-
-
Notifications
You must be signed in to change notification settings - Fork 1.8k
Description
There is a super small typo on line 56 the Sec. 3 of the Quickstart, the image
should become images
in the model call, as reported at the bottom:
OLD:
for images, gt_masks in dataloader:
predicted_mask = model(image)
loss = loss_fn(predicted_mask, gt_masks)
loss.backward()
optimizer.step()
proposed NEW
for images, gt_masks in dataloader:
predicted_mask = model(images)
loss = loss_fn(predicted_mask, gt_masks)
loss.backward()
optimizer.step()
Metadata
Metadata
Assignees
Labels
No labels