Why checkpointing preserves progress in PyTorch - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to save the model's state dictionary.

PyTorch

torch.save(model[1], 'model.pth')

Drag options to blanks, or click blank then click option'

A.gradients

B.parameters()

C.weights

D.state_dict()

Attempts:

3 left

💡 Hint

Common Mistakes

Saving model.parameters() instead of state_dict()

Saving model.weights which does not exist

Saving the state_dict() stores the model's learned parameters, which is essential for checkpointing.

2fill in blank

medium

Complete the code to load the saved state dictionary into the model.

PyTorch

model.load_state_dict(torch.load('[1]'))

Drag options to blanks, or click blank then click option'

Astate.pth

Bmodel.pth

Cweights.pth

Dcheckpoint.pth

Attempts:

3 left

💡 Hint

Common Mistakes

Using a different filename than the one used to save

Forgetting to load the state dictionary

The file model.pth is where the model's state dictionary was saved, so it must be loaded from there.

3fill in blank

hard

Fix the error in saving the optimizer state for checkpointing.

PyTorch

torch.save(optimizer[1], 'optimizer.pth')

Drag options to blanks, or click blank then click option'

A.gradients

B.parameters()

C.state_dict()

D.weights

Attempts:

3 left

💡 Hint

Common Mistakes

Saving optimizer.parameters() which does not exist

Saving optimizer.weights which is invalid

The optimizer's state_dict() contains all the information needed to resume training exactly.

4fill in blank

hard

Fill both blanks to save both model and optimizer states in one checkpoint dictionary.

PyTorch

checkpoint = {'model': model[1], 'optimizer': optimizer[2], 'epoch': epoch}
torch.save(checkpoint, 'checkpoint.pth')

Drag options to blanks, or click blank then click option'

A.state_dict()

B.parameters()

C.gradients

D.weights

Attempts:

3 left

💡 Hint

Common Mistakes

Using different methods for model and optimizer

Using .parameters() or .weights instead of .state_dict()

Both model and optimizer states are saved using their state_dict() methods to preserve training progress.

5fill in blank

hard

Fill all three blanks to load checkpoint and resume training with model, optimizer, and epoch.

PyTorch

checkpoint = torch.load('checkpoint.pth')
model.load_state_dict(checkpoint['[1]'])
optimizer.load_state_dict(checkpoint['[2]'])
epoch = checkpoint['[3]']

Drag options to blanks, or click blank then click option'

Amodel

Boptimizer

Cepoch

Dstate

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong keys like 'state' instead of 'model'

Forgetting to load epoch to resume training correctly

The checkpoint dictionary stores keys 'model', 'optimizer', and 'epoch' to restore training state.