1. Start a new TAP session with 2 Nodes and 2 Tasks, instead of 1 Node and 1 Task as documented in Lab Setup
2. If you have not done so, download the notebooks as instructed in Lab Setup
3. Navigate to the sciml-course/cnn directory and open cnn_part5.ipynb
4. Do not change the kernel for this lab

If you are running the notebook and encounter an error with mpirun, double check that you started a TAP session with 2 Nodes and 2 Tasks and are using the default kernel.

This notebook will use the same hyperparameters as used in part 1:

• Learning Rate (lr): how much model parameters are updated at each batch/epoch
• Batch Size: number of data points used to estimate gradients at each iteration
• Epochs: Number of times to iterate over our entire dataset in optimization process

These hyperparameters will be used throughout the notebook.