mozilla
/
TTS
зеркало из https://github.com/mozilla/TTS.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Merge branch 'dev' of github.com:mozilla/TTS into stylemel_in_testing

This commit is contained in:
Thomas Werkmeister 2019-07-25 13:23:41 +02:00
Родитель 4a23354d3c 0261ce79a6
Коммит 11f9edd849
7 изменённых файлов: 30 добавлений и 32 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

12
config.json
Просмотреть файл

@ -40,12 +40,12 @@
"windowing": false, // Enables attention windowing. Used only in eval mode.
"memory_size": 5, // ONLY TACOTRON - memory queue size used to queue network predictions to feed autoregressive connection. Useful if r < 5.
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet.
"use_forward_attn": true, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"forward_attn_mask": false,
"transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention.
"location_attn": false, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"use_forward_attn": true, // enable/disable forward attention. In general, it aligns faster.
"forward_attn_mask": false, // Apply forward attention mask at inference to prevent bad modes. Try it if your model does not align well.
"transition_agent": false, // enable/disable transition agent of forward attention.
"location_attn": false, // enable_disable location sensitive attention.
"loss_masking": true, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": true, // Train stopnet predicting the end of synthesis.

10
config_libritts.json
Просмотреть файл

@ -39,13 +39,13 @@
"warmup_steps": 4000, // Noam decay steps to increase the learning rate from 0 to "lr"
"memory_size": 5, // ONLY TACOTRON - memory queue size used to queue network predictions to feed autoregressive connection. Useful if r < 5.
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"windowing": false, // Enables attention windowing. Used only in eval mode.
"use_forward_attn": false, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"use_forward_attn": false, // enable/disable forward attention. In general, it aligns faster.
"forward_attn_mask": false,
"transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention.
"location_attn": true, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"transition_agent": false, // enable/disable transition agent of forward attention.
"location_attn": true, // enable_disable location sensitive attention.
"loss_masking": true, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": true, // Train stopnet predicting the end of synthesis.

4
config_tacotron.json
Просмотреть файл

@ -42,10 +42,10 @@
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"use_forward_attn": true, // if it uses forward attention. In general, it aligns faster.
"use_forward_attn": true, // enable/disable forward attention. In general, it aligns faster.
"forward_attn_mask": false, // Apply forward attention mask af inference to prevent bad modes. Try it if your model does not align well.
"transition_agent": true, // enable/disable transition agent of forward attention.
"location_attn": false, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"location_attn": false, // enable_disable location sensitive attention.
"loss_masking": true, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": true, // Train stopnet predicting the end of synthesis.

10
config_tacotron2.json
Просмотреть файл

@ -39,12 +39,12 @@
"warmup_steps": 4000, // Noam decay steps to increase the learning rate from 0 to "lr"
"memory_size": 5, // ONLY TACOTRON - memory queue size used to queue network predictions to feed autoregressive connection. Useful if r < 5.
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet.
"use_forward_attn": true, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"use_forward_attn": true, // enable/disable forward attention. In general, it aligns faster.
"forward_attn_mask": false, // Apply forward attention mask af inference to prevent bad modes. Try it if your model does not align well.
"transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention.
"location_attn": false, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"transition_agent": false, // enable/disable transition agent of forward attention.
"location_attn": false, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"loss_masking": true, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": true, // Train stopnet predicting the end of synthesis.

12
config_tacotron_de.json
Просмотреть файл

@ -40,12 +40,12 @@
"windowing": false, // Enables attention windowing. Used only in eval mode.
"memory_size": 5, // ONLY TACOTRON - memory queue size used to queue network predictions to feed autoregressive connection. Useful if r < 5.
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet.
"use_forward_attn": false, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention.
"forward_attn_mask": false,
"location_attn": true, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"use_forward_attn": false, // enable/disable forward attention. In general, it aligns faster.
"transition_agent": false, // enable/disable transition agent of forward attention.
"forward_attn_mask": false, // Apply forward attention mask at inference to prevent bad modes. Try it if your model does not align well.
"location_attn": true, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"loss_masking": true, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": true, // Train stopnet predicting the end of synthesis.

4
config_tacotron_gst.json
Просмотреть файл

@ -42,8 +42,8 @@
"attention_norm": "sigmoid", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // "original" or "bn".
"prenet_dropout": true, // enable/disable dropout at prenet.
"use_forward_attn": true, // if it uses forward attention. In general, it aligns faster.
"forward_attn_mask": false, // Apply forward attention mask af inference to prevent bad modes. Try it if your model does not align well.
"use_forward_attn": true, // enable/disable forward attention. In general, it aligns faster.
"forward_attn_mask": false, // Apply forward attention mask at inference to prevent bad modes. Try it if your model does not align well.
"transition_agent": false, // enable/disable transition agent of forward attention.
"location_attn": false, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"loss_masking": true, // enable / disable loss masking against the sequence padding.

10
datasets/preprocess.py
Просмотреть файл

@ -75,21 +75,19 @@ def mailabs(root_path, meta_files=None):
speaker_regex = re.compile("by_book/(male|female)/(?P<speaker_name>[^/]+)/")
if meta_files is None:
csv_files = glob(root_path+"/**/metadata.csv", recursive=True)
folders = [os.path.dirname(f) for f in csv_files]
else:
csv_files = meta_files
folders = [f.strip().split("by_book")[1][1:] for f in csv_files]
# meta_files = [f.strip() for f in meta_files.split(",")]
items = []
for idx, csv_file in enumerate(csv_files):
for csv_file in csv_files:
txt_file = os.path.join(root_path, csv_file)
folder = os.path.dirname(txt_file)
# determine speaker based on folder structure...
speaker_name_match = speaker_regex.search(csv_file)
speaker_name_match = speaker_regex.search(txt_file)
if speaker_name_match is None:
continue
speaker_name = speaker_name_match.group("speaker_name")
print(" | > {}".format(csv_file))
folder = folders[idx]
txt_file = os.path.join(root_path, csv_file)
with open(txt_file, 'r') as ttf:
for line in ttf:
cols = line.split('|')