Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

config update

This commit is contained in:
Eren Golge 2019-04-08 09:56:16 +02:00
parent fc57ac92c2
commit 5afd3d980b
4 changed files with 13 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
.compute
View File

@ -3,5 +3,5 @@ ls ${SHARED_DIR}/data/
pip3 install https://download.pytorch.org/whl/cu100/torch-1.0.1.post2-cp36-cp36m-linux_x86_64.whl
yes | apt-get install espeak
python3 setup.py develop
python3 distribute.py --config_path config_cluster.json --data_path ${SHARED_DIR}/data/Blizzard/Nancy/ --restore_path ${USER_DIR}/best_model_4467.pth.tar
# python3 distribute.py --config_path config_cluster.json --data_path ${SHARED_DIR}/data/Blizzard/Nancy/
# python3 distribute.py --config_path config_cluster.json --data_path ${SHARED_DIR}/data/Blizzard/Nancy/ --restore_path ${USER_DIR}/best_model_4467.pth.tar
python3 distribute.py --config_path config_cluster.json --data_path ${USER_DIR}/Mozilla/

2
config.json
View File

@ -60,7 +60,7 @@
"meta_file_val": "prompts_val.data", // DATASET-RELATED: metafile for evaluation dataloader.
"dataset": "nancy", // 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": 0, // DATASET-RELATED: minimum text length to use in training
"max_seq_len": 120, // DATASET-RELATED: maximum text length
"max_seq_len": 150, // DATASET-RELATED: maximum text length
"output_path": "/media/erogol/data_ssd/Data/models/ljspeech_models/", // DATASET-RELATED: output path for all training outputs.
"num_loader_workers": 8, // 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.

18
config_cluster.json
View File

@ -1,12 +1,12 @@
{
"run_name": "nancy-bn",
"run_description": "Finetune 4467. No ending character ^. and no starting character",
"run_name": "mozilla-fattn",
"run_description": "Mozilla with 0 batch group size and fattn",
"audio":{
// Audio processing parameters
"num_mels": 80, // size of the mel spec frame.
"num_freq": 1025, // number of stft frequency levels. Size of the linear spectogram frame.
"sample_rate": 16000, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled.
"sample_rate": 22050, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled.
"frame_length_ms": 50, // stft window length in ms.
"frame_shift_ms": 12.5, // stft window hop-lengh in ms.
"preemphasis": 0.98, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
@ -41,9 +41,9 @@
"memory_size": 5, // ONLY TACOTRON - memory queue size used to queue network predictions to feed autoregressive connection. Useful if r < 5.
"attention_norm": "softmax", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"use_forward_attn": false, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"use_forward_attn": true, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"batch_size": 32, // Batch size for training. Lower values than 32 might cause hard to learn attention.
"batch_size": 16, // Batch size for training. Lower values than 32 might cause hard to learn attention.
"eval_batch_size":16,
"r": 1, // Number of frames to predict for step.
"wd": 0.000001, // Weight decay weight.
@ -51,16 +51,16 @@
"save_step": 1000, // Number of training steps expected to save traning stats and checkpoints.
"print_step": 10, // Number of steps to log traning on console.
"tb_model_param_stats": true, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging.
"batch_group_size": 8, //Number of batches to shuffle after bucketing.
"batch_group_size": 0, //Number of batches to shuffle after bucketing.
"run_eval": true,
"run_eval": false,
"test_delay_epochs": 10, //Until attention is aligned, testing only wastes computation time.
"data_path": "/media/erogol/data_ssd/Data/LJSpeech-1.1", // DATASET-RELATED: can overwritten from command argument
"meta_file_train": "prompts_train.data", // DATASET-RELATED: metafile for training dataloader.
"meta_file_val": "prompts_val.data", // DATASET-RELATED: metafile for evaluation dataloader.
"dataset": "nancy", // DATASET-RELATED: one of TTS.dataset.preprocessors depending on your target dataset. Use "tts_cache" for pre-computed dataset by extract_features.py
"dataset": "mozilla", // 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": 0, // DATASET-RELATED: minimum text length to use in training
"max_seq_len": 120, // DATASET-RELATED: maximum text length
"max_seq_len": 150, // DATASET-RELATED: maximum text length
"output_path": "../keep/", // DATASET-RELATED: output path for all training outputs.
"num_loader_workers": 8, // 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.

3
datasets/preprocess.py
View File

@ -45,11 +45,10 @@ def tweb(root_path, meta_file):
def mozilla(root_path, meta_files):
"""Normalizes Mozilla meta data files to TTS format"""
import glob
meta_files = glob.glob(root_path + "**/batch*.txt", recursive=True)
meta_files = glob.glob(root_path + "/**/batch*.txt", recursive=True)
folders = [os.path.dirname(f.strip()) for f in meta_files]
items = []
for idx, meta_file in enumerate(meta_files):
print(" | > {}".format(meta_file))
folder = folders[idx]
txt_file = os.path.join(root_path, meta_file)
with open(txt_file, 'r') as ttf: