mirror of https://github.com/coqui-ai/TTS.git
config updates, m-ai-labs processor update. If no metafileis given parse the ffodler to use all the speakers
This commit is contained in:
Eren Golge
parent
b35b6e8e98
commit
118fe61028
2
.compute
2
.compute
|
|
@ -10,7 +10,7 @@ wget https://www.dropbox.com/s/wqn5v3wkktw9lmo/install.sh?dl=0 -O install.sh
|
|||
sudo sh install.sh
|
||||
python3 setup.py develop
|
||||
# cp -R ${USER_DIR}/GermanData ../tmp/
|
||||
python3 distribute.py --config_path config_tacotron_de.json --data_path ${USER_DIR}/GermanData/karlsson/
|
||||
python3 distribute.py --config_path config_tacotron_de.json --data_path /data/rw/home/de_DE/
|
||||
# cp -R ${USER_DIR}/Mozilla_22050 ../tmp/
|
||||
# python3 distribute.py --config_path config_tacotron_gst.json --data_path ../tmp/Mozilla_22050/
|
||||
while true; do sleep 1000000; done
|
||||
|
|
|
|||
|
|
@ -1,7 +1,6 @@
|
|||
{
|
||||
"github_branch":"tacotron-gst",
|
||||
"run_name": "german-karlsson-tacotron-loc_attn",
|
||||
"run_description": "train german with all of the german dataset",
|
||||
"run_name": "german-all-tacotrongst",
|
||||
"run_description": "train with all the german dataset using gst",
|
||||
|
||||
"audio":{
|
||||
// Audio processing parameters
|
||||
|
|
@ -32,7 +31,7 @@
|
|||
|
||||
"reinit_layers": [],
|
||||
|
||||
"model": "Tacotron", // one of the model in models/
|
||||
"model": "TacotronGST", // one of the model in models/
|
||||
"grad_clip": 1, // upper limit for gradients for clipping.
|
||||
"epochs": 10000, // total number of epochs to train.
|
||||
"lr": 0.0001, // Initial learning rate. If Noam decay is active, maximum learning rate.
|
||||
|
|
@ -43,9 +42,9 @@
|
|||
"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.
|
||||
"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": true,
|
||||
"forward_attn_mask": false,
|
||||
"location_attn": true, // ONLY TACOTRON2 - 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.
|
||||
|
|
@ -66,51 +65,16 @@
|
|||
"test_sentences_file": "de_sentences.txt", // set a file to load sentences to be used for testing. If it is null then we use default english sentences.
|
||||
"test_delay_epochs": 5, //Until attention is aligned, testing only wastes computation time.
|
||||
"data_path": "/home/erogol/Data/m-ai-labs/de_DE/by_book/" , // DATASET-RELATED: can overwritten from command argument
|
||||
"meta_file_train": [
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/mix/erzaehlungen_poe/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/mix/auf_zwei_planeten/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/kleinzaches/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/spiegel_kaetzchen/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/herrnarnesschatz/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/maedchen_von_moorhof/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/koenigsgaukler/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/altehous/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/odysseus/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/undine/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/reise_tilsit/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/schmied_seines_glueckes/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/kammmacher/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/unterm_birnbaum/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/liebesbriefe/metadata.csv",
|
||||
"/home/erogol/Data/m-ai-labs/de_DE/by_book/male/karlsson/sandmann/metadata.csv"
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/eva_k/kleine_lord/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/eva_k/toten_seelen/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/eva_k/werde_die_du_bist/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/eva_k/grune_haus/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/rebecca_braunert_plunkett/das_letzte_marchen/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/rebecca_braunert_plunkett/ferien_vom_ich/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/rebecca_braunert_plunkett/maerchen/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/rebecca_braunert_plunkett/mein_weg_als_deutscher_und_jude/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/caspar/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/sterben/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/weihnachtsabend/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/frankenstein/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/tschun/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/menschenhasser/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/grune_gesicht/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/tom_sawyer/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/ramona_deininger/alter_afrikaner/metadata.csv",
|
||||
// "/home/erogol/Data/m-ai-labs/de_DE/by_book/female/angela_merkel/merkel_alone/metadata.csv"
|
||||
], // DATASET-RELATED: metafile for training dataloader.
|
||||
"meta_file_train": null, // DATASET-RELATED: metafile for training dataloader.
|
||||
"meta_file_val": null, // DATASET-RELATED: metafile for evaluation dataloader.
|
||||
"dataset": "mailabs", // DATASET-RELATED: one of TTS.dataset.preprocessors depending on your target dataset. Use "tts_cache" for pre-computed dataset by extract_features.py
|
||||
"min_seq_len": 15, // DATASET-RELATED: minimum text length to use in training
|
||||
"max_seq_len": 200, // DATASET-RELATED: maximum text length
|
||||
"output_path": "/media/erogol/data_ssd/Data/models/mozilla_models/", // DATASET-RELATED: output path for all training outputs.
|
||||
"output_path": "/home/erogol/Models/mozilla_models/", // DATASET-RELATED: output path for all training outputs.
|
||||
"num_loader_workers": 4, // number of training data loader processes. Don't set it too big. 4-8 are good values.
|
||||
"num_val_loader_workers": 4, // number of evaluation data loader processes.
|
||||
"phoneme_cache_path": "phoneme_cache", // phoneme computation is slow, therefore, it caches results in the given folder.
|
||||
"use_phonemes": true, // use phonemes instead of raw characters. It is suggested for better pronounciation.
|
||||
"use_phonemes": false, // use phonemes instead of raw characters. It is suggested for better pronounciation.
|
||||
"phoneme_language": "de", // depending on your target language, pick one from https://github.com/bootphon/phonemizer#languages
|
||||
"text_cleaner": "phoneme_cleaners"
|
||||
}
|
||||
|
|
|
|||
|
|
@ -62,20 +62,24 @@ def mozilla(root_path, meta_file):
|
|||
def mailabs(root_path, meta_files):
|
||||
"""Normalizes M-AI-Labs meta data files to TTS format"""
|
||||
if meta_files is None:
|
||||
meta_files = glob(root_path+"/**/metadata.csv", recursive=True)
|
||||
folders = [f.strip().split("/")[-2] for f in meta_files]
|
||||
csv_files = glob(root_path+"/**/metadata.csv", recursive=True)
|
||||
folders = [os.path.dirname(f) for f in csv_files]
|
||||
else:
|
||||
folders = [f.strip().split("by_book")[1][1:] for f in meta_files]
|
||||
csv_files = meta_files
|
||||
folders = [f.strip().split("by_book")[1][1:] for f in csv_file]
|
||||
# meta_files = [f.strip() for f in meta_files.split(",")]
|
||||
items = []
|
||||
for idx, meta_file in enumerate(meta_files):
|
||||
print(" | > {}".format(meta_file))
|
||||
for idx, csv_file in enumerate(csv_files):
|
||||
print(" | > {}".format(csv_file))
|
||||
folder = folders[idx]
|
||||
txt_file = os.path.join(root_path, meta_file)
|
||||
txt_file = os.path.join(root_path, csv_file)
|
||||
with open(txt_file, 'r') as ttf:
|
||||
for line in ttf:
|
||||
cols = line.split('|')
|
||||
wav_file = os.path.join(root_path, folder.replace("metadata.csv", ""), 'wavs', cols[0] + '.wav')
|
||||
if meta_files is None:
|
||||
wav_file = os.path.join(folder, 'wavs', cols[0] + '.wav')
|
||||
else:
|
||||
wav_file = os.path.join(root_path, folder.replace("metadata.csv", ""), 'wavs', cols[0] + '.wav')
|
||||
if os.path.isfile(wav_file):
|
||||
text = cols[1].strip()
|
||||
items.append([text, wav_file])
|
||||
|
|
|
|||
Loading…
Reference in New Issue