Skip to content
Discussion options

You must be logged in to vote

Hi, this sounds fine. for 3D data, you want the shape [C, H, W, D]. The data loader will take care of creating the batch. You can use AddChannel to prepend a singleton channel to your images as part of your transforms.

Replies: 1 comment 1 reply

Comment options

You must be logged in to vote
1 reply
@PCaesar18
Comment options

Answer selected by PCaesar18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
2 participants