Zhang, Kevin Swersky, Yarin Gal, and Geoffrey E. Hinton

## Table of Contents
- [Requirements](#requirements)
- [Quick Start](#quick-start)
- [Experiments](#experiments)

## Requirements
- Python 3
- Tensorflow 1.8

## Quick Start
1. Train a model: `python -m TD.train --hparams=resnet_default`
2. Prune that model: `python -m TD.scripts.prune.eval --hparams=resnet_default --prune_percent 0.0,0.25,0.5,0.75,0.95`

### Flags
- `--env`: one of `local`, `gcp` (GPU instances), or `tpu` (TPU instances). Feel free to add more if necessary.
- `--hparams`: the hparam set you want to run.
- `--hparam_override`: manually specify hparams to be overridden (e.g `--hparam_override 'drop_rate=0.66'`)