Using AdaptiveLogsoftmaxWithLoss

salesforce / awd-lstm-lm

LSTM and QRNN Language Model Toolkit for PyTorch

BSD 3-Clause "New" or "Revised" License

1.96k stars 488 forks source link

Using AdaptiveLogsoftmaxWithLoss #69

Open akurniawan opened 6 years ago

akurniawan commented 6 years ago

Since now AdaptiveLogsoftmaxWithLoss is released, is it worth to just use their implementation instead of using the SplitCrossEntropyLoss? It seems we can also use the split method by using cutoffs parameters. Or is there any fundamental difference between those two?