Skip to content

Latest commit

 

History

History
13 lines (10 loc) · 978 Bytes

README.md

File metadata and controls

13 lines (10 loc) · 978 Bytes

MNIST-GANS

Generative Adversarial Networks on MNIST dataset using pytorch. This code is tested on python 3.7. For training tips and network architecture on GANs I took the help from here

Steps to run the project

  • Download dataset using download_dataset.sh. Rest of the code assumes the dataset is present in data directory

This downloads the dataset from the official site, unzips it and writes all the image files in the format counter-label.jpg With counter varying from 00000 to 60000 and label varying from 1 to 9

  • Run training using src/main.py. You need to give a unique id to your run with option --uid UNIQUE_ID
Generated sample after 200 epochs

Generated Sample

If you find a bug or you have a suggestion regarding the code structure or coding style please create a issue