PyTorch Tutorial in 2102570 Introduction to Computer Vision, CUEE

  This tutorial will contain example with description for image classification, custom datasets, transfer learning, and object detection task using PyTorch step by step.

Table of contents

• Prepare system environment for deep learning
  • Set up on your own machine
  • Online environment
• Reference and more resource

Prepare system environment for deep learning

Set up on your own machine

  If you want to set up on your machine, the machine must have at least 8 GB of RAM and a dedicated graphics card from NVIDIA (another brand is still okay, but it may be hard to set up).

• linux/ubuntu - I will update the complete command soon.
• Windows

Online environment

  If you don't have a machine, you can still develop your work by using an online environment instead.

• Google Colab - Now Colab has limited access credit.
• Kaggle - Limited GPU access but it will refill every week. Before access GPU you have to verify your account by following this tutorial.
• Microsoft Azure - Hard to use, but it's more adjustable. Azure has 100 free credits for student, activate by this link.
• Deepnote

Reference and more resource

Reference

• Deep Learning with Pytorch, all code in this book available in Github
• Zero to Mastery Learn PyTorch for Deep Learning, video version of this course Freecodecamp or Author's youtube channels

More Resource

• pytorch_notebooks: A collection of PyTorch notebooks for studying and practicing deep learning.
• pytorch-image-models: Collection of pytorch pretrained-weight models for computer vision from HuggingFace
• pytorch-tutorial
• Deep Learning With Pytorch: a 60 minute blitz: An Official's Pytorch tutorial.
• Deep Learning for Coders with Fastai and Pytorch, official's free version on Github or Official's site