Open ZeroOne96 opened 4 years ago
I re-implement this model by torch. But when i train this model using Adam, loss become nan soon although i try several learning rates. So I want to ask some loss info in the training process for further re-implement.
Really thanks.
I re-implement this model by torch. But when i train this model using Adam, loss become nan soon although i try several learning rates. So I want to ask some loss info in the training process for further re-implement.
Really thanks.