Open SiyuanWei opened 4 years ago
And I cProfile the SE script and it shows the pykaldi asr decode funtion is the bottleneck
Hi, this is not normal. It should not be as slow as this. Can you let me know your configs, like the batch_size, decoding graph, etc?
I use 8 batch size and decoding graph and transition model under a tri4b system of kaldi librispeech examples. Other configurations shouldn't infulence the decoding speed...emm
The normal training speed should be like
:18:19.000Z /container_e09_1568924618046_16257_01_000002: [1,0]
The batchsize here is 4, and I used 1 V100 GPU. It took less than 30 seconds for 10 updates. I feel that your model has been diverged in the early beginning of the training, which means it would take much longer time to decode the utterances since the model is poor. I would suggest to pay attention to the learning rate, the decoding configs, CE and/or L2 regularization. Note that, use a larger momentum is very helpful for SE training from my experiments.
Hi, I have completed the gmm stage and excute the CE training stage of transformer based on a tri4b system,. But when I'm trying a SE training on a Tesla P100 ,it is extremely slow as the following pict shows. I wonder is it normal?![Capture](https://user-images.githubusercontent.com/20530918/71152849-861ec100-2272-11ea-89b1-3de86df47577.PNG)