Implementing the training step
In this exercise, you'll implement the training_step() method in a PyTorch Lightning module designed for an image classification task.
Your implementation should unpack a batch of images and labels, compute the model predictions via the forward pass, calculate the cross entropy loss, and log the training loss.
This exercise is part of the course
Scalable AI Models with PyTorch Lightning
Exercise instructions
- Ensure that you compute predictions using the forward pass.
- Calculate the cross entropy loss.
- Log the training loss.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
from torch.nn.functional import cross_entropy
def training_step(self, batch, batch_idx):
x, y = batch
# Ensure that you compute predictions using the forward pass
y_hat = ____
# Calculate the cross entropy loss
loss = ____
# Log the loss
self.____("train_loss", loss)
return loss