Validation loop in PyTorch - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to set the model to evaluation mode before validation.

PyTorch

model.[1]()

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

Aeval

Btrain

Cfit

Dvalidate

Attempts:

3 left

💡 Hint

Common Mistakes

Using model.train() instead of model.eval() during validation.

Calling a non-existent method like model.validate().

Calling model.eval() sets the model to evaluation mode, which disables dropout and batch normalization updates during validation.

2fill in blank

medium

Complete the code to disable gradient calculation during validation.

PyTorch

with torch.[1]():
    # validation code here

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

Ano_grad

Bgrad

Cenable_grad

Dautograd

Attempts:

3 left

💡 Hint

Common Mistakes

Not disabling gradients during validation, causing unnecessary computation.

Using torch.grad() which does not exist.

Using torch.no_grad() disables gradient calculation, which saves memory and computations during validation.

3fill in blank

hard

Fix the error in the validation loop to correctly accumulate the total loss.

PyTorch

total_loss = 0
for inputs, labels in val_loader:
    outputs = model(inputs)
    loss = criterion(outputs, labels)
    total_loss [1] loss.item()

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

A=

B*=

C-=

D+=

Attempts:

3 left

💡 Hint

Common Mistakes

Using = which overwrites total_loss each iteration.

Using -= or *= which changes the meaning incorrectly.

Use total_loss += loss.item() to add each batch's loss to the total loss.

4fill in blank

hard

Fill both blanks to compute the average validation loss after the loop.

PyTorch

avg_loss = total_loss [1] len([2])

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

A/

B*

Cval_loader

Dtrain_loader

Attempts:

3 left

💡 Hint

Common Mistakes

Using train_loader length instead of val_loader.

Multiplying instead of dividing.

Divide the total loss by the number of batches in val_loader to get the average loss.

5fill in blank

hard

Fill all three blanks to complete the validation loop with accuracy calculation.

PyTorch

correct = 0
total = 0
with torch.no_grad():
    for inputs, labels in [1]:
        outputs = model(inputs)
        _, predicted = torch.max(outputs.data, [2])
        total += labels.size([3])
        correct += (predicted == labels).sum().item()

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

Aval_loader

Bdim=1

C0

Dtrain_loader

Attempts:

3 left

💡 Hint

Common Mistakes

Using train_loader instead of val_loader.

Using wrong dimension for torch.max or labels.size.

Use val_loader to loop over validation data, dim=1 to get max over class scores, and 0 to get batch size dimension.