Closed NorwegianRockCat closed 3 years ago
The documentation for this function is that if you provide an int, scale will scale the smallest size to the int. Creating a tuple here ruins the checks in functional. So, just store the int.
This seems to match what PyTorch does.
The documentation for this function is that if you provide an int, scale will scale the smallest size to the int. Creating a tuple here ruins the checks in functional. So, just store the int.
This seems to match what PyTorch does.