From 27b67b7945ca7cd8c7ad1ce60744892a4eb47716 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Eren=20G=C3=B6lge?= Date: Wed, 2 Mar 2022 09:15:20 +0100 Subject: [PATCH] Fix import --- TTS/tts/configs/tacotron_config.py | 3 +++ TTS/tts/models/forward_tts.py | 2 +- tests/data_tests/test_loader.py | 6 +++--- tests/tts_tests/test_align_tts_train.py | 2 +- tests/tts_tests/test_fast_pitch_speaker_emb_train.py | 4 ++-- tests/tts_tests/test_fast_pitch_train.py | 2 +- tests/tts_tests/test_glow_tts_d-vectors_train.py | 3 +-- tests/tts_tests/test_glow_tts_speaker_emb_train.py | 3 +-- tests/tts_tests/test_glow_tts_train.py | 2 +- tests/tts_tests/test_speedy_speech_train.py | 2 +- tests/tts_tests/test_tacotron2_d-vectors_train.py | 2 +- tests/tts_tests/test_tacotron2_speaker_emb_train.py | 3 ++- tests/tts_tests/test_tacotron2_train.py | 2 +- tests/tts_tests/test_tacotron_train.py | 2 +- tests/tts_tests/test_vits_multilingual_speaker_emb_train.py | 2 +- tests/tts_tests/test_vits_multilingual_train-d_vectors.py | 2 +- tests/tts_tests/test_vits_speaker_emb_train.py | 2 +- tests/tts_tests/test_vits_train.py | 2 +- 18 files changed, 24 insertions(+), 22 deletions(-) diff --git a/TTS/tts/configs/tacotron_config.py b/TTS/tts/configs/tacotron_config.py index d6edd267..5193c224 100644 --- a/TTS/tts/configs/tacotron_config.py +++ b/TTS/tts/configs/tacotron_config.py @@ -83,6 +83,8 @@ class TacotronConfig(BaseTTSConfig): ddc_r (int): reduction rate used by the coarse decoder when `double_decoder_consistency` is in use. Set this as a multiple of the `r` value. Defaults to 6. + speakers_file (str): + Path to the speaker mapping file for the Speaker Manager. Defaults to None. use_speaker_embedding (bool): enable / disable using speaker embeddings for multi-speaker models. If set True, the model is in the multi-speaker mode. Defaults to False. @@ -176,6 +178,7 @@ class TacotronConfig(BaseTTSConfig): ddc_r: int = 6 # multi-speaker settings + speakers_file: str = None use_speaker_embedding: bool = False speaker_embedding_dim: int = 512 use_d_vector_file: bool = False diff --git a/TTS/tts/models/forward_tts.py b/TTS/tts/models/forward_tts.py index db8fef2d..a1273f7f 100644 --- a/TTS/tts/models/forward_tts.py +++ b/TTS/tts/models/forward_tts.py @@ -261,7 +261,7 @@ class ForwardTTS(BaseTTS): # init speaker embedding layer if config.use_speaker_embedding and not config.use_d_vector_file: print(" > Init speaker_embedding layer.") - self.emb_g = nn.Embedding(self.args.num_speakers, self.args.hidden_channels) + self.emb_g = nn.Embedding(self.num_speakers, self.args.hidden_channels) nn.init.uniform_(self.emb_g.weight, -0.1, 0.1) @staticmethod diff --git a/tests/data_tests/test_loader.py b/tests/data_tests/test_loader.py index 4d8cc68a..2727bbdd 100644 --- a/tests/data_tests/test_loader.py +++ b/tests/data_tests/test_loader.py @@ -44,13 +44,13 @@ class TestTTSDataset(unittest.TestCase): self.max_loader_iter = 4 self.ap = AudioProcessor(**c.audio) - def _create_dataloader(self, batch_size, r, bgs): + def _create_dataloader(self, batch_size, r, bgs, start_by_longest=False): # load dataset meta_data_train, meta_data_eval = load_tts_samples(dataset_config, eval_split=True, eval_split_size=0.2) items = meta_data_train + meta_data_eval - tokenizer = TTSTokenizer.init_from_config(c) + tokenizer, _ = TTSTokenizer.init_from_config(c) dataset = TTSDataset( outputs_per_step=r, compute_linear_spec=True, @@ -77,7 +77,7 @@ class TestTTSDataset(unittest.TestCase): def test_loader(self): if ok_ljspeech: - dataloader, dataset = self._create_dataloader(2, c.r, 0) + dataloader, dataset = self._create_dataloader(1, 1, 0) for i, data in enumerate(dataloader): if i == self.max_loader_iter: diff --git a/tests/tts_tests/test_align_tts_train.py b/tests/tts_tests/test_align_tts_train.py index d5115af6..6c68d8c9 100644 --- a/tests/tts_tests/test_align_tts_train.py +++ b/tests/tts_tests/test_align_tts_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.align_tts_config import AlignTTSConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_fast_pitch_speaker_emb_train.py b/tests/tts_tests/test_fast_pitch_speaker_emb_train.py index 59e90e0a..88505988 100644 --- a/tests/tts_tests/test_fast_pitch_speaker_emb_train.py +++ b/tests/tts_tests/test_fast_pitch_speaker_emb_train.py @@ -5,9 +5,9 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.config.shared_configs import BaseAudioConfig from TTS.tts.configs.fast_pitch_config import FastPitchConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint -config_path = os.path.join(get_tests_output_path(), "test_model_config.json") +config_path = os.path.join(get_tests_output_path(), "fast_pitch_speaker_emb_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") audio_config = BaseAudioConfig( diff --git a/tests/tts_tests/test_fast_pitch_train.py b/tests/tts_tests/test_fast_pitch_train.py index bbfbb823..5a51f0bb 100644 --- a/tests/tts_tests/test_fast_pitch_train.py +++ b/tests/tts_tests/test_fast_pitch_train.py @@ -5,7 +5,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.config.shared_configs import BaseAudioConfig from TTS.tts.configs.fast_pitch_config import FastPitchConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_glow_tts_d-vectors_train.py b/tests/tts_tests/test_glow_tts_d-vectors_train.py index c85e6bcd..dd5e954e 100644 --- a/tests/tts_tests/test_glow_tts_d-vectors_train.py +++ b/tests/tts_tests/test_glow_tts_d-vectors_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.glow_tts_config import GlowTTSConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") @@ -17,7 +17,6 @@ config = GlowTTSConfig( num_eval_loader_workers=0, text_cleaner="english_cleaners", use_phonemes=True, - use_espeak_phonemes=True, phoneme_language="en-us", phoneme_cache_path="tests/data/ljspeech/phoneme_cache/", run_eval=True, diff --git a/tests/tts_tests/test_glow_tts_speaker_emb_train.py b/tests/tts_tests/test_glow_tts_speaker_emb_train.py index 7e6aabde..df86cf05 100644 --- a/tests/tts_tests/test_glow_tts_speaker_emb_train.py +++ b/tests/tts_tests/test_glow_tts_speaker_emb_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.glow_tts_config import GlowTTSConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") @@ -17,7 +17,6 @@ config = GlowTTSConfig( num_eval_loader_workers=0, text_cleaner="english_cleaners", use_phonemes=True, - use_espeak_phonemes=True, phoneme_language="en-us", phoneme_cache_path="tests/data/ljspeech/phoneme_cache/", run_eval=True, diff --git a/tests/tts_tests/test_glow_tts_train.py b/tests/tts_tests/test_glow_tts_train.py index e5dc44ee..3a1c4a68 100644 --- a/tests/tts_tests/test_glow_tts_train.py +++ b/tests/tts_tests/test_glow_tts_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.glow_tts_config import GlowTTSConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_speedy_speech_train.py b/tests/tts_tests/test_speedy_speech_train.py index 7e938a40..98cf8e09 100644 --- a/tests/tts_tests/test_speedy_speech_train.py +++ b/tests/tts_tests/test_speedy_speech_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.speedy_speech_config import SpeedySpeechConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_speedy_speech_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_tacotron2_d-vectors_train.py b/tests/tts_tests/test_tacotron2_d-vectors_train.py index 0bc31449..e5f83804 100644 --- a/tests/tts_tests/test_tacotron2_d-vectors_train.py +++ b/tests/tts_tests/test_tacotron2_d-vectors_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.tacotron2_config import Tacotron2Config -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_tacotron2_speaker_emb_train.py b/tests/tts_tests/test_tacotron2_speaker_emb_train.py index 653933dd..2dd50c73 100644 --- a/tests/tts_tests/test_tacotron2_speaker_emb_train.py +++ b/tests/tts_tests/test_tacotron2_speaker_emb_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.tacotron2_config import Tacotron2Config -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") @@ -28,6 +28,7 @@ config = Tacotron2Config( "Be a voice, not an echo.", ], use_speaker_embedding=True, + num_speakers=4, max_decoder_steps=50, ) diff --git a/tests/tts_tests/test_tacotron2_train.py b/tests/tts_tests/test_tacotron2_train.py index 76727edf..a45065b2 100644 --- a/tests/tts_tests/test_tacotron2_train.py +++ b/tests/tts_tests/test_tacotron2_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.tacotron2_config import Tacotron2Config -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_tacotron_train.py b/tests/tts_tests/test_tacotron_train.py index 02491e64..96c63162 100644 --- a/tests/tts_tests/test_tacotron_train.py +++ b/tests/tts_tests/test_tacotron_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.tacotron_config import TacotronConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_vits_multilingual_speaker_emb_train.py b/tests/tts_tests/test_vits_multilingual_speaker_emb_train.py index afa60a1b..c09f8498 100644 --- a/tests/tts_tests/test_vits_multilingual_speaker_emb_train.py +++ b/tests/tts_tests/test_vits_multilingual_speaker_emb_train.py @@ -5,7 +5,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.config.shared_configs import BaseDatasetConfig from TTS.tts.configs.vits_config import VitsConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_vits_multilingual_train-d_vectors.py b/tests/tts_tests/test_vits_multilingual_train-d_vectors.py index b0744103..8607a8f7 100644 --- a/tests/tts_tests/test_vits_multilingual_train-d_vectors.py +++ b/tests/tts_tests/test_vits_multilingual_train-d_vectors.py @@ -5,7 +5,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.config.shared_configs import BaseDatasetConfig from TTS.tts.configs.vits_config import VitsConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_vits_speaker_emb_train.py b/tests/tts_tests/test_vits_speaker_emb_train.py index 1aecc596..8a586076 100644 --- a/tests/tts_tests/test_vits_speaker_emb_train.py +++ b/tests/tts_tests/test_vits_speaker_emb_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.vits_config import VitsConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs") diff --git a/tests/tts_tests/test_vits_train.py b/tests/tts_tests/test_vits_train.py index ec9a5915..76c88682 100644 --- a/tests/tts_tests/test_vits_train.py +++ b/tests/tts_tests/test_vits_train.py @@ -4,7 +4,7 @@ import shutil from tests import get_device_id, get_tests_output_path, run_cli from TTS.tts.configs.vits_config import VitsConfig -from TTS.utils.trainer_utils import get_last_checkpoint +from trainer import get_last_checkpoint config_path = os.path.join(get_tests_output_path(), "test_model_config.json") output_path = os.path.join(get_tests_output_path(), "train_outputs")