Would you please show some major training parameters of the pre trained model

[JRMeyer]

>>> jackhuang
[January 12, 2018, 3:17am]

I want to know some major parameters of the pre-trained model since my
model doesn't perform well and I want some help.

[This is an archived TTS discussion thread from discourse.mozilla.org/t/would-you-please-show-some-major-training-parameters-of-the-pre-trained-model]

[JRMeyer]

>>> kdavis
[January 13, 2018, 6:33am]

The parameters need to be tuned to the data you are training on.

So a question in return is what data set are you training with?

[Archived Post]

[JRMeyer]

>>> jackhuang
[January 15, 2018, 11:34am]

I use Common voice, Librivox-360 clean and TED as the training set,
would you please give me some advice on choosing the parameters?

[Archived Post]

[JRMeyer]

>>> kdavis
[January 17, 2018, 8:29am]

Here are some of the parameters we used to train the released model

...
--train_batch_size 12 slash
--dev_batch_size 8 slash
--test_batch_size 8 slash
--epoch 13 slash
--learning_rate 0.0001 slash
--display_step 0 slash
--validation_step 1 slash
--dropout_rate 0.2367 slash
--default_stddev 0.046875 slash
--checkpoint_step 1 slash
--log_level 0 slash
...

The batch sizes depend upon the memory of your graphics card. You'll
want to make them as big as possible while not running out of memory.

For your data set I'd suggest using more epochs, maybe something around
20-30.

The only other parameter you may want to play with a bit is the
dropout_rate. You may want to increase it a bit to somewhere around
0.25 to 0.27 but you will have to play around to find the optimal
value.

[Archived Post]

[JRMeyer]

>>> jackhuang
[January 17, 2018, 10:21am]

Thank you very much! slash
And may I ask if the parameter 'n_hidden' is 2048 ?

[Archived Post]

[JRMeyer]

>>> kdavis
[January 17, 2018, 10:26am]

Yes the n_hidden parameter is 2048.

[Archived Post]

[JRMeyer]

>>> jackhuang
[January 17, 2018, 10:32am]

Thank
you!![:grinning:](

[Archived Post]

[JRMeyer]

>>> mark2
[February 2, 2018, 1:20pm]

Is there any guidelines which parameters affect speed, and which model
accuracy? I just want to get rough model quickly to verify the whole
concept, but still want some reasonable outputs...

[Archived Post]

[JRMeyer]

>>> jackhuang
[February 2, 2018, 1:28pm]

n_hidden and train_batch_size are the main parameters that affect
speed(if the n_hidden is smaller and train_batch_size is larger, the
speed will be faster). n_hidden and epoch will affect model accuracy.

[Archived Post]