Description
The goal of the previous section was to exploit PyTorch to build a network. Here you’ll build your framework for denoising images without using autograd or torch.nn modules. We allow
1 from torch import empty , cat , arange
-
from torch . nn . f u n c t i o n a l import fold , unfold
and the Python Standard Library1 only. We could consider additional operations if you demonstrate a convincing use-case (please ask on slack or by email to the TAs). Note that you can achieve most elementary tensor operations
7
