Tom Hale's PyTorch Udacity / Facebook scholarship course notes

These notes were written in Markdown with formulae.

The formulae are converted to images by TeXify.

I believe that I have attributed all materials, except the Udacity PyTorch Scholarship course, from which these notes are flagrantly copied ;)

My notes

• Lesson 2
• Lesson 3
• Lesson 4

Google Colab starter code for Image Classifier Project.ipynb

It took me a while to work out how to get Google Colab setup for this project:

• Install newly released PyTorch 1.0.0
• Automatically download and unzip the dataset
• Check and warn if GPU is not enabled
• Helper functions
• image_show(): Convert and show an image without the default gridlines
• label_to_text(): Convert a tensor label to its label from cat_to_name.json

Here's the setup code: Google_Colab_startup_PyTorch_Challenge.ipynb

Buy me a beer^H^H^H^Hcoconut

I took many hours creating this content which I hope will make your life easier.

If you have found these notes useful and feel moved to, please donate so I can buy a coconut:

Other resources

List of all PyTorch Challenge notes

• Lesson 2
  • Tatiana Gaponova's 12-page PDF notes
  • 73 page Google Doc (author didn't put their name at the top)
• Lesson 4
  • Tatiana Gaponova's 11-page PDF notes