Convolution Layers #7
Description
Great work @upskyy Very easy to follow implementation. A couple of questions and I appreciate the clarification:
- Did you verify whether this implementation is trainable?
- I see that most implementations out there use SeparableConv instead of Conv. Have you used SeparableConv to see whether it makes a difference?
- It seems the squeeze excitation block is missing one conv layer right in the beginning of the block