Could you mind tell me the WER of librispeech datasets? thanks

Alexander-H-Liu / End-to-end-ASR-Pytorch

This is an open source project (formerly named Listen, Attend and Spell - PyTorch Implementation) for end-to-end ASR implemented with Pytorch, the well known deep learning toolkit.

MIT License

1.18k stars 317 forks source link

Could you mind tell me the WER of librispeech datasets? thanks #27

Closed iamxiaoyubei closed 5 years ago

iamxiaoyubei commented 5 years ago

Hi, I have seen that you share some example log files with Tensorboard:log/log_url.txt.

So, could you mind tell me what is error rate/dev and error rate/train? Is it WER(word error rate)? Which sub-dataset is it test on?(dev-clean/dev-other/test-clean/test-other/or both of them?)

If it is not WER, could you mind tell me the WER of librispeech datasets like this:

========= =====
Dev 4.3%
DevOther 13.0%
Test 4.5%
TestOther 13.2%
========= =====

Thank you a lot!!