ReduceLROnPlateau in PyTorch - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a ReduceLROnPlateau scheduler that monitors validation loss.

PyTorch

scheduler = torch.optim.lr_scheduler.[1](optimizer, mode='min')

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

AReduceLROnPlateau

BExponentialLR

CCosineAnnealingLR

DStepLR

Attempts:

3 left

💡 Hint

Common Mistakes

Using StepLR instead of ReduceLROnPlateau

Forgetting to set mode='min' for validation loss

The ReduceLROnPlateau scheduler reduces the learning rate when a metric has stopped improving. Here, it monitors validation loss, so mode='min' is used.

2fill in blank

medium

Complete the code to call the scheduler step function with the validation loss value.

PyTorch

scheduler.[1](val_loss)

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

Aadjust

Bupdate

Creduce

Dstep

Attempts:

3 left

💡 Hint

Common Mistakes

Calling step() without passing the metric value

Using incorrect method names like 'update' or 'reduce'

The step() method of ReduceLROnPlateau is called with the current validation loss to check if learning rate should be reduced.

3fill in blank

hard

Fix the error in the scheduler initialization by filling the blank with the correct patience value.

PyTorch

scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, patience=[1])

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

A5

B'5'

CNone

D-1

Attempts:

3 left

💡 Hint

Common Mistakes

Passing patience as a string instead of integer

Using negative or None values

The patience parameter expects an integer number of epochs to wait before reducing the learning rate. It should be 5, not a string or invalid value.

4fill in blank

hard

Fill both blanks to create a scheduler that reduces learning rate by a factor of 0.1 after 3 epochs without improvement.

PyTorch

scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, factor=[1], patience=[2])

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

A0.1

B3

C0.5

D5

Attempts:

3 left

💡 Hint

Common Mistakes

Using factor greater than 1

Setting patience too high or as a float

The factor controls how much the learning rate is reduced (0.1 means reduce to 10%). The patience is the number of epochs to wait (3 here) before reducing.

5fill in blank

hard

Fill all three blanks to create a training loop that updates the scheduler with validation loss and prints the learning rate.

PyTorch

for epoch in range(num_epochs):
    train()
    val_loss = validate()
    scheduler.[1](val_loss)
    lr = optimizer.param_groups[0]['[2]']
    print(f"Epoch {epoch+1}, Learning Rate: {lr:.6f}")

# The scheduler step method is called with the validation loss, and the learning rate key is '[3]'.

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

Astep

Blr

Dlearning_rate

Attempts:

3 left

💡 Hint

Common Mistakes

Using incorrect method name instead of 'step'

Accessing learning rate with wrong key like 'learning_rate'

The scheduler's step() method is called with the validation loss. The learning rate is accessed via the key 'lr' in the optimizer's parameter groups.