neu-teng

How to faster the training process

1

I run this version transformer to train my nmt system with four GPUS. The data is about 3 million parallel between Chinese and Japanese. But I found the time costing...

Is it possible for decoding the real slot-value?

2

Hi, thanks for sharing your code I had run some experiments on the several data set, and the model replies with the slot_token just like name_SLOT rather than the real...