adaptive-softmax
adaptive-softmax copied to clipboard
How to put 1 billion word dataset into train.txt, valid.txt and test.txt
In order to reproduce the result, could you please describe how to make the 1 billion word dataset into train.txt, valid.txt and test.txt? Or an example showing first a few lines of the train/valid/test files?