TAO pruning stage for CNN (YOLO)

Hey all,
I explored the different steps at the TAO sdk, and i could not find explation how actually the prune stage in tao is done ( only description of the API call ). i understand that this step removes the less contribute neurons relatively to the score of the target function. But in cnn we have feature maps and filters, so how the prune stage is affects them?
many thanks!

Refer to What is -eq in tlt yolo_v4 prune - #5 by Morganh

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.