Closed usptact closed 6 years ago
To learn on custom data (not images), are the shapes hard-coded to 8x8 or I can change them? I essentially have flat vectors.
The models certainly expect images for now unfortunately, shouldn't be too hard to hack a version which takes arbitrary network structures (ideally in pytorch ;))
To learn on custom data (not images), are the shapes hard-coded to 8x8 or I can change them? I essentially have flat vectors.