Handy Google Colab notebook for Transfer Learning exercise (Lesson 4)

Summary

The website provides a modified Google Colab notebook for participants of Udacity's PyTorch Scholarship Challenge to classify cat and dog images using transfer learning with GPU support.

Abstract

The web content introduces a Google Colab notebook tailored for the Udacity PyTorch Scholarship Challenge's Lesson 4 exercise. This exercise involves classifying images of cats and dogs using a pre-trained model, which necessitates GPU access. The notebook has been modified to allow direct execution on Google Colab, a free platform that provides GPU resources. The steps to use the notebook include downloading it from a provided link, uploading it to Google Colab, and then running it to focus on the exercise without worrying about dataset management. The modifications made to the notebook include installing PyTorch version 0.4, setting the runtime to use a GPU, downloading and unzipping the dataset, and integrating methods from helper.py to eliminate the need for uploading it separately. The author, Avinash, invites users to contact him on Slack for assistance with the notebook.

Opinions

The author believes that Google Colab is a valuable resource for those without access to a GPU.
The modified notebook is seen as a facilitator for the exercise, removing the need to handle dataset downloads and other preparatory steps.
The author is open to providing support, indicating a community-oriented approach and a willingness to help peers.
The use of a specific PyTorch version (0.4) suggests the author's familiarity with the compatibility requirements of the exercise.

Handy Google Colab notebook for Transfer Learning exercise (Lesson 4)

This blog post is for people who are doing Udacity’s PyTorch Scholarship Challenge course: Deep Learning with PyTorch.

In Lesson 4, we get to classify cats and dogs images using a pretrained model. This exercise requires access to a GPU. Since not many of us do not have a GPU, a good and free alternative is Google Colab.

So, I created this notebook with all the modifications required so that you can run this on Colab directly. Here are the brief steps:

Download the modified Notebook from this link

Visit Google Colab

You will be prompted with a modal, select Upload

If modal doesn’t appear and instead if it opens a new Notebook, then from menu File > Upload Notebook

Upload the modified notebook

That’s it!

Now you can focus on completing the exercise instead of worrying about how to download the images etc.

If you are interested in knowing changes I have made:

Intsall PyTorch v.0.4

Set the Runtime to GPU, so that GPU is enabled

Download the dataset and unzip it

Add the methods from helper.py so that you don’t need to upload it

Hope this helps! If you have got any issues with running this notebook, then feel free to message me on Slack for any help. My slack username is avinash

All the best for the exercise!