Help on creating the model from the common voice mozilla

[JRMeyer]

>>> javi.rahman
[August 17, 2019, 11:50am]

Hi,

I was trying to create a new model from the common voice corpus
(English).

As per the document I have downloaded the corpus en tar file.

Then extracted

bin/import_cv2.py slash --filter_alphabet path/to/some/alphabet.txt
/path/to/extracted/language/archive

I am on this command slash -- slash >

./DeepSpeech.py slash --train_files
CommonVoice-2.0/corpus3-en/clips/train.csv slash --dev_files
CommonVoice-2.0/corpus3-en/clips/dev.csv slash --test_files
CommonVoice-2.0/corpus3-en/clips/test.csv slash --train_batch_size 12
slash --dev_batch_size 12 slash --test_batch_size 12 slash --learning_rate 0.0001
slash --epoch 95 slash --validation_step 5 slash --dropout_rate 0.30 slash --default_stddev
0.046875 slash --export_dir /opt/deepspeech/exportmodels slash --checkpoint
/opt/deepspeech/checkpoint

But it is taking too long.

Please confirm once the command finishes it will bring like what inside
the slash -- slash > deepspeech-0.5.1-models

I assume it will create a new models file ( lm.binary, output_graph.pb,
output_graph.pbmm, output_graph.tflite).

[This is an archived TTS discussion thread from discourse.mozilla.org/t/help-on-creating-the-model-from-the-common-voice-mozilla]

[JRMeyer]

>>> lissyx
[August 18, 2019, 11:11am]



> But it is taking too long.

Please share details, we do document that training is an intensive
process, so without knowing your hardware as well as what you consider
'too long', we can't do anything.



> I assume it will create a new models file ( lm.binary,
> output_graph.pb, output_graph.pbmm, output_graph.tflite).

Don't assume, read the documentation and the help.

[Archived Post]

[JRMeyer]

>>> javi.rahman
[August 18, 2019, 12:47pm]

,

I am using Core i7 7th generation, 8gb ram, 4gb Nvidia graphics.

I will check the document and let you know if I have any doubts.

[Archived Post]

[JRMeyer]

>>> lissyx
[August 18, 2019, 1:26pm]



> I am using Core i7 7th generation, 8gb ram, 4gb Nvidia graphics.

Then no surprise, 250h of french audio takes slash ~4h for training on 2x
RTX2080Ti GPUs.

[Archived Post]

[JRMeyer]

>>> lissyx
[August 19, 2019, 7:23am]

Just for the sake of
completeness, can you make sure you are running tensorflow-gpu ?
pip uninstall tensorflow && pip install --upgrade tensorflow-gpu==1.14.0
in your virtualenv.

[Archived Post]