search

arthurdouillard / incremental_learning.pytorch

A collection of incremental learning paper implementations including PODNet (ECCV20) and Ghost (CVPR-W21).
MIT License
383 stars 60 forks source link

When to set use_multi_fc over and bias flags? #10

Closed Akella17 closed 5 years ago

Akella17 commented 5 years ago

Can you mention the reason behind the use of multi_fc (E2E) and single_fc (iCaRL and LwF)? Also, when does one set the bias in the classifier (True for E2E and iCaRL while False for LwF)?

arthurdouillard commented 5 years ago

  1. I've tried multi_fc and single_fc. They should have comparable results. I've used multi_fc because I thought it was how authors implemented it (although I can't read matlab...). Use single_fc by default.

  2. Kinda same answer for bias. Use it if it improves your model performance, else don't.