This workshop will introduce participants to deep learning using PyTorch. Learn the basic elements of a PyTorch script and see how to train a simple model using the MNIST hand-written digits dataset.
For more PICSciE/RC workshops on machine learning see the Fall 2024 line-up.
Click on a notebook above. After the notebook loads, click on the "Open in Colab" button. Run the cells. If you encounter the message "Warning: This notebook was not authored by Google" then choose "Run anyway".
- Tensors
- Dataset and Dataloader
- Transforms
- Building the Neural Network
- Autograd
- Optimization
- Saving and Loading Models
One can also replace github.com with githubtocolab.com in the URL to open the notebook in Google Colab.
PyTorch Website
PyTorch Tutorials
PyTorch - Learn the Basics
PyTorch Examples on GitHub
PyTorch on the Princeton Research Computing Clusters
Multi-GPU Training with PyTorch on the Princeton Research Computing Clusters
Videos on YouTube by Andrej Karpathy
Intro to Deep Learning (MIT 6.S191)
Short Courses on deeplearning.ai
LLLM inference
Hugging Face for Princeton Research Computing
Multi-GPU Training
If you encounter any difficulties with the material in this guide then please attend a help session or submit a support ticket.
The material in this repository is derived from the PyTorch tutorial. This workshop was created by David Turner, Jonathan Halverson and members of Princeton Research Computing.