Flight Lab / Learning machines / 06
Lesson 06 of 06 · about 4 min
A convolution slides a kernel across the input. With no padding the kernel has to sit fully inside, so the output is shorter than the input. The width is (input − kernel + 2 pad) / stride + 1.
The input is 8 wide, the kernel is 3, the stride is 1, and the pad is 0. Enter the output width.