search

hirofumi0810 / neural_sp

End-to-end ASR/LM implementation with PyTorch
Apache License 2.0
589 stars 140 forks source link

librispeech dict issue vocab #348

Closed qkrguswn2401 closed 2 years ago

qkrguswn2401 commented 2 years ago

I an running librispeech example. under dict folder, there are bpe10000.vocab and wpbpe100000.txt.
Vocab file is the vocab file The wpbpe10000.txt contains only , , . When running language model, the dict.txt of the model is the wpbpe10000.txt (3 vocabs), which does not make sense!
Did I make the wpbpe10000.txt wrong?

qkrguswn2401 commented 2 years ago

image

qkrguswn2401 commented 2 years ago

image this is the dict.txt that Language model uses

MM-0712 commented 2 years ago

Yes, the dict seems error. It need to check carefully again.

Best Guodong