Yubei

What's the WER if training on full librispeech train set?

5

Has someone trained it on full librispeech train sets(train-clean-100, train-clean-360, train-other-500)? Could you tell the WER if training on them? Thank you!

I found that you directly adapt all of the entire network layers from ADDA instead of the feature extraction layers. I am confused, is this still feature adaptation?

Would you like share your PER on TIMIT and CER on chinese copurs?

3

Hi~ Thank you for sharing! Would you like share your PER on TIMIT and CER on chinese copurs? Thank you!

GPU utilization down to 0% without any error infos

13

Hi, I've been training models for almost two days. Today, the GPU utilization dropped suddenly to 0%, but all GPU memory were still occupied by the experiment. Besides, the experimental...

Question about low GPU utilization in async training in ASR task

1

I have a total of two GPU servers and one CPU server. I ran 4 trainers on each GPU server. I tried to run the ps and controller jobs on...

Why the results of LM fusion worse than without LM fusion?

19

Hi! I finished shallow LM fusion in ASR using your `ComputeLogitsWithLM` in asr/fusion.py. Then, I have tested it many times using different parameters which maybe affect LM fusion results. However,...

Could you share your pretained model? Thank you!

2

Could you share your pretained model? Thank you!

Thank you for sharing your code! Could you share your WER on librispeech?