Lab - PyTorch Hands-on Lab

[carmengg]

PyTorch Hands-on Lab

Goal

To provide participants with practical experience in using PyTorch, allowing them to implement and experiment with the concepts introduced in the previous session.

Breakdown

1. Tensor Operations Exercise
   • Task: Create tensors, perform basic operations, and move tensors between CPU and GPU
2. Data Loading and Preprocessing Exercise
   • Task: Load a small subset of RTS data using Dataset and DataLoader
3. Model Building Exercise
   • Task: Construct a simple neural network using nn.Module
4. Optimization Exercise
   • Task: Define a loss function, an optimizer, and a learning rate scheduler
5. Mini Training Loop Exercise
   • Task: Implement a basic training loop to train the model on the small subset of RTS data
6. Discussion and Troubleshooting
   • Discuss potential improvements or extensions to the exercises