5,303
edits
PyTorch: Difference between revisions
→TensorBoard
| Line 62: | Line 62: | ||
writer = SummaryWriter(log_dir="./runs") | writer = SummaryWriter(log_dir="./runs") | ||
writer.add_scalar("train_loss", | # Calculate loss. Increment the step. | ||
writer.add_scalar("train_loss", loss.item(), step) | |||
# Optionally flush e.g. at checkpoints | # Optionally flush e.g. at checkpoints | ||