From cfaa51fddc19c7ac8b9c4607494fd9270eddd60c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Eren=20G=C3=B6lge?= Date: Tue, 7 Dec 2021 12:57:51 +0000 Subject: [PATCH] Update BaseTTS config --- TTS/tts/configs/shared_configs.py | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/TTS/tts/configs/shared_configs.py b/TTS/tts/configs/shared_configs.py index 98461bdd..ad3bbe70 100644 --- a/TTS/tts/configs/shared_configs.py +++ b/TTS/tts/configs/shared_configs.py @@ -78,7 +78,7 @@ class CharactersConfig(Coqpit): is_unique (bool): remove any duplicate characters in the character lists. It is a bandaid for compatibility with the old - models trained with character lists with duplicates. + models trained with character lists with duplicates. Defaults to True. is_sorted (bool): Sort the characters in alphabetical order. Defaults to True. @@ -166,6 +166,9 @@ class BaseTTSConfig(BaseTrainingConfig): compute_linear_spec (bool): If True data loader computes and returns linear spectrograms alongside the other data. + precompute_num_workers (int): + Number of workers to precompute features. Defaults to 0. + use_noise_augment (bool): Augment the input audio with random noise. @@ -214,6 +217,7 @@ class BaseTTSConfig(BaseTrainingConfig): phoneme_cache_path: str = None # vocabulary parameters characters: CharactersConfig = None + add_blank: bool = False # training params batch_group_size: int = 0 loss_masking: bool = None @@ -225,8 +229,8 @@ class BaseTTSConfig(BaseTrainingConfig): max_text_len: int = float("inf") compute_f0: bool = False compute_linear_spec: bool = False + precompute_num_workers: int = 0 use_noise_augment: bool = False - add_blank: bool = False # dataset datasets: List[BaseDatasetConfig] = field(default_factory=lambda: [BaseDatasetConfig()]) # optimizer