A 2-hour long lecture by Dr. Alessio Ansuini, data scientist @ AREA Research & Technology, Trieste, Italy.
This repo contains two notebooks covering some notions that were explained during the lecture.
In both of them, you'll learn how to build and train a Neural Network in PyTorch.
The two notebooks are different in the way they present and approach the code within them:
- The
basicnotebook offers a more hands-on approach without giving detailed explanations about what's happening - The
mediumnotebook instead presents everything in detail and is thought of for people willing to understand clearly the meaning of the lines of code they're executing
These notebooks are for an introductory-level course about Neural Networks. The main idea is that attendees, after reading going through the notebooks, are able to grasp the concepts of building a vanilla Fully-connected Neural Network in PyTorch understanding what's happening. We do not cover more advanced stuff, like Convolutional Neural Networks or Dropout, or GPU training. We hope that the attendees who are willing to experiment more by themselves have enough knowledge to explore by themselves the vast world of Deep Learning in PyTorch thanks also to the myriad of tutorials which are freely present on the web.
If you don't not wish to execute the code on your local machine, you can freely make use of Google Colab, a resource which is available to everyone having a Google account.
Once logged in, click on File → Open notebook. In the form which will open up, select GitHub on the upper bar.
Click CANCEL on the next form, then paste the URL of this repository into the search bar to load one of the notebooks into Colab.
Executing a code cell (with Ctrl+Enter or Shift+Enter) might open up a warning telling that the code has been imported from GitHub, ignore it by clicking on Run anyway.
Enjoy!
These notebooks were created by Andrea Gasparin and Marco Zullich with the kind help of Alessio Ansuini and Alisea Stroligo.