search

HawkAaron / E2E-ASR

PyTorch Implementations for End-to-End Automatic Speech Recognition
126 stars 27 forks source link

Any results? #2

Closed smolendawid closed 5 years ago

smolendawid commented 6 years ago

Are there any results on any standard dataset?

HawkAaron commented 6 years ago

Since I'm working on RNN Transducer loss implementation, after finish that, I'll create a repo release the TIMIT training & decoding code on TensorFlow, MXNet and PyTorch.

holianh commented 6 years ago

in this repo, you use local/ folder, contain some code, can you provide me full your code? I am glad waiting for your full code here or email: tuananhktmt@gmail.com thank you so much!

HawkAaron commented 6 years ago

@holianh The local/ folder is from kaldi, so you can find it in kaldi timit recipe.

guangsenw commented 5 years ago

is the rnnt loss finished? could you please release some numbers on the TIMIT set also the parameter settings for training the rnnt model?

HawkAaron commented 5 years ago

The results for CTC and RNN-T are released.