How long does is take time to train the model with librispeech

[JRMeyer]

>>> Branden_Stark
[January 15, 2018, 7:23am]

I use three GPUs (GTX 1080 , Memory 8G) to train model with librispeech
( I use the script ./bin/run-librivox.sh ) and have for 4 days ,but it
just displayed 'I STARTING Optimization' , why so low? Then , I found my
GPU Utilization is always 0% ，just a few moments is 100%. How to use
the GPU?

[This is an archived TTS discussion thread from discourse.mozilla.org/t/how-long-does-is-take-time-to-train-the-model-with-librispeech]

[JRMeyer]

>>> lissyx
[January 15, 2018, 8:22am]

You might need to tweak the script to change batch size to improve GPU
usage?

[Archived Post]

[JRMeyer]

>>> Branden_Stark
[January 15, 2018, 8:36am]

now, my GPU's memory usage is 7797Mib( total is 8113Mib), if I increase
the batch size , is it OK ? or will over memory?

[Archived Post]

[JRMeyer]

>>> Branden_Stark
[January 15, 2018, 8:38am]

my batch size is the script's default : train_batch_size = 12,
dev_batch_size=12 test_batch_size=12

[Archived Post]

[JRMeyer]

>>> lissyx
[January 15, 2018, 8:40am]

I'm not an oracle, you'll only know if you try. It might also be that
your CPUs is not fast enough to feed the GPUs ?

[Archived Post]

[JRMeyer]

>>> Branden_Stark
[January 15, 2018, 8:44am]

ok,I will try it ,thank you.

[Archived Post]

[JRMeyer]

>>> elpimous_robot
[January 15, 2018, 3:29pm]

Lissyx, you're an oracle...sometimes LOL
![:joy:](

[Archived Post]

[JRMeyer]

>>> Branden_Stark
[January 16, 2018, 9:15am]

Yeah,you are right.Thank you.

[Archived Post]