ResNET50 - Gachon-Graduation-work/Muzzle_Detection_Project GitHub Wiki

ResNET50

resnet50

It is the most basic structure of resnet, a resisual block, and is mainly called a botleneck architecture in the paper. For each resisual function, it is implemented in three layer stacks.

reference : https://jisuhan.tistory.com/71#Motivation

ResNET50 layer

resnet50layer It has a total of 50 layers and specifies that [] in each conv chunk is a residual block. That is, it may be seen that the resnet50 undergoes a process in which different residual block forms are repeatedly learned for each layer

Performance

performance