Fast Dropout Training

[matheo_r]

Hi,

Does anyone know why fast dropout training (http://proceedings.mlr.press/v28/wang13a.pdf) is not implemented in PyTorch ?

Thanks,
Matheo