get_weight_decays doesn't work properly for Bidirectional RNNs
OverLordGoldDragon opened this issue · comments
In particular IndRNNs and likely RNNs, maybe also LSTMs - looking into it, will fix
OverLordGoldDragon / keras-adamw
Keras/TF implementation of AdamW, SGDW, NadamW, Warm Restarts, and Learning Rate multipliers