Discussions

You can download the discussions here. We will try to upload discussion material prior to their corresponding sessions. Future discussions listed below likely have broken links.

• Discussion 1 -- PyTorch Intro
  tl;dr: We will learn to use PyTorch, the framework we will be using for implementing deep neural networks.
  
  

  Notebook:

  • discussion_1.ipynb
  • disussion 1 colab notebook

  Suggested Readings:

  • Introduction to PyTorch Tensors
  • What’s a Tensor? - Dan Fleisch
• Discussion 2 -- PyTorchAutoGrad
  tl;dr: We will use PyTorch's automatic differentiation to fit some simple models.
  
  

  Notebook:

  • discussion_2.ipynb
  • discussion 2 colab notebook

  Suggested Readings:

  • A Gentle Introduction to torch.autograd.
• Discussion 3 -- SCC Practice
  tl;dr: We will practice running notebooks on the shared compute cluster.
  
  

  Notebook: discussion_3.ipynb

  Suggested Readings:

  • SCC OnDemand Web Portal
• Discussion 4 -- PyTorch Dataset Classes
  tl;dr: We will build deep neural networks and compare training them with good and bad initialization strategies.
  
  

  Notebook:

  • discussion_4.ipynb
  • discussion 4 colab notebook

  Suggested Readings:

  • Building Models with PyTorch
• Discussion 5 -- Neural Net Initialization
  tl;dr: We will build deep neural networks and compare training them with good and bad initialization strategies.
  
  

  Notebook:

  • discussion_05.ipynb
  • discussion 5 colab notebook

  Suggested Readings:

  • Building Models with PyTorch
• Discussion 6 -- Regularization to Improve Test Accuracy
  tl;dr: We'll consider early stopping and implicit and explicit regularization strategies.
  
  

  Notebook:

  • discussion_06.ipynb
  • discussion 06 colab notebook

  Suggested Readings:

  • TBD
• Discussion 7 -- Regularization
  tl;dr: We will test regularization strategies and measure their impact on our models.
  
  

  Notebook:

  • discussion_07.ipynb
  • discussion 07 colab notebook

  Suggested Readings:

  • TBD