1. Learn
    2. /
  2. Courses
    3. /
  3. Introduction to Deep Learning with PyTorch
Connected

Exercise

Calculating cross entropy loss

Cross-entropy loss is a widely used method to measure classification loss. In this exercise, you’ll calculate cross-entropy loss in PyTorch using:

  • y: the ground truth label.
  • scores: a vector of predictions before softmax.

Loss functions help neural networks learn by measuring prediction errors. Create a one-hot encoded vector for y, define the cross-entropy loss function, and compute the loss using scores and the encoded label. The result will be a single float representing the sample's loss.

torch, CrossEntropyLoss, and torch.nn.functional as F have already been imported for you.

Instructions 1/3

undefined XP
    1
    2
    3
  • Create the one-hot encoded vector of the ground truth label y, with 4 features (one for each class), and assign it to one_hot_label.