mirror of
https://github.com/wassname/TTS.git
synced 2026-09-11 12:00:24 +08:00
Mass refactoring
This commit is contained in:
erogol
@@ -1,8 +1,8 @@
|
||||
import unittest
|
||||
import torch as T
|
||||
|
||||
from TTS.utils.generic_utils import save_checkpoint, save_best_model
|
||||
from TTS.layers.tacotron import Prenet
|
||||
from TTS.tts.utils.generic_utils import save_checkpoint, save_best_model
|
||||
from TTS.tts.layers.tacotron import Prenet
|
||||
|
||||
OUT_PATH = '/tmp/test.pth.tar'
|
||||
|
||||
|
||||
@@ -0,0 +1,151 @@
|
||||
{
|
||||
"model": "Tacotron2",
|
||||
"run_name": "test_sample_dataset_run",
|
||||
"run_description": "sample dataset test run",
|
||||
|
||||
// AUDIO PARAMETERS
|
||||
"audio":{
|
||||
// stft parameters
|
||||
"fft_size": 1024, // number of stft frequency levels. Size of the linear spectogram frame.
|
||||
"win_length": 1024, // stft window length in ms.
|
||||
"hop_length": 256, // stft window hop-lengh in ms.
|
||||
"frame_length_ms": null, // stft window length in ms.If null, 'win_length' is used.
|
||||
"frame_shift_ms": null, // stft window hop-lengh in ms. If null, 'hop_length' is used.
|
||||
|
||||
// Audio processing parameters
|
||||
"sample_rate": 22050, // DATASET-RELATED: wav sample-rate.
|
||||
"preemphasis": 0.0, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
|
||||
"ref_level_db": 20, // reference level db, theoretically 20db is the sound of air.
|
||||
|
||||
// Silence trimming
|
||||
"do_trim_silence": true,// enable trimming of slience of audio as you load it. LJspeech (true), TWEB (false), Nancy (true)
|
||||
"trim_db": 60, // threshold for timming silence. Set this according to your dataset.
|
||||
|
||||
// Griffin-Lim
|
||||
"power": 1.5, // value to sharpen wav signals after GL algorithm.
|
||||
"griffin_lim_iters": 60,// #griffin-lim iterations. 30-60 is a good range. Larger the value, slower the generation.
|
||||
|
||||
// MelSpectrogram parameters
|
||||
"num_mels": 80, // size of the mel spec frame.
|
||||
"mel_fmin": 0.0, // minimum freq level for mel-spec. ~50 for male and ~95 for female voices. Tune for dataset!!
|
||||
"mel_fmax": 8000.0, // maximum freq level for mel-spec. Tune for dataset!!
|
||||
"spec_gain": 20.0,
|
||||
|
||||
// Normalization parameters
|
||||
"signal_norm": true, // normalize spec values. Mean-Var normalization if 'stats_path' is defined otherwise range normalization defined by the other params.
|
||||
"min_level_db": -100, // lower bound for normalization
|
||||
"symmetric_norm": true, // move normalization to range [-1, 1]
|
||||
"max_norm": 4.0, // scale normalization to range [-max_norm, max_norm] or [0, max_norm]
|
||||
"clip_norm": true, // clip normalized values into the range.
|
||||
"stats_path": null // DO NOT USE WITH MULTI_SPEAKER MODEL. scaler stats file computed by 'compute_statistics.py'. If it is defined, mean-std based notmalization is used and other normalization params are ignored
|
||||
},
|
||||
|
||||
// VOCABULARY PARAMETERS
|
||||
// if custom character set is not defined,
|
||||
// default set in symbols.py is used
|
||||
// "characters":{
|
||||
// "pad": "_",
|
||||
// "eos": "~",
|
||||
// "bos": "^",
|
||||
// "characters": "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!'(),-.:;? ",
|
||||
// "punctuations":"!'(),-.:;? ",
|
||||
// "phonemes":"iyɨʉɯuɪʏʊeøɘəɵɤoɛœɜɞʌɔæɐaɶɑɒᵻʘɓǀɗǃʄǂɠǁʛpbtdʈɖcɟkɡqɢʔɴŋɲɳnɱmʙrʀⱱɾɽɸβfvθðszʃʒʂʐçʝxɣχʁħʕhɦɬɮʋɹɻjɰlɭʎʟˈˌːˑʍwɥʜʢʡɕʑɺɧɚ˞ɫ"
|
||||
// },
|
||||
|
||||
// DISTRIBUTED TRAINING
|
||||
"distributed":{
|
||||
"backend": "nccl",
|
||||
"url": "tcp:\/\/localhost:54321"
|
||||
},
|
||||
|
||||
"reinit_layers": [], // give a list of layer names to restore from the given checkpoint. If not defined, it reloads all heuristically matching layers.
|
||||
|
||||
// TRAINING
|
||||
"batch_size": 1, // Batch size for training. Lower values than 32 might cause hard to learn attention. It is overwritten by 'gradual_training'.
|
||||
"eval_batch_size":1,
|
||||
"r": 7, // Number of decoder frames to predict per iteration. Set the initial values if gradual training is enabled.
|
||||
"gradual_training": [[0, 7, 4]], //set gradual training steps [first_step, r, batch_size]. If it is null, gradual training is disabled. For Tacotron, you might need to reduce the 'batch_size' as you proceeed.
|
||||
"loss_masking": true, // enable / disable loss masking against the sequence padding.
|
||||
"ga_alpha": 10.0, // weight for guided attention loss. If > 0, guided attention is enabled.
|
||||
|
||||
// VALIDATION
|
||||
"run_eval": true,
|
||||
"test_delay_epochs": 0, //Until attention is aligned, testing only wastes computation time.
|
||||
"test_sentences_file": null, // set a file to load sentences to be used for testing. If it is null then we use default english sentences.
|
||||
|
||||
// OPTIMIZER
|
||||
"noam_schedule": false, // use noam warmup and lr schedule.
|
||||
"grad_clip": 1.0, // upper limit for gradients for clipping.
|
||||
"epochs": 1, // total number of epochs to train.
|
||||
"lr": 0.0001, // Initial learning rate. If Noam decay is active, maximum learning rate.
|
||||
"wd": 0.000001, // Weight decay weight.
|
||||
"warmup_steps": 4000, // Noam decay steps to increase the learning rate from 0 to "lr"
|
||||
"seq_len_norm": false, // Normalize eash sample loss with its length to alleviate imbalanced datasets. Use it if your dataset is small or has skewed distribution of sequence lengths.
|
||||
|
||||
// TACOTRON PRENET
|
||||
"memory_size": -1, // ONLY TACOTRON - size of the memory queue used fro storing last decoder predictions for auto-regression. If < 0, memory queue is disabled and decoder only uses the last prediction frame.
|
||||
"prenet_type": "bn", // "original" or "bn".
|
||||
"prenet_dropout": false, // enable/disable dropout at prenet.
|
||||
|
||||
// TACOTRON ATTENTION
|
||||
"attention_type": "original", // 'original' or 'graves'
|
||||
"attention_heads": 4, // number of attention heads (only for 'graves')
|
||||
"attention_norm": "sigmoid", // softmax or sigmoid.
|
||||
"windowing": false, // Enables attention windowing. Used only in eval mode.
|
||||
"use_forward_attn": false, // if it uses forward attention. In general, it aligns faster.
|
||||
"forward_attn_mask": false, // Additional masking forcing monotonicity only in eval mode.
|
||||
"transition_agent": false, // enable/disable transition agent of forward attention.
|
||||
"location_attn": true, // enable_disable location sensitive attention. It is enabled for TACOTRON by default.
|
||||
"bidirectional_decoder": false, // use https://arxiv.org/abs/1907.09006. Use it, if attention does not work well with your dataset.
|
||||
"double_decoder_consistency": true, // use DDC explained here https://erogol.com/solving-attention-problems-of-tts-models-with-double-decoder-consistency-draft/
|
||||
"ddc_r": 7, // reduction rate for coarse decoder.
|
||||
|
||||
// STOPNET
|
||||
"stopnet": true, // Train stopnet predicting the end of synthesis.
|
||||
"separate_stopnet": true, // Train stopnet seperately if 'stopnet==true'. It prevents stopnet loss to influence the rest of the model. It causes a better model, but it trains SLOWER.
|
||||
|
||||
// TENSORBOARD and LOGGING
|
||||
"print_step": 1, // Number of steps to log training on console.
|
||||
"tb_plot_step": 100, // Number of steps to plot TB training figures.
|
||||
"print_eval": false, // If True, it prints intermediate loss values in evalulation.
|
||||
"save_step": 10000, // Number of training steps expected to save traninpg stats and checkpoints.
|
||||
"checkpoint": true, // If true, it saves checkpoints per "save_step"
|
||||
"tb_model_param_stats": false, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging.
|
||||
|
||||
// DATA LOADING
|
||||
"text_cleaner": "phoneme_cleaners",
|
||||
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
|
||||
"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.
|
||||
"batch_group_size": 0, //Number of batches to shuffle after bucketing.
|
||||
"min_seq_len": 6, // DATASET-RELATED: minimum text length to use in training
|
||||
"max_seq_len": 153, // DATASET-RELATED: maximum text length
|
||||
|
||||
// PATHS
|
||||
"output_path": "tests/train_outputs/",
|
||||
|
||||
// PHONEMES
|
||||
"phoneme_cache_path": "tests/train_outputs/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.
|
||||
"phoneme_language": "en-us", // depending on your target language, pick one from https://github.com/bootphon/phonemizer#languages
|
||||
|
||||
// MULTI-SPEAKER and GST
|
||||
"use_speaker_embedding": false, // use speaker embedding to enable multi-speaker learning.
|
||||
"style_wav_for_test": null, // path to style wav file to be used in TacotronGST inference.
|
||||
"use_gst": false, // TACOTRON ONLY: use global style tokens
|
||||
|
||||
// DATASETS
|
||||
"train_portion": 0.1, // dataset portion used for training. It is mainly for internal experiments.
|
||||
"eval_portion": 0.1, // dataset portion used for training. It is mainly for internal experiments.
|
||||
"datasets": // List of datasets. They all merged and they get different speaker_ids.
|
||||
[
|
||||
{
|
||||
"name": "ljspeech",
|
||||
"path": "tests/data/ljspeech/",
|
||||
"meta_file_train": "metadata.csv",
|
||||
"meta_file_val": "metadata.csv"
|
||||
}
|
||||
]
|
||||
|
||||
}
|
||||
|
||||
@@ -0,0 +1,24 @@
|
||||
{
|
||||
"audio":{
|
||||
"num_mels": 80, // size of the mel spec frame.
|
||||
"num_freq": 513, // number of stft frequency levels. Size of the linear spectogram frame.
|
||||
"sample_rate": 22050, // wav sample-rate. If different than the original data, it is resampled.
|
||||
"frame_length_ms": null, // stft window length in ms.
|
||||
"frame_shift_ms": null, // stft window hop-lengh in ms.
|
||||
"hop_length": 256,
|
||||
"win_length": 1024,
|
||||
"preemphasis": 0.97, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
|
||||
"min_level_db": -100, // normalization range
|
||||
"ref_level_db": 20, // reference level db, theoretically 20db is the sound of air.
|
||||
"power": 1.5, // value to sharpen wav signals after GL algorithm.
|
||||
"griffin_lim_iters": 30,// #griffin-lim iterations. 30-60 is a good range. Larger the value, slower the generation.
|
||||
"signal_norm": true, // normalize the spec values in range [0, 1]
|
||||
"symmetric_norm": true, // move normalization to range [-1, 1]
|
||||
"clip_norm": true, // clip normalized values into the range.
|
||||
"max_norm": 4, // scale normalization to range [-max_norm, max_norm] or [0, max_norm]
|
||||
"mel_fmin": 0, // minimum freq level for mel-spec. ~50 for male and ~95 for female voices. Tune for dataset!!
|
||||
"mel_fmax": 8000, // maximum freq level for mel-spec. Tune for dataset!!
|
||||
"do_trim_silence": false
|
||||
}
|
||||
}
|
||||
|
||||
@@ -0,0 +1,144 @@
|
||||
{
|
||||
"run_name": "multiband-melgan",
|
||||
"run_description": "multiband melgan mean-var scaling",
|
||||
|
||||
// AUDIO PARAMETERS
|
||||
"audio":{
|
||||
"fft_size": 1024, // number of stft frequency levels. Size of the linear spectogram frame.
|
||||
"win_length": 1024, // stft window length in ms.
|
||||
"hop_length": 256, // stft window hop-lengh in ms.
|
||||
"frame_length_ms": null, // stft window length in ms.If null, 'win_length' is used.
|
||||
"frame_shift_ms": null, // stft window hop-lengh in ms. If null, 'hop_length' is used.
|
||||
|
||||
// Audio processing parameters
|
||||
"sample_rate": 22050, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled.
|
||||
"preemphasis": 0.0, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
|
||||
"ref_level_db": 0, // reference level db, theoretically 20db is the sound of air.
|
||||
|
||||
// Silence trimming
|
||||
"do_trim_silence": true,// enable trimming of slience of audio as you load it. LJspeech (false), TWEB (false), Nancy (true)
|
||||
"trim_db": 60, // threshold for timming silence. Set this according to your dataset.
|
||||
|
||||
// MelSpectrogram parameters
|
||||
"num_mels": 80, // size of the mel spec frame.
|
||||
"mel_fmin": 50.0, // minimum freq level for mel-spec. ~50 for male and ~95 for female voices. Tune for dataset!!
|
||||
"mel_fmax": 7600.0, // maximum freq level for mel-spec. Tune for dataset!!
|
||||
"spec_gain": 1.0, // scaler value appplied after log transform of spectrogram.
|
||||
|
||||
// Normalization parameters
|
||||
"signal_norm": true, // normalize spec values. Mean-Var normalization if 'stats_path' is defined otherwise range normalization defined by the other params.
|
||||
"min_level_db": -100, // lower bound for normalization
|
||||
"symmetric_norm": true, // move normalization to range [-1, 1]
|
||||
"max_norm": 4.0, // scale normalization to range [-max_norm, max_norm] or [0, max_norm]
|
||||
"clip_norm": true, // clip normalized values into the range.
|
||||
"stats_path": null // DO NOT USE WITH MULTI_SPEAKER MODEL. scaler stats file computed by 'compute_statistics.py'. If it is defined, mean-std based notmalization is used and other normalization params are ignored
|
||||
},
|
||||
|
||||
// DISTRIBUTED TRAINING
|
||||
// "distributed":{
|
||||
// "backend": "nccl",
|
||||
// "url": "tcp:\/\/localhost:54321"
|
||||
// },
|
||||
|
||||
// MODEL PARAMETERS
|
||||
"use_pqmf": true,
|
||||
|
||||
// LOSS PARAMETERS
|
||||
"use_stft_loss": true,
|
||||
"use_subband_stft_loss": true,
|
||||
"use_mse_gan_loss": true,
|
||||
"use_hinge_gan_loss": false,
|
||||
"use_feat_match_loss": false, // use only with melgan discriminators
|
||||
|
||||
// loss weights
|
||||
"stft_loss_weight": 0.5,
|
||||
"subband_stft_loss_weight": 0.5,
|
||||
"mse_G_loss_weight": 2.5,
|
||||
"hinge_G_loss_weight": 2.5,
|
||||
"feat_match_loss_weight": 25,
|
||||
|
||||
// multiscale stft loss parameters
|
||||
"stft_loss_params": {
|
||||
"n_ffts": [1024, 2048, 512],
|
||||
"hop_lengths": [120, 240, 50],
|
||||
"win_lengths": [600, 1200, 240]
|
||||
},
|
||||
|
||||
// subband multiscale stft loss parameters
|
||||
"subband_stft_loss_params":{
|
||||
"n_ffts": [384, 683, 171],
|
||||
"hop_lengths": [30, 60, 10],
|
||||
"win_lengths": [150, 300, 60]
|
||||
},
|
||||
|
||||
"target_loss": "avg_G_loss", // loss value to pick the best model to save after each epoch
|
||||
|
||||
// DISCRIMINATOR
|
||||
"discriminator_model": "melgan_multiscale_discriminator",
|
||||
"discriminator_model_params":{
|
||||
"base_channels": 16,
|
||||
"max_channels":512,
|
||||
"downsample_factors":[4, 4, 4]
|
||||
},
|
||||
"steps_to_start_discriminator": 200000, // steps required to start GAN trainining.1
|
||||
|
||||
// GENERATOR
|
||||
"generator_model": "multiband_melgan_generator",
|
||||
"generator_model_params": {
|
||||
"upsample_factors":[8, 4, 2],
|
||||
"num_res_blocks": 4
|
||||
},
|
||||
|
||||
// DATASET
|
||||
"data_path": "tests/data/ljspeech/wavs/",
|
||||
"feature_path": null,
|
||||
"seq_len": 16384,
|
||||
"pad_short": 2000,
|
||||
"conv_pad": 0,
|
||||
"use_noise_augment": false,
|
||||
"use_cache": true,
|
||||
|
||||
"reinit_layers": [], // give a list of layer names to restore from the given checkpoint. If not defined, it reloads all heuristically matching layers.
|
||||
|
||||
// TRAINING
|
||||
"batch_size": 4, // Batch size for training. Lower values than 32 might cause hard to learn attention. It is overwritten by 'gradual_training'.
|
||||
|
||||
// VALIDATION
|
||||
"run_eval": true,
|
||||
"test_delay_epochs": 10, //Until attention is aligned, testing only wastes computation time.
|
||||
"test_sentences_file": null, // set a file to load sentences to be used for testing. If it is null then we use default english sentences.
|
||||
|
||||
// OPTIMIZER
|
||||
"epochs": 1, // total number of epochs to train.
|
||||
"wd": 0.0, // Weight decay weight.
|
||||
"gen_clip_grad": -1, // Generator gradient clipping threshold. Apply gradient clipping if > 0
|
||||
"disc_clip_grad": -1, // Discriminator gradient clipping threshold.
|
||||
"lr_scheduler_gen": "MultiStepLR", // one of the schedulers from https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
|
||||
"lr_scheduler_gen_params": {
|
||||
"gamma": 0.5,
|
||||
"milestones": [100000, 200000, 300000, 400000, 500000, 600000]
|
||||
},
|
||||
"lr_scheduler_disc": "MultiStepLR", // one of the schedulers from https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
|
||||
"lr_scheduler_disc_params": {
|
||||
"gamma": 0.5,
|
||||
"milestones": [100000, 200000, 300000, 400000, 500000, 600000]
|
||||
},
|
||||
"lr_gen": 1e-4, // Initial learning rate. If Noam decay is active, maximum learning rate.
|
||||
"lr_disc": 1e-4,
|
||||
|
||||
// TENSORBOARD and LOGGING
|
||||
"print_step": 1, // Number of steps to log traning on console.
|
||||
"print_eval": false, // If True, it prints loss values for each step in eval run.
|
||||
"save_step": 25000, // Number of training steps expected to plot training stats on TB and save model checkpoints.
|
||||
"checkpoint": true, // If true, it saves checkpoints per "save_step"
|
||||
"tb_model_param_stats": false, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging.
|
||||
|
||||
// DATA LOADING
|
||||
"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.
|
||||
"eval_split_size": 10,
|
||||
|
||||
// PATHS
|
||||
"output_path": "tests/outputs/train_outputs/"
|
||||
}
|
||||
|
||||
@@ -1,88 +0,0 @@
|
||||
{
|
||||
"run_name": "mozilla-no-loc-fattn-stopnet-sigmoid-loss_masking",
|
||||
"run_description": "using forward attention, with original prenet, loss masking,separate stopnet, sigmoid. Compare this with 4817. Pytorch DPP",
|
||||
|
||||
"audio":{
|
||||
// Audio processing parameters
|
||||
"num_mels": 80, // size of the mel spec frame.
|
||||
"fft_size": 1024, // number of stft frequency levels. Size of the linear spectogram frame.
|
||||
"sample_rate": 22050, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled.
|
||||
"hop_length": 256,
|
||||
"win_length": 1024,
|
||||
"preemphasis": 0.98, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis.
|
||||
"min_level_db": -100, // normalization range
|
||||
"ref_level_db": 20, // reference level db, theoretically 20db is the sound of air.
|
||||
"power": 1.5, // value to sharpen wav signals after GL algorithm.
|
||||
"griffin_lim_iters": 60,// #griffin-lim iterations. 30-60 is a good range. Larger the value, slower the generation.
|
||||
// Normalization parameters
|
||||
"signal_norm": true, // normalize the spec values in range [0, 1]
|
||||
"symmetric_norm": false, // move normalization to range [-1, 1]
|
||||
"max_norm": 1, // scale normalization to range [-max_norm, max_norm] or [0, max_norm]
|
||||
"clip_norm": true, // clip normalized values into the range.
|
||||
"mel_fmin": 0.0, // minimum freq level for mel-spec. ~50 for male and ~95 for female voices. Tune for dataset!!
|
||||
"mel_fmax": 8000.0, // maximum freq level for mel-spec. Tune for dataset!!
|
||||
"do_trim_silence": true // enable trimming of slience of audio as you load it. LJspeech (false), TWEB (false), Nancy (true)
|
||||
},
|
||||
|
||||
"distributed":{
|
||||
"backend": "nccl",
|
||||
"url": "tcp:\/\/localhost:54321"
|
||||
},
|
||||
|
||||
"reinit_layers": [],
|
||||
|
||||
"model": "Tacotron2", // one of the model in models/
|
||||
"grad_clip": 1, // upper limit for gradients for clipping.
|
||||
"epochs": 1000, // total number of epochs to train.
|
||||
"lr": 0.0001, // Initial learning rate. If Noam decay is active, maximum learning rate.
|
||||
"lr_decay": false, // if true, Noam learning rate decaying is applied through training.
|
||||
"warmup_steps": 4000, // Noam decay steps to increase the learning rate from 0 to "lr"
|
||||
"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,
|
||||
"attention_type": "original",
|
||||
"attention_heads": 5,
|
||||
"bidirectional_decoder": 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.
|
||||
"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.
|
||||
"separate_stopnet": true, // Train stopnet seperately if 'stopnet==true'. It prevents stopnet loss to influence the rest of the model. It causes a better model, but it trains SLOWER.
|
||||
"tb_model_param_stats": false, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging.
|
||||
"use_gst": false,
|
||||
"double_decoder_consistency": true, // use DDC explained here https://erogol.com/solving-attention-problems-of-tts-models-with-double-decoder-consistency-draft/
|
||||
"ddc_r": 7, // reduction rate for coarse decoder.
|
||||
|
||||
"batch_size": 32, // 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.
|
||||
"checkpoint": true, // If true, it saves checkpoints per "save_step"
|
||||
"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.
|
||||
"batch_group_size": 0, //Number of batches to shuffle after bucketing.
|
||||
|
||||
"run_eval": true,
|
||||
"test_delay_epochs": 5, //Until attention is aligned, testing only wastes computation time.
|
||||
"test_sentences_file": null, // set a file to load sentences to be used for testing. If it is null then we use default english sentences.
|
||||
"data_path": "/media/erogol/data_ssd/Data/Mozilla/", // DATASET-RELATED: can overwritten from command argument
|
||||
"meta_file_train": "metadata_train.txt", // DATASET-RELATED: metafile for training dataloader.
|
||||
"meta_file_val": "metadata_val.txt", // DATASET-RELATED: metafile for evaluation dataloader.
|
||||
"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": 150, // DATASET-RELATED: maximum text length
|
||||
"output_path": "../keep/", // 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": "mozilla_us_phonemes", // phoneme computation is slow, therefore, it caches results in the given folder.
|
||||
"use_phonemes": false, // use phonemes instead of raw characters. It is suggested for better pronounciation.
|
||||
"phoneme_language": "en-us", // depending on your target language, pick one from https://github.com/bootphon/phonemizer#languages
|
||||
"text_cleaner": "phoneme_cleaners",
|
||||
"use_speaker_embedding": false // whether to use additional embeddings for separate speakers
|
||||
}
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
import unittest
|
||||
|
||||
from TTS.utils.text import phonemes
|
||||
from TTS.tts.utils.text import phonemes
|
||||
|
||||
class SymbolsTest(unittest.TestCase):
|
||||
def test_uniqueness(self): #pylint: disable=no-self-use
|
||||
|
||||
+2
-2
@@ -1,7 +1,7 @@
|
||||
import os
|
||||
import unittest
|
||||
|
||||
from TTS.tests import get_tests_path, get_tests_input_path, get_tests_output_path
|
||||
from tests import get_tests_path, get_tests_input_path, get_tests_output_path
|
||||
from TTS.utils.audio import AudioProcessor
|
||||
from TTS.utils.io import load_config
|
||||
|
||||
@@ -10,7 +10,7 @@ OUT_PATH = os.path.join(get_tests_output_path(), "audio_tests")
|
||||
WAV_FILE = os.path.join(get_tests_input_path(), "example_1.wav")
|
||||
|
||||
os.makedirs(OUT_PATH, exist_ok=True)
|
||||
conf = load_config(os.path.join(TESTS_PATH, 'test_config.json'))
|
||||
conf = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
|
||||
# pylint: disable=protected-access
|
||||
|
||||
@@ -4,10 +4,11 @@ import unittest
|
||||
import torch as T
|
||||
|
||||
from TTS.server.synthesizer import Synthesizer
|
||||
from TTS.tests import get_tests_input_path, get_tests_output_path
|
||||
from TTS.utils.text.symbols import make_symbols, phonemes, symbols
|
||||
from TTS.utils.generic_utils import setup_model
|
||||
from TTS.utils.io import load_config, save_checkpoint
|
||||
from tests import get_tests_input_path, get_tests_output_path
|
||||
from TTS.tts.utils.text.symbols import make_symbols, phonemes, symbols
|
||||
from TTS.tts.utils.generic_utils import setup_model
|
||||
from TTS.tts.utils.io import save_checkpoint
|
||||
from TTS.utils.io import load_config
|
||||
|
||||
|
||||
class DemoServerTest(unittest.TestCase):
|
||||
|
||||
@@ -0,0 +1,88 @@
|
||||
import os
|
||||
import unittest
|
||||
import torch as T
|
||||
|
||||
from tests import get_tests_path, get_tests_input_path
|
||||
from TTS.speaker_encoder.model import SpeakerEncoder
|
||||
from TTS.speaker_encoder.loss import GE2ELoss
|
||||
from TTS.utils.io import load_config
|
||||
|
||||
|
||||
file_path = get_tests_input_path()
|
||||
c = load_config(os.path.join(file_path, "test_config.json"))
|
||||
|
||||
|
||||
class SpeakerEncoderTests(unittest.TestCase):
|
||||
# pylint: disable=R0201
|
||||
def test_in_out(self):
|
||||
dummy_input = T.rand(4, 20, 80) # B x T x D
|
||||
dummy_hidden = [T.rand(2, 4, 128), T.rand(2, 4, 128)]
|
||||
model = SpeakerEncoder(
|
||||
input_dim=80, proj_dim=256, lstm_dim=768, num_lstm_layers=3
|
||||
)
|
||||
# computing d vectors
|
||||
output = model.forward(dummy_input)
|
||||
assert output.shape[0] == 4
|
||||
assert output.shape[1] == 256
|
||||
output = model.inference(dummy_input)
|
||||
assert output.shape[0] == 4
|
||||
assert output.shape[1] == 256
|
||||
# compute d vectors by passing LSTM hidden
|
||||
# output = model.forward(dummy_input, dummy_hidden)
|
||||
# assert output.shape[0] == 4
|
||||
# assert output.shape[1] == 20
|
||||
# assert output.shape[2] == 256
|
||||
# check normalization
|
||||
output_norm = T.nn.functional.normalize(output, dim=1, p=2)
|
||||
assert_diff = (output_norm - output).sum().item()
|
||||
assert output.type() == "torch.FloatTensor"
|
||||
assert (
|
||||
abs(assert_diff) < 1e-4
|
||||
), f" [!] output_norm has wrong values - {assert_diff}"
|
||||
# compute d for a given batch
|
||||
dummy_input = T.rand(1, 240, 80) # B x T x D
|
||||
output = model.compute_embedding(dummy_input, num_frames=160, overlap=0.5)
|
||||
assert output.shape[0] == 1
|
||||
assert output.shape[1] == 256
|
||||
assert len(output.shape) == 2
|
||||
|
||||
|
||||
class GE2ELossTests(unittest.TestCase):
|
||||
# pylint: disable=R0201
|
||||
def test_in_out(self):
|
||||
# check random input
|
||||
dummy_input = T.rand(4, 5, 64) # num_speaker x num_utterance x dim
|
||||
loss = GE2ELoss(loss_method="softmax")
|
||||
output = loss.forward(dummy_input)
|
||||
assert output.item() >= 0.0
|
||||
# check all zeros
|
||||
dummy_input = T.ones(4, 5, 64) # num_speaker x num_utterance x dim
|
||||
loss = GE2ELoss(loss_method="softmax")
|
||||
output = loss.forward(dummy_input)
|
||||
# check speaker loss with orthogonal d-vectors
|
||||
dummy_input = T.empty(3, 64)
|
||||
dummy_input = T.nn.init.orthogonal(dummy_input)
|
||||
dummy_input = T.cat(
|
||||
[
|
||||
dummy_input[0].repeat(5, 1, 1).transpose(0, 1),
|
||||
dummy_input[1].repeat(5, 1, 1).transpose(0, 1),
|
||||
dummy_input[2].repeat(5, 1, 1).transpose(0, 1),
|
||||
]
|
||||
) # num_speaker x num_utterance x dim
|
||||
loss = GE2ELoss(loss_method="softmax")
|
||||
output = loss.forward(dummy_input)
|
||||
assert output.item() < 0.005
|
||||
|
||||
|
||||
# class LoaderTest(unittest.TestCase):
|
||||
# def test_output(self):
|
||||
# items = libri_tts("/home/erogol/Data/Libri-TTS/train-clean-360/")
|
||||
# ap = AudioProcessor(**c['audio'])
|
||||
# dataset = MyDataset(ap, items, 1.6, 64, 10)
|
||||
# loader = DataLoader(dataset, batch_size=32, shuffle=False, num_workers=0, collate_fn=dataset.collate_fn)
|
||||
# count = 0
|
||||
# for mel, spk in loader:
|
||||
# print(mel.shape)
|
||||
# if count == 4:
|
||||
# break
|
||||
# count += 1
|
||||
@@ -1,9 +1,9 @@
|
||||
import unittest
|
||||
import torch as T
|
||||
|
||||
from TTS.layers.tacotron import Prenet, CBHG, Decoder, Encoder
|
||||
from TTS.layers.losses import L1LossMasked
|
||||
from TTS.utils.generic_utils import sequence_mask
|
||||
from TTS.tts.layers.tacotron import Prenet, CBHG, Decoder, Encoder
|
||||
from TTS.tts.layers.losses import L1LossMasked
|
||||
from TTS.tts.utils.generic_utils import sequence_mask
|
||||
|
||||
# pylint: disable=unused-variable
|
||||
|
||||
|
||||
@@ -4,18 +4,18 @@ import shutil
|
||||
import torch
|
||||
import numpy as np
|
||||
|
||||
from tests import get_tests_path, get_tests_input_path, get_tests_output_path
|
||||
from torch.utils.data import DataLoader
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.utils.audio import AudioProcessor
|
||||
from TTS.datasets import TTSDataset
|
||||
from TTS.datasets.preprocess import ljspeech
|
||||
from TTS.tts.datasets import TTSDataset
|
||||
from TTS.tts.datasets.preprocess import ljspeech
|
||||
|
||||
#pylint: disable=unused-variable
|
||||
|
||||
file_path = os.path.dirname(os.path.realpath(__file__))
|
||||
OUTPATH = os.path.join(file_path, "outputs/loader_tests/")
|
||||
OUTPATH = os.path.join(get_tests_output_path(), "loader_tests/")
|
||||
os.makedirs(OUTPATH, exist_ok=True)
|
||||
c = load_config(os.path.join(file_path, 'test_config.json'))
|
||||
c = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
ok_ljspeech = os.path.exists(c.data_path)
|
||||
|
||||
DATA_EXIST = True
|
||||
|
||||
@@ -1,8 +1,8 @@
|
||||
import unittest
|
||||
import os
|
||||
from TTS.tests import get_tests_input_path
|
||||
from tests import get_tests_input_path
|
||||
|
||||
from TTS.datasets.preprocess import common_voice
|
||||
from TTS.tts.datasets.preprocess import common_voice
|
||||
|
||||
|
||||
class TestPreprocessors(unittest.TestCase):
|
||||
|
||||
@@ -1,14 +1,14 @@
|
||||
import os
|
||||
import copy
|
||||
import torch
|
||||
import os
|
||||
import unittest
|
||||
import numpy as np
|
||||
|
||||
from torch import optim
|
||||
from torch import nn
|
||||
import torch
|
||||
from tests import get_tests_input_path
|
||||
from torch import nn, optim
|
||||
|
||||
from TTS.tts.layers.losses import MSELossMasked
|
||||
from TTS.tts.models.tacotron2 import Tacotron2
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.layers.losses import MSELossMasked
|
||||
from TTS.models.tacotron2 import Tacotron2
|
||||
|
||||
#pylint: disable=unused-variable
|
||||
|
||||
@@ -16,8 +16,7 @@ torch.manual_seed(1)
|
||||
use_cuda = torch.cuda.is_available()
|
||||
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
||||
|
||||
file_path = os.path.dirname(os.path.realpath(__file__))
|
||||
c = load_config(os.path.join(file_path, 'test_config.json'))
|
||||
c = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
|
||||
class TacotronTrainTest(unittest.TestCase):
|
||||
|
||||
@@ -0,0 +1,135 @@
|
||||
import os
|
||||
import torch
|
||||
import unittest
|
||||
import numpy as np
|
||||
import tensorflow as tf
|
||||
tf.get_logger().setLevel('INFO')
|
||||
|
||||
from tests import get_tests_path, get_tests_input_path, get_tests_output_path
|
||||
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.tts.tf.models.tacotron2 import Tacotron2
|
||||
from TTS.tts.tf.utils.tflite import convert_tacotron2_to_tflite, load_tflite_model
|
||||
|
||||
#pylint: disable=unused-variable
|
||||
|
||||
torch.manual_seed(1)
|
||||
use_cuda = torch.cuda.is_available()
|
||||
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
||||
|
||||
c = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
|
||||
class TacotronTFTrainTest(unittest.TestCase):
|
||||
|
||||
@staticmethod
|
||||
def generate_dummy_inputs():
|
||||
chars_seq = torch.randint(0, 24, (8, 128)).long().to(device)
|
||||
chars_seq_lengths = torch.randint(100, 128, (8, )).long().to(device)
|
||||
chars_seq_lengths = torch.sort(chars_seq_lengths, descending=True)[0]
|
||||
mel_spec = torch.rand(8, 30, c.audio['num_mels']).to(device)
|
||||
mel_postnet_spec = torch.rand(8, 30, c.audio['num_mels']).to(device)
|
||||
mel_lengths = torch.randint(20, 30, (8, )).long().to(device)
|
||||
stop_targets = torch.zeros(8, 30, 1).float().to(device)
|
||||
speaker_ids = torch.randint(0, 5, (8, )).long().to(device)
|
||||
|
||||
chars_seq = tf.convert_to_tensor(chars_seq.cpu().numpy())
|
||||
chars_seq_lengths = tf.convert_to_tensor(chars_seq_lengths.cpu().numpy())
|
||||
mel_spec = tf.convert_to_tensor(mel_spec.cpu().numpy())
|
||||
return chars_seq, chars_seq_lengths, mel_spec, mel_postnet_spec, mel_lengths,\
|
||||
stop_targets, speaker_ids
|
||||
|
||||
def test_train_step(self):
|
||||
''' test forward pass '''
|
||||
chars_seq, chars_seq_lengths, mel_spec, mel_postnet_spec, mel_lengths,\
|
||||
stop_targets, speaker_ids = self.generate_dummy_inputs()
|
||||
|
||||
for idx in mel_lengths:
|
||||
stop_targets[:, int(idx.item()):, 0] = 1.0
|
||||
|
||||
stop_targets = stop_targets.view(chars_seq.shape[0],
|
||||
stop_targets.size(1) // c.r, -1)
|
||||
stop_targets = (stop_targets.sum(2) > 0.0).unsqueeze(2).float().squeeze()
|
||||
|
||||
model = Tacotron2(num_chars=24, r=c.r, num_speakers=5)
|
||||
# training pass
|
||||
output = model(chars_seq, chars_seq_lengths, mel_spec, training=True)
|
||||
|
||||
# check model output shapes
|
||||
assert np.all(output[0].shape == mel_spec.shape)
|
||||
assert np.all(output[1].shape == mel_spec.shape)
|
||||
assert output[2].shape[2] == chars_seq.shape[1]
|
||||
assert output[2].shape[1] == (mel_spec.shape[1] // model.decoder.r)
|
||||
assert output[3].shape[1] == (mel_spec.shape[1] // model.decoder.r)
|
||||
|
||||
# inference pass
|
||||
output = model(chars_seq, training=False)
|
||||
|
||||
def test_forward_attention(self,):
|
||||
chars_seq, chars_seq_lengths, mel_spec, mel_postnet_spec, mel_lengths,\
|
||||
stop_targets, speaker_ids = self.generate_dummy_inputs()
|
||||
|
||||
for idx in mel_lengths:
|
||||
stop_targets[:, int(idx.item()):, 0] = 1.0
|
||||
|
||||
stop_targets = stop_targets.view(chars_seq.shape[0],
|
||||
stop_targets.size(1) // c.r, -1)
|
||||
stop_targets = (stop_targets.sum(2) > 0.0).unsqueeze(2).float().squeeze()
|
||||
|
||||
model = Tacotron2(num_chars=24, r=c.r, num_speakers=5, forward_attn=True)
|
||||
# training pass
|
||||
output = model(chars_seq, chars_seq_lengths, mel_spec, training=True)
|
||||
|
||||
# check model output shapes
|
||||
assert np.all(output[0].shape == mel_spec.shape)
|
||||
assert np.all(output[1].shape == mel_spec.shape)
|
||||
assert output[2].shape[2] == chars_seq.shape[1]
|
||||
assert output[2].shape[1] == (mel_spec.shape[1] // model.decoder.r)
|
||||
assert output[3].shape[1] == (mel_spec.shape[1] // model.decoder.r)
|
||||
|
||||
# inference pass
|
||||
output = model(chars_seq, training=False)
|
||||
|
||||
def test_tflite_conversion(self, ): #pylint:disable=no-self-use
|
||||
model = Tacotron2(num_chars=24,
|
||||
num_speakers=0,
|
||||
r=3,
|
||||
postnet_output_dim=80,
|
||||
decoder_output_dim=80,
|
||||
attn_type='original',
|
||||
attn_win=False,
|
||||
attn_norm='sigmoid',
|
||||
prenet_type='original',
|
||||
prenet_dropout=True,
|
||||
forward_attn=False,
|
||||
trans_agent=False,
|
||||
forward_attn_mask=False,
|
||||
location_attn=True,
|
||||
attn_K=0,
|
||||
separate_stopnet=True,
|
||||
bidirectional_decoder=False,
|
||||
enable_tflite=True)
|
||||
model.build_inference()
|
||||
convert_tacotron2_to_tflite(model, output_path='test_tacotron2.tflite', experimental_converter=True)
|
||||
# init tflite model
|
||||
tflite_model = load_tflite_model('test_tacotron2.tflite')
|
||||
# fake input
|
||||
inputs = tf.random.uniform([1, 4], maxval=10, dtype=tf.int32) #pylint:disable=unexpected-keyword-arg
|
||||
# run inference
|
||||
# get input and output details
|
||||
input_details = tflite_model.get_input_details()
|
||||
output_details = tflite_model.get_output_details()
|
||||
# reshape input tensor for the new input shape
|
||||
tflite_model.resize_tensor_input(input_details[0]['index'], inputs.shape) #pylint:disable=unexpected-keyword-arg
|
||||
tflite_model.allocate_tensors()
|
||||
detail = input_details[0]
|
||||
input_shape = detail['shape']
|
||||
tflite_model.set_tensor(detail['index'], inputs)
|
||||
# run the tflite_model
|
||||
tflite_model.invoke()
|
||||
# collect outputs
|
||||
decoder_output = tflite_model.get_tensor(output_details[0]['index'])
|
||||
postnet_output = tflite_model.get_tensor(output_details[1]['index'])
|
||||
# remove tflite binary
|
||||
os.remove('test_tacotron2.tflite')
|
||||
|
||||
@@ -1,13 +1,14 @@
|
||||
import os
|
||||
import copy
|
||||
import torch
|
||||
import os
|
||||
import unittest
|
||||
|
||||
from torch import optim
|
||||
from torch import nn
|
||||
import torch
|
||||
from tests import get_tests_input_path
|
||||
from torch import nn, optim
|
||||
|
||||
from TTS.tts.layers.losses import L1LossMasked
|
||||
from TTS.tts.models.tacotron import Tacotron
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.layers.losses import L1LossMasked
|
||||
from TTS.models.tacotron import Tacotron
|
||||
|
||||
#pylint: disable=unused-variable
|
||||
|
||||
@@ -15,8 +16,7 @@ torch.manual_seed(1)
|
||||
use_cuda = torch.cuda.is_available()
|
||||
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
||||
|
||||
file_path = os.path.dirname(os.path.realpath(__file__))
|
||||
c = load_config(os.path.join(file_path, 'test_config.json'))
|
||||
c = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
|
||||
def count_parameters(model):
|
||||
|
||||
@@ -3,12 +3,12 @@ import os
|
||||
# pylint: disable=wildcard-import
|
||||
# pylint: disable=unused-import
|
||||
import unittest
|
||||
from TTS.utils.text import *
|
||||
from TTS.tests import get_tests_path
|
||||
from tests import get_tests_input_path
|
||||
from TTS.tts.utils.text import *
|
||||
from tests import get_tests_path
|
||||
from TTS.utils.io import load_config
|
||||
|
||||
TESTS_PATH = get_tests_path()
|
||||
conf = load_config(os.path.join(TESTS_PATH, 'test_config.json'))
|
||||
conf = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
def test_phoneme_to_sequence():
|
||||
text = "Recent research at Harvard has shown meditating for as little as 8 weeks can actually increase, the grey matter in the parts of the brain responsible for emotional regulation and learning!"
|
||||
@@ -19,7 +19,7 @@ def test_phoneme_to_sequence():
|
||||
sequence_with_params = phoneme_to_sequence(text, text_cleaner, lang, tp=conf.characters)
|
||||
text_hat_with_params = sequence_to_phoneme(sequence, tp=conf.characters)
|
||||
gt = "ɹiːsənt ɹɪsɜːtʃ æt hɑːɹvɚd hɐz ʃoʊn mɛdᵻteɪɾɪŋ fɔːɹ æz lɪɾəl æz eɪt wiːks kæn æktʃuːəli ɪnkɹiːs, ðə ɡɹeɪ mæɾɚɹ ɪnðə pɑːɹts ʌvðə bɹeɪn ɹɪspɑːnsəbəl fɔːɹ ɪmoʊʃənəl ɹɛɡjuːleɪʃən ænd lɜːnɪŋ!"
|
||||
assert text_hat == text_hat_with_params == gt
|
||||
assert text_hat == text_hat_with_params == gt
|
||||
|
||||
# multiple punctuations
|
||||
text = "Be a voice, not an! echo?"
|
||||
|
||||
Executable
+13
@@ -0,0 +1,13 @@
|
||||
#!/usr/bin/env bash
|
||||
|
||||
BASEDIR=$(dirname "$0")
|
||||
echo "$BASEDIR"
|
||||
# run training
|
||||
CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tts.py --config_path $BASEDIR/inputs/test_train_config.json
|
||||
# find the training folder
|
||||
LATEST_FOLDER=$(ls $BASEDIR/train_outputs/| sort | tail -1)
|
||||
echo $LATEST_FOLDER
|
||||
# continue the previous training
|
||||
CUDA_VISIBLE_DEVICES="" python TTS/bin/train_tts.py --continue_path $BASEDIR/train_outputs/$LATEST_FOLDER
|
||||
# remove all the outputs
|
||||
rm -rf $BASEDIR/train_outputs/
|
||||
@@ -0,0 +1,95 @@
|
||||
import os
|
||||
|
||||
import numpy as np
|
||||
from tests import get_tests_path, get_tests_input_path, get_tests_output_path
|
||||
from torch.utils.data import DataLoader
|
||||
|
||||
from TTS.utils.audio import AudioProcessor
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.vocoder.datasets.gan_dataset import GANDataset
|
||||
from TTS.vocoder.datasets.preprocess import load_wav_data
|
||||
|
||||
file_path = os.path.dirname(os.path.realpath(__file__))
|
||||
OUTPATH = os.path.join(get_tests_output_path(), "loader_tests/")
|
||||
os.makedirs(OUTPATH, exist_ok=True)
|
||||
|
||||
C = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
|
||||
test_data_path = os.path.join(get_tests_path(), "data/ljspeech/")
|
||||
ok_ljspeech = os.path.exists(test_data_path)
|
||||
|
||||
|
||||
def gan_dataset_case(batch_size, seq_len, hop_len, conv_pad, return_segments, use_noise_augment, use_cache, num_workers):
|
||||
''' run dataloader with given parameters and check conditions '''
|
||||
ap = AudioProcessor(**C.audio)
|
||||
_, train_items = load_wav_data(test_data_path, 10)
|
||||
dataset = GANDataset(ap,
|
||||
train_items,
|
||||
seq_len=seq_len,
|
||||
hop_len=hop_len,
|
||||
pad_short=2000,
|
||||
conv_pad=conv_pad,
|
||||
return_segments=return_segments,
|
||||
use_noise_augment=use_noise_augment,
|
||||
use_cache=use_cache)
|
||||
loader = DataLoader(dataset=dataset,
|
||||
batch_size=batch_size,
|
||||
shuffle=True,
|
||||
num_workers=num_workers,
|
||||
pin_memory=True,
|
||||
drop_last=True)
|
||||
|
||||
max_iter = 10
|
||||
count_iter = 0
|
||||
|
||||
# return random segments or return the whole audio
|
||||
if return_segments:
|
||||
for item1, _ in loader:
|
||||
feat1, wav1 = item1
|
||||
# feat2, wav2 = item2
|
||||
expected_feat_shape = (batch_size, ap.num_mels, seq_len // hop_len + conv_pad * 2)
|
||||
|
||||
# check shapes
|
||||
assert np.all(feat1.shape == expected_feat_shape), f" [!] {feat1.shape} vs {expected_feat_shape}"
|
||||
assert (feat1.shape[2] - conv_pad * 2) * hop_len == wav1.shape[2]
|
||||
|
||||
# check feature vs audio match
|
||||
if not use_noise_augment:
|
||||
for idx in range(batch_size):
|
||||
audio = wav1[idx].squeeze()
|
||||
feat = feat1[idx]
|
||||
mel = ap.melspectrogram(audio)
|
||||
# the first 2 and the last 2 frames are skipped due to the padding
|
||||
# differences in stft
|
||||
assert (feat - mel[:, :feat1.shape[-1]])[:, 2:-2].sum() <= 0, f' [!] {(feat - mel[:, :feat1.shape[-1]])[:, 2:-2].sum()}'
|
||||
|
||||
count_iter += 1
|
||||
# if count_iter == max_iter:
|
||||
# break
|
||||
else:
|
||||
for item in loader:
|
||||
feat, wav = item
|
||||
expected_feat_shape = (batch_size, ap.num_mels, (wav.shape[-1] // hop_len) + (conv_pad * 2))
|
||||
assert np.all(feat.shape == expected_feat_shape), f" [!] {feat.shape} vs {expected_feat_shape}"
|
||||
assert (feat.shape[2] - conv_pad * 2) * hop_len == wav.shape[2]
|
||||
count_iter += 1
|
||||
if count_iter == max_iter:
|
||||
break
|
||||
|
||||
|
||||
def test_parametrized_gan_dataset():
|
||||
''' test dataloader with different parameters '''
|
||||
params = [
|
||||
[32, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, True, False, True, 0],
|
||||
[32, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, True, False, True, 4],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, True, True, True, 0],
|
||||
[1, C.audio['hop_length'], C.audio['hop_length'], 0, True, True, True, 0],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 2, True, True, True, 0],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, False, True, True, 0],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, True, False, True, 0],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, True, True, False, 0],
|
||||
[1, C.audio['hop_length'] * 10, C.audio['hop_length'], 0, False, False, False, 0],
|
||||
]
|
||||
for param in params:
|
||||
print(param)
|
||||
gan_dataset_case(*param)
|
||||
@@ -0,0 +1,54 @@
|
||||
import os
|
||||
|
||||
import torch
|
||||
from tests import get_tests_input_path, get_tests_output_path, get_tests_path
|
||||
|
||||
from TTS.utils.audio import AudioProcessor
|
||||
from TTS.utils.io import load_config
|
||||
from TTS.vocoder.layers.losses import MultiScaleSTFTLoss, STFTLoss, TorchSTFT
|
||||
|
||||
TESTS_PATH = get_tests_path()
|
||||
|
||||
OUT_PATH = os.path.join(get_tests_output_path(), "audio_tests")
|
||||
os.makedirs(OUT_PATH, exist_ok=True)
|
||||
|
||||
WAV_FILE = os.path.join(get_tests_input_path(), "example_1.wav")
|
||||
|
||||
C = load_config(os.path.join(get_tests_input_path(), 'test_config.json'))
|
||||
ap = AudioProcessor(**C.audio)
|
||||
|
||||
|
||||
def test_torch_stft():
|
||||
torch_stft = TorchSTFT(ap.fft_size, ap.hop_length, ap.win_length)
|
||||
# librosa stft
|
||||
wav = ap.load_wav(WAV_FILE)
|
||||
M_librosa = abs(ap._stft(wav)) # pylint: disable=protected-access
|
||||
# torch stft
|
||||
wav = torch.from_numpy(wav[None, :]).float()
|
||||
M_torch = torch_stft(wav)
|
||||
# check the difference b/w librosa and torch outputs
|
||||
assert (M_librosa - M_torch[0].data.numpy()).max() < 1e-5
|
||||
|
||||
|
||||
def test_stft_loss():
|
||||
stft_loss = STFTLoss(ap.fft_size, ap.hop_length, ap.win_length)
|
||||
wav = ap.load_wav(WAV_FILE)
|
||||
wav = torch.from_numpy(wav[None, :]).float()
|
||||
loss_m, loss_sc = stft_loss(wav, wav)
|
||||
assert loss_m + loss_sc == 0
|
||||
loss_m, loss_sc = stft_loss(wav, torch.rand_like(wav))
|
||||
assert loss_sc < 1.0
|
||||
assert loss_m + loss_sc > 0
|
||||
|
||||
|
||||
def test_multiscale_stft_loss():
|
||||
stft_loss = MultiScaleSTFTLoss([ap.fft_size//2, ap.fft_size, ap.fft_size*2],
|
||||
[ap.hop_length // 2, ap.hop_length, ap.hop_length * 2],
|
||||
[ap.win_length // 2, ap.win_length, ap.win_length * 2])
|
||||
wav = ap.load_wav(WAV_FILE)
|
||||
wav = torch.from_numpy(wav[None, :]).float()
|
||||
loss_m, loss_sc = stft_loss(wav, wav)
|
||||
assert loss_m + loss_sc == 0
|
||||
loss_m, loss_sc = stft_loss(wav, torch.rand_like(wav))
|
||||
assert loss_sc < 1.0
|
||||
assert loss_m + loss_sc > 0
|
||||
@@ -0,0 +1,26 @@
|
||||
import numpy as np
|
||||
import torch
|
||||
|
||||
from TTS.vocoder.models.melgan_discriminator import MelganDiscriminator
|
||||
from TTS.vocoder.models.melgan_multiscale_discriminator import MelganMultiscaleDiscriminator
|
||||
|
||||
|
||||
def test_melgan_discriminator():
|
||||
model = MelganDiscriminator()
|
||||
print(model)
|
||||
dummy_input = torch.rand((4, 1, 256 * 10))
|
||||
output, _ = model(dummy_input)
|
||||
assert np.all(output.shape == (4, 1, 10))
|
||||
|
||||
|
||||
def test_melgan_multi_scale_discriminator():
|
||||
model = MelganMultiscaleDiscriminator()
|
||||
print(model)
|
||||
dummy_input = torch.rand((4, 1, 256 * 16))
|
||||
scores, feats = model(dummy_input)
|
||||
assert len(scores) == 3
|
||||
assert len(scores) == len(feats)
|
||||
assert np.all(scores[0].shape == (4, 1, 64))
|
||||
assert np.all(feats[0][0].shape == (4, 16, 4096))
|
||||
assert np.all(feats[0][1].shape == (4, 64, 1024))
|
||||
assert np.all(feats[0][2].shape == (4, 256, 256))
|
||||
@@ -0,0 +1,14 @@
|
||||
import numpy as np
|
||||
import torch
|
||||
|
||||
from TTS.vocoder.models.melgan_generator import MelganGenerator
|
||||
|
||||
def test_melgan_generator():
|
||||
model = MelganGenerator()
|
||||
print(model)
|
||||
dummy_input = torch.rand((4, 80, 64))
|
||||
output = model(dummy_input)
|
||||
assert np.all(output.shape == (4, 1, 64 * 256))
|
||||
output = model.inference(dummy_input)
|
||||
assert np.all(output.shape == (4, 1, (64 + 4) * 256))
|
||||
|
||||
@@ -0,0 +1,28 @@
|
||||
import os
|
||||
import torch
|
||||
|
||||
import soundfile as sf
|
||||
from librosa.core import load
|
||||
|
||||
from tests import get_tests_path, get_tests_input_path
|
||||
from TTS.vocoder.layers.pqmf import PQMF
|
||||
|
||||
|
||||
TESTS_PATH = get_tests_path()
|
||||
WAV_FILE = os.path.join(get_tests_input_path(), "example_1.wav")
|
||||
|
||||
|
||||
def test_pqmf():
|
||||
w, sr = load(WAV_FILE)
|
||||
|
||||
layer = PQMF(N=4, taps=62, cutoff=0.15, beta=9.0)
|
||||
w, sr = load(WAV_FILE)
|
||||
w2 = torch.from_numpy(w[None, None, :])
|
||||
b2 = layer.analysis(w2)
|
||||
w2_ = layer.synthesis(b2)
|
||||
|
||||
print(w2_.max())
|
||||
print(w2_.min())
|
||||
print(w2_.mean())
|
||||
sf.write('pqmf_output.wav', w2_.flatten().detach(), sr)
|
||||
|
||||
@@ -0,0 +1,21 @@
|
||||
import torch
|
||||
import numpy as np
|
||||
|
||||
from TTS.vocoder.models.random_window_discriminator import RandomWindowDiscriminator
|
||||
|
||||
|
||||
def test_rwd():
|
||||
layer = RandomWindowDiscriminator(cond_channels=80,
|
||||
window_sizes=(512, 1024, 2048, 4096,
|
||||
8192),
|
||||
cond_disc_downsample_factors=[
|
||||
(8, 4, 2, 2, 2), (8, 4, 2, 2),
|
||||
(8, 4, 2), (8, 4), (4, 2, 2)
|
||||
],
|
||||
hop_length=256)
|
||||
x = torch.rand([4, 1, 22050])
|
||||
c = torch.rand([4, 80, 22050 // 256])
|
||||
|
||||
scores, _ = layer(x, c)
|
||||
assert len(scores) == 10
|
||||
assert np.all(scores[0].shape == (4, 1, 1))
|
||||
Executable
+15
@@ -0,0 +1,15 @@
|
||||
#!/usr/bin/env bash
|
||||
|
||||
BASEDIR=$(dirname "$0")
|
||||
echo "$BASEDIR"
|
||||
# create run dir
|
||||
mkdir $BASEDIR/train_outputs
|
||||
# run training
|
||||
CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder.py --config_path $BASEDIR/inputs/test_vocoder_multiband_melgan_config.json
|
||||
# find the training folder
|
||||
LATEST_FOLDER=$(ls $BASEDIR/outputs/train_outputs/| sort | tail -1)
|
||||
echo $LATEST_FOLDER
|
||||
# continue the previous training
|
||||
CUDA_VISIBLE_DEVICES="" python TTS/bin/train_vocoder.py --continue_path $BASEDIR/outputs/train_outputs/$LATEST_FOLDER
|
||||
# remove all the outputs
|
||||
rm -rf $BASEDIR/train_outputs/
|
||||
Reference in New Issue
Block a user