[Srikanth R]

Can you write in detailed what other steps were you able to run other than synthesis?

DNNGEN or MCD calculation?

[Srikanth R]

use gen_wav_features as a variable in configuration file, similar to output features:
[‘mgc’]

[Srikanth R]

yeah, everything should be in numpy.float32.

I can’t certainly say that this caused the error but still to be corrected.

Also, check norm_info_mgc_251_MVN.dat file. It should contain 512 values with first 251 values representing mean of the data and next 251 values representing variance of the data.

[Srikanth R]

Did you modify the output_feature streams in configuration file?
output_features : [‘mgc’, ‘lf0’, ‘vuv’, ‘bap’] to
output_features : [‘mgc’]

and change the dimension?
[Outputs]
mgc : 60 -> did you change this number?
dmgc : 180 -> did you change this number?

and did you check the number of frames in modified files same as in original data? If yes, did you use the same command as below?
x2x +fa mgc/herald_001.mgc | wc -l

If the answer to all the above questions is Yes, please do check one of the files in the final directories before training:
x2x +fa nn_no_silence_lab_norm_601/herald_001.lab | wc -l (divide this number by 601)
x2x +fa nn_norm_mgc_{dim}/herald_001.cmp | wc -l (divide this number by corresponding dim)

If both remains still same, please let us know.

[Srikanth R]

Yes, the number of frames for input and output should be same. A minor difference (< 5) is allowed.

In your case, I suggest using LSTMs with input as phone-level features (excluding duration features) with number of frames matching the number in output.

[Srikanth R]

Try installing bandmat in virtual environment of python by following below steps:

1. If you don’t have one, you can setup the python virtual environment following the steps in below link:
http://computing.help.inf.ed.ac.uk/python

2. Install bandmat: pip install bandmat

3. If you are running into any gcc errors with numpy while installing bandmat, use below export command (assuming you are using python version 2.7) and run the install command again:
export CFLAGS=-I/usr/lib64/python2.7/site-packages/numpy/core/include/

[Srikanth R]

1. Some of the variables are not being used in the code and hence commented in configuration/configuration.py – you can search for “private_l2_reg” in hyper params section in the file and uncomment it. However, there is no reason to use “run_dnn.py” now, as everything updated in “run_lstm.py”. Please set to “sequential_training” variable to False if you want to run DNN and set to True if you are having either RNN/LSTM in hidden layers. Also, please update the “configuration/configuration.py” with the current version.

2. Please use smaller learning rate for training — or else your training may converge before 10th epoch and hence model won’n be saved. Alternatively, you can comment the line number 313 in current version of the script “run_lstm.py”: which says save the model only after 10th epoch.

if epoch>10: ## comment this line
cPickle.dump(best_dnn_model, open(nnets_file_name, ‘wb’))

[Srikanth R]

This is a bug which was corrected in latest version, please do svn update and you should be able to run without errors.

more specifically about the error:

1. In first step, the script creates its numeric representation and saves it to directory binary_label_189
2. In second step, it removes silence frames from directory “binary_label_189” and saves it to “nn_no_silence_lab_189”. But, if there are no silence segments in label file, it stops writing into this directory. Please comment the line numbers 69, 70 in “frontend/silence_remover.py” in order to correct the bug without updating the repository.

<pre class=”lang:python”>
#if len(nonsilence_indices) == frame_number:
# continue

[Srikanth R]

Also, please use the latest configuration file:
https://svn.ecdf.ed.ac.uk/repo/inf/dnn_tts/configuration/configuration.py

There was a small bug (incorrect data type initialization) which prevented training in sequential mode even when set to True. That was corrected and updated later (from Revision no. 252).

[Srikanth R]

Change the Ossian path in the below files and configure $OSSIAN to your own local path and re-run the step:

Ossian/tools/dnn_tts/recipes/dnn/feed_forward_dnn_ossian.conf
Ossian/tools/dnn_tts/recipes/dnn/feed_forward_dnn_ossian_DUR.conf

[Author]

Posts