search

ShipengWang / Adam-NSCL

PyTorch implementation of our Adam-NSCL algorithm from our CVPR2021 (oral) paper "Training Networks in Null Space for Continual Learning"
MIT License
50 stars 2 forks source link

deeper network #2

Closed YoungSun3Pan closed 3 years ago

YoungSun3Pan commented 3 years ago

Hi, how does the method perform with a deeper network(such as resnet101). Is the deep network more sensitive to parameter changes?

ShipengWang commented 3 years ago

We did not perform with resnet101, but it seems not sensitive to parameter changes for deep network.