This tutorial in meant to be run using Google colab.
Note: you can install pytorch+matplotlib+jupyter to run the notebooks locally if you prefer, however this is at your own risks as we will not be able to provide any support for the installation during the tutorial.
The first 30 minutes of the session will be dedicated to building and training a small neural network from scratch in PyTorch.
To start please follow the next steps:
- Open the notebook on Google colab: https://colab.research.google.com/github/gasse/cors_2021_tutorial/blob/main/from_scratch.ipynb
- Select File -> Save a copy to Drive
- Activate GPU execution by choosing Runtime -> Change runtime type -> GPU
For the second half of the session, we will train a state-of-the-art neural network in PyTorch, to tackle the MNIST digit recogition task.
To start please open Google colab, follow the next steps:
- Open the notebook on Google colab: https://colab.research.google.com/github/gasse/cors_2021_tutorial/blob/main/mnist.ipynb
- Select File -> Save a copy to Drive
- Activate GPU execution by choosing Runtime -> Change runtime type -> GPU