Pytorch-1.0-Implementation-of-Google-Street-View-House-Numbers-SVHN-classifier

This is an implementation of Google Street View House Numbers (SVHN) classifier using Pytorch 1.0 and Python 3. Some snippets of code were borrowed from the following repositories:

• potterhsu
• Bjarten
• imadmali

Step 1

Go through /data/preprocessing.ipynb to download and convert the digitStruct.mat file in each dataset to a CSV file

Step 2

Go through /src/training.ipynb to train the SVHN classifier

Step 2

Go through /src/inference.ipynb to make inference with the saved model checkpoint

Prediction Results