ttslearn.train_util

Specific utilities for training

Dataset

class ttslearn.train_util.Dataset(in_paths, out_paths)

Dataset for numpy files

Parameters

• in_paths (list) – List of paths to input files

• out_paths (list) – List of paths to output files

__getitem__(idx)

Get a pair of input and target

Parameters

idx (int) – index of the pair

Returns

input and target in numpy format

Return type

tuple

__len__()

Returns the size of the dataset

Returns

size of the dataset

Return type

int

DataLoader

collate_fn_dnntts	Collate function for DNN-TTS.
collate_fn_wavenet	Collate function for WaveNet.
collate_fn_tacotron	Collate function for Tacotron.
get_data_loaders	Get data loaders for training and validation.

Model parameters

moving_average_	Exponential moving average (EMA) of model parameters.
num_trainable_params	Count the number of trainable parameters in the model.

Helper for training

setup	Setup for traiining
save_checkpoint	Save a checkpoint.
num_trainable_params	Count the number of trainable parameters in the model.
get_epochs_with_optional_tqdm	Get epochs with optional progress bar.

Plotting

plot_attention	Plot attention.
plot_2d_feats	Plot 2D features.