This 30 seconds is amazing. Been spending hours trying to actually see an example of how 1x1 conv reduces dimensions to a fixed number of output channels. Thank you!!!!
Месяц назад
But these filters (colored blocks) should have the same depth as the input, so they should be represented as a 1 x 1 x input_depth Matrices.
This 30 seconds is amazing. Been spending hours trying to actually see an example of how 1x1 conv reduces dimensions to a fixed number of output channels.
Thank you!!!!
But these filters (colored blocks) should have the same depth as the input, so they should be represented as a 1 x 1 x input_depth Matrices.
Awesome 👍
wow.
you ruined it by rotating the axis