Back to notes PyTorch December 31, 2025 4 words

Classification

Single Label

loss_fn = nn.CrossEntropyLoss()

# Forward pass
logits = model(inputs)

# Backward pass
optimizer.zero_grad()
loss = loss_fn(logits, targets)
loss.backward()
optimizer.step()

Multi Label

loss_fn = nn.BCEWithLogitsLoss() # Binary Cross Entropy

# Forward pass
logits = model(inputs)

# Backward pass
optimizer.zero_grad()
loss = loss_fn(logits, targets)
loss.backward()
optimizer.step()