Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix the Bug in Synthesizer

This commit is contained in:
Edresson Casanova 2022-03-30 15:32:35 -03:00
parent e409f3588b
commit a8292c7c03
2 changed files with 17 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
TTS/bin/compute_embeddings.py
View File

@ -86,6 +86,10 @@ for idx, wav_file in enumerate(tqdm(wav_files)):
class_mapping[wav_file_name]["name"] = class_name class_mapping[wav_file_name]["name"] = class_name
class_mapping[wav_file_name]["embedding"] = embedd class_mapping[wav_file_name]["embedding"] = embedd
if args.old_file:
# merge the embeddings dict
class_mapping = {**encoder_manager.embeddings, **class_mapping}
if class_mapping: if class_mapping:
# save class_mapping if target dataset is defined # save class_mapping if target dataset is defined
if ".json" not in args.output_path or ".pth" not in args.output_path: if ".json" not in args.output_path or ".pth" not in args.output_path:

14
TTS/utils/synthesizer.py
View File

@ -123,9 +123,21 @@ class Synthesizer(object):
if use_cuda: if use_cuda:
self.tts_model.cuda() self.tts_model.cuda()
if self.encoder_checkpoint and hasattr(self.tts_model, "speaker_manager"): if self.encoder_checkpoint and hasattr(self.tts_model, "speaker_manager") and self.tts_model.speaker_manager is not None:
self.tts_model.speaker_manager.init_encoder(self.encoder_checkpoint, self.encoder_config, use_cuda) self.tts_model.speaker_manager.init_encoder(self.encoder_checkpoint, self.encoder_config, use_cuda)
if self.tts_emotions_file and hasattr(self.tts_model, "emotion_manager") and self.tts_model.emotion_manager is not None:
if getattr(self.tts_config, "use_external_emotions_embeddings", False) or (getattr(self.tts_config, "model_args", None) and getattr(self.tts_config.model_args, "use_external_emotions_embeddings", False)):
self.tts_model.emotion_manager.load_embeddings_from_file(self.tts_emotions_file)
else:
self.tts_model.emotion_manager.load_ids_from_file(self.tts_emotions_file)
if self.tts_speakers_file and hasattr(self.tts_model, "speaker_manager") and self.tts_model.speaker_manager is not None:
if getattr(self.tts_config, "use_d_vector_file", False) or (getattr(self.tts_config, "model_args", None) and getattr(self.tts_config.model_args, "use_d_vector_file", False)):
self.tts_model.speaker_manager.load_embeddings_from_file(self.tts_speakers_file)
else:
self.tts_model.speaker_manager.load_ids_from_file(self.tts_speakers_file)
def _set_speaker_encoder_paths_from_tts_config(self): def _set_speaker_encoder_paths_from_tts_config(self):
"""Set the encoder paths from the tts model config for models with speaker encoders.""" """Set the encoder paths from the tts model config for models with speaker encoders."""
if hasattr(self.tts_config, "model_args") and hasattr( if hasattr(self.tts_config, "model_args") and hasattr(