search

michetonu / gradient_reversal_keras_tf

Keras implementation of a gradient reversal layer for the Tensorflow backend
GNU General Public License v3.0
90 stars 24 forks source link

Lambda schedule #7

Open eliottbrion opened 4 years ago

eliottbrion commented 4 years ago

Hello, Thanks for sharing your code! In the paper, they propose to vary the value of lambda during the training. Do you know how to expend your code so that it can allow lambda to vary during training, for example following lambda(p) = 2/(1+exp(-10p))-1 where p is the learning procedure linearly changing from 0 to 1. Thanks

michetonu commented 4 years ago

@eliottbrion that's a good idea, but off the top of my head I don't have a good idea on how to implement it. I'll give it some thought – if you think of anything please feel free to submit a PR!

eliottbrion commented 4 years ago

@michetonu great, thanks. Looking forward to your thoughts.