mirror of https://github.com/coqui-ai/TTS.git
Remove old Tacotron recipes
This commit is contained in:
Eren Gölge
parent
9631aab0e7
commit
5fa78ee69f
|
|
@ -1,22 +0,0 @@
|
||||||
#!/bin/bash
|
|
||||||
# take the scripts's parent's directory to prefix all the output paths.
|
|
||||||
RUN_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
|
|
||||||
echo $RUN_DIR
|
|
||||||
# # download LJSpeech dataset
|
|
||||||
# wget http://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2
|
|
||||||
# # extract
|
|
||||||
# tar -xjf LJSpeech-1.1.tar.bz2
|
|
||||||
# # create train-val splits
|
|
||||||
# shuf LJSpeech-1.1/metadata.csv > LJSpeech-1.1/metadata_shuf.csv
|
|
||||||
# head -n 12000 LJSpeech-1.1/metadata_shuf.csv > LJSpeech-1.1/metadata_train.csv
|
|
||||||
# tail -n 1100 LJSpeech-1.1/metadata_shuf.csv > LJSpeech-1.1/metadata_val.csv
|
|
||||||
# mv LJSpeech-1.1 $RUN_DIR/
|
|
||||||
# rm LJSpeech-1.1.tar.bz2
|
|
||||||
# # compute dataset mean and variance for normalization
|
|
||||||
# python TTS/bin/compute_statistics.py $RUN_DIR/tacotron2-DDC.json $RUN_DIR/scale_stats.npy --data_path $RUN_DIR/LJSpeech-1.1/wavs/
|
|
||||||
# training ....
|
|
||||||
# change the GPU id if needed
|
|
||||||
CUDA_VISIBLE_DEVICES="0" python TTS/bin/train_tts.py --config_path $RUN_DIR/tacotron2-DCA.json \
|
|
||||||
--coqpit.output_path $RUN_DIR \
|
|
||||||
--coqpit.datasets.0.path /media/erogol/nvme_linux/gdrive/Projects/TTS/recipes/ljspeech/tacotron2-DDC/LJSpeech-1.1/ \
|
|
||||||
--coqpit.audio.stats_path $RUN_DIR/scale_stats.npy \
|
|
||||||
Binary file not shown.
|
|
@ -1,85 +0,0 @@
|
||||||
{
|
|
||||||
"datasets": [
|
|
||||||
{
|
|
||||||
"name": "ljspeech",
|
|
||||||
"path": "DEFINE THIS",
|
|
||||||
"meta_file_train": "metadata.csv",
|
|
||||||
"meta_file_val": null
|
|
||||||
}
|
|
||||||
],
|
|
||||||
"audio": {
|
|
||||||
"fft_size": 1024,
|
|
||||||
"win_length": 1024,
|
|
||||||
"hop_length": 256,
|
|
||||||
"frame_length_ms": null,
|
|
||||||
"frame_shift_ms": null,
|
|
||||||
"sample_rate": 22050,
|
|
||||||
"preemphasis": 0.0,
|
|
||||||
"ref_level_db": 20,
|
|
||||||
"do_trim_silence": true,
|
|
||||||
"trim_db": 60,
|
|
||||||
"power": 1.5,
|
|
||||||
"griffin_lim_iters": 60,
|
|
||||||
"num_mels": 80,
|
|
||||||
"mel_fmin": 50.0,
|
|
||||||
"mel_fmax": 7600.0,
|
|
||||||
"spec_gain": 1,
|
|
||||||
"signal_norm": true,
|
|
||||||
"min_level_db": -100,
|
|
||||||
"symmetric_norm": true,
|
|
||||||
"max_norm": 4.0,
|
|
||||||
"clip_norm": true,
|
|
||||||
"stats_path": "scale_stats.npy"
|
|
||||||
},
|
|
||||||
"distributed_backend": "nlcc",
|
|
||||||
"distributed_url": "tcp:\/\/localhost:54321",
|
|
||||||
"model": "Tacotron2",
|
|
||||||
"run_name": "ljspeech-dca",
|
|
||||||
"run_description": "tacotron2 with dynamic conv attention.",
|
|
||||||
"batch_size": 64,
|
|
||||||
"eval_batch_size": 16,
|
|
||||||
"mixed_precision": true,
|
|
||||||
"loss_masking": true,
|
|
||||||
"decoder_loss_alpha": 0.25,
|
|
||||||
"postnet_loss_alpha": 0.25,
|
|
||||||
"postnet_diff_spec_alpha": 0.25,
|
|
||||||
"decoder_diff_spec_alpha": 0.25,
|
|
||||||
"decoder_ssim_alpha": 0.25,
|
|
||||||
"postnet_ssim_alpha": 0.25,
|
|
||||||
"ga_alpha": 5.0,
|
|
||||||
"stopnet_pos_weight": 15.0,
|
|
||||||
"run_eval": true,
|
|
||||||
"test_delay_epochs": 10,
|
|
||||||
"max_decoder_steps": 1000,
|
|
||||||
"noam_schedule": true,
|
|
||||||
"grad_clip": 0.05,
|
|
||||||
"epochs": 1000,
|
|
||||||
"lr": 0.001,
|
|
||||||
"wd": 1e-06,
|
|
||||||
"warmup_steps": 4000,
|
|
||||||
"memory_size": -1,
|
|
||||||
"prenet_type": "original",
|
|
||||||
"prenet_dropout": true,
|
|
||||||
"attention_type": "dynamic_convolution",
|
|
||||||
"location_attn": true,
|
|
||||||
"attention_norm": "sigmoid",
|
|
||||||
"r": 2,
|
|
||||||
"stopnet": true,
|
|
||||||
"separate_stopnet": true,
|
|
||||||
"print_step": 25,
|
|
||||||
"tb_plot_step": 100,
|
|
||||||
"print_eval": false,
|
|
||||||
"save_step": 10000,
|
|
||||||
"checkpoint": true,
|
|
||||||
"text_cleaner": "phoneme_cleaners",
|
|
||||||
"num_loader_workers": 4,
|
|
||||||
"num_val_loader_workers": 4,
|
|
||||||
"batch_group_size": 4,
|
|
||||||
"min_seq_len": 6,
|
|
||||||
"max_seq_len": 180,
|
|
||||||
"compute_input_seq_cache": true,
|
|
||||||
"output_path": "DEFINE THIS",
|
|
||||||
"phoneme_cache_path": "DEFINE THIS",
|
|
||||||
"use_phonemes": false,
|
|
||||||
"phoneme_language": "en-us"
|
|
||||||
}
|
|
||||||
|
|
@ -1,22 +0,0 @@
|
||||||
#!/bin/bash
|
|
||||||
# take the scripts's parent's directory to prefix all the output paths.
|
|
||||||
RUN_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
|
|
||||||
echo $RUN_DIR
|
|
||||||
# download LJSpeech dataset
|
|
||||||
wget http://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2
|
|
||||||
# extract
|
|
||||||
tar -xjf LJSpeech-1.1.tar.bz2
|
|
||||||
# create train-val splits
|
|
||||||
shuf LJSpeech-1.1/metadata.csv > LJSpeech-1.1/metadata_shuf.csv
|
|
||||||
head -n 12000 LJSpeech-1.1/metadata_shuf.csv > LJSpeech-1.1/metadata_train.csv
|
|
||||||
tail -n 1100 LJSpeech-1.1/metadata_shuf.csv > LJSpeech-1.1/metadata_val.csv
|
|
||||||
mv LJSpeech-1.1 $RUN_DIR/
|
|
||||||
rm LJSpeech-1.1.tar.bz2
|
|
||||||
# compute dataset mean and variance for normalization
|
|
||||||
python TTS/bin/compute_statistics.py $RUN_DIR/tacotron2-DDC.json $RUN_DIR/scale_stats.npy --data_path $RUN_DIR/LJSpeech-1.1/wavs/
|
|
||||||
# training ....
|
|
||||||
# change the GPU id if needed
|
|
||||||
CUDA_VISIBLE_DEVICES="0" python TTS/bin/train_tts.py --config_path $RUN_DIR/tacotron2-DDC.json \
|
|
||||||
--coqpit.output_path $RUN_DIR \
|
|
||||||
--coqpit.datasets.0.path $RUN_DIR/LJSpeech-1.1/ \
|
|
||||||
--coqpit.audio.stats_path $RUN_DIR/scale_stats.npy \
|
|
||||||
Binary file not shown.
|
|
@ -1,94 +0,0 @@
|
||||||
{
|
|
||||||
"model": "Tacotron2",
|
|
||||||
"datasets": [
|
|
||||||
{
|
|
||||||
"name": "ljspeech",
|
|
||||||
"path": "DEFINE THIS",
|
|
||||||
"meta_file_train": "metadata.csv",
|
|
||||||
"meta_file_val": null
|
|
||||||
}
|
|
||||||
],
|
|
||||||
"audio": {
|
|
||||||
"fft_size": 1024,
|
|
||||||
"win_length": 1024,
|
|
||||||
"hop_length": 256,
|
|
||||||
"frame_length_ms": null,
|
|
||||||
"frame_shift_ms": null,
|
|
||||||
"sample_rate": 22050,
|
|
||||||
"preemphasis": 0.0,
|
|
||||||
"ref_level_db": 20,
|
|
||||||
"do_trim_silence": true,
|
|
||||||
"trim_db": 60,
|
|
||||||
"power": 1.5,
|
|
||||||
"griffin_lim_iters": 60,
|
|
||||||
"num_mels": 80,
|
|
||||||
"mel_fmin": 50.0,
|
|
||||||
"mel_fmax": 7600.0,
|
|
||||||
"spec_gain": 1,
|
|
||||||
"signal_norm": true,
|
|
||||||
"min_level_db": -100,
|
|
||||||
"symmetric_norm": true,
|
|
||||||
"max_norm": 4.0,
|
|
||||||
"clip_norm": true,
|
|
||||||
"stats_path": "scale_stats.npy"
|
|
||||||
},
|
|
||||||
"gst":{
|
|
||||||
"gst_embedding_dim": 256,
|
|
||||||
"gst_num_heads": 4,
|
|
||||||
"gst_num_style_tokens": 10
|
|
||||||
},
|
|
||||||
"distributed_backend": "gloo",
|
|
||||||
"distributed_url": "tcp:\/\/localhost:54321",
|
|
||||||
"run_name": "ljspeech-ddc",
|
|
||||||
"run_description": "tacotron2 with double decoder consistency.",
|
|
||||||
"batch_size": 64,
|
|
||||||
"eval_batch_size": 16,
|
|
||||||
"mixed_precision": false,
|
|
||||||
"loss_masking": true,
|
|
||||||
"decoder_loss_alpha": 0.25,
|
|
||||||
"postnet_loss_alpha": 0.25,
|
|
||||||
"postnet_diff_spec_alpha": 0.25,
|
|
||||||
"decoder_diff_spec_alpha": 0.25,
|
|
||||||
"decoder_ssim_alpha": 0.25,
|
|
||||||
"postnet_ssim_alpha": 0.25,
|
|
||||||
"ga_alpha": 5.0,
|
|
||||||
"stopnet_pos_weight": 15.0,
|
|
||||||
"run_eval": true,
|
|
||||||
"test_delay_epochs": 10,
|
|
||||||
"test_sentences_file": null,
|
|
||||||
"max_decoder_steps": 1000,
|
|
||||||
"noam_schedule": true,
|
|
||||||
"grad_clip": 0.05,
|
|
||||||
"epochs": 1000,
|
|
||||||
"lr": 0.001,
|
|
||||||
"wd": 1e-06,
|
|
||||||
"warmup_steps": 4000,
|
|
||||||
"memory_size": -1,
|
|
||||||
"prenet_type": "original",
|
|
||||||
"prenet_dropout": true,
|
|
||||||
"attention_type": "original",
|
|
||||||
"location_attn": true,
|
|
||||||
"double_decoder_consistency": true,
|
|
||||||
"ddc_r": 6,
|
|
||||||
"attention_norm": "sigmoid",
|
|
||||||
"r": 6,
|
|
||||||
"gradual_training": [[0, 6, 64], [10000, 4, 32], [50000, 3, 32], [100000, 2, 32]],
|
|
||||||
"stopnet": true,
|
|
||||||
"separate_stopnet": true,
|
|
||||||
"print_step": 25,
|
|
||||||
"tb_plot_step": 100,
|
|
||||||
"print_eval": false,
|
|
||||||
"save_step": 10000,
|
|
||||||
"checkpoint": true,
|
|
||||||
"text_cleaner": "phoneme_cleaners",
|
|
||||||
"num_loader_workers": 4,
|
|
||||||
"num_val_loader_workers": 4,
|
|
||||||
"batch_group_size": 4,
|
|
||||||
"min_seq_len": 6,
|
|
||||||
"max_seq_len": 180,
|
|
||||||
"compute_input_seq_cache": true,
|
|
||||||
"output_path": "DEFINE THIS",
|
|
||||||
"phoneme_cache_path": "DEFINE THIS",
|
|
||||||
"use_phonemes": false,
|
|
||||||
"phoneme_language": "en-us"
|
|
||||||
}
|
|
||||||
Loading…
Reference in New Issue