Your question is Standard vs Separable Convolutions. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're comparing CNN architectures for an image classification task and need to choose between a conventional convolution block and a lighter alternative.
Explain the difference between standard convolutions and depthwise separable convolutions, and when you would use each.