"unpack_output" function takes too much time

[bestamigunay]

Hello,

When I deploy a CNN architecture on Pynq-Z2 board, unpack_output function from the driver_base.py takes too much time. How can I make it faster? I think the solution is in the folding_config_file. If I maximize the PE value of the last layer, unpack_output function runs faster, but still takes too much time (~500ms). And only the last layer's PE value affects the speed of that function. Which extra modifications do I need to do in the folding_config_file for getting higher speed? Thank you!