Custom Google Colab notebooks for Udacity’s Deep Learning with PyTorch

When I started this amazing course, I was trying to get running the notebooks on my local computer and I ran into so many issues, mainly:

1. Dealing with pip or conda
2. CUDA drivers issues

I spent more than 2 days to get CUDA working and I finally gave up. Then I discovered this amazing Google Colab

• It’s free!
• You can keep the notebook running for 12 hours
• Comes with GPU K80
• Did I tell you it’s free?

However, getting the Udacity’s notebooks working on Colab again required some effort. PyTorch needed to be installed, appropriate version of Pillow needed to be installed, required data to downloaded etc. So I started making these custom notebooks and now I am sharing all these so that you don’t have to spend any extra minute setting up Google Colab. You can just import them, get started and focus only on solving the exercise.

Lesson 4

• Cats and Dogs classification
• MNIST Digit Classification
• Fashion MNIST Classification
• Transfer Learning

Lesson 5

• CIFAR Data Classification

Lesson 6

• Style Transfer

Lesson 7

• Character Level

Lesson 8

• Sentiment Analysis

Lesson 9

• PyTorch v1 (nightly)
• PyTorch v1 (stable)

Final Lab Project

• Image Classification Project

If you run into any issues with these notebooks, feel free to ping me on Slack. My user id is avinashss Let’s brainstorm and exchange ideas!