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
3,624 Commits 80 Branches 98 Tags 314 MiB
Commit Graph

591 Commits

Author SHA1 Message Date
Eren Gölge 1ddf245b08 Use speaker_encoder from speaker manager in Vits 2021-12-16 14:56:34 +00:00
Eren Gölge 6d7199d559 Rename setup_model to setup_speaker_encoder_model 2021-12-13 16:28:54 +00:00
Eren Gölge bbea9b3f9f Remove redundant code 2021-12-10 07:53:19 +00:00
Eren Gölge 66b6e9bc99 Make style 2021-12-10 07:53:10 +00:00
WeberJulian 4706583452 Add support for multi-lingual models in CLI 2021-12-09 13:43:08 +00:00
WeberJulian 3f3505c1ca Prevent weighted sampler use when num_gpus > 1 2021-12-09 13:42:42 +00:00
WeberJulian 0f64d45e04 Revert init multispeaker change 2021-12-09 13:42:42 +00:00
WeberJulian 4001322e50 Fix trailing space 2021-12-09 13:42:42 +00:00
WeberJulian 352b4be104 Move multilingual logic out of the trainer 2021-12-09 13:42:42 +00:00
Edresson be8f444636 Add the SCL resample TODO 2021-12-09 13:41:56 +00:00
WeberJulian eff0a5ca10 Fix merge bug 2021-12-09 13:41:56 +00:00
WeberJulian 9c1bec86a4 Fix tests 2021-12-09 13:41:28 +00:00
Edresson 79f75924de Fix pylint checks 2021-12-09 13:41:28 +00:00
WeberJulian 93dbb67c52 Remove self.audio_config from VITS 2021-12-09 13:41:28 +00:00
Edresson 0359cab4fa Remove torchaudio requeriment 2021-12-09 13:39:13 +00:00
WeberJulian 36ddf32972 Fix trailing whitespace 2021-12-09 13:38:18 +00:00
WeberJulian 9d8d4e6fb3 Update docstring 2021-12-09 13:38:18 +00:00
Edresson 87059e3bbb Add the language embedding dim in the duration predictor class 2021-12-09 13:38:18 +00:00
Edresson aa1a070d58 Rename ununsed_speakers to ignored_speakers 2021-12-09 13:38:18 +00:00
Edresson 1251d04387 Fix function name 2021-12-09 13:37:50 +00:00
Edresson 9781e4d516 Lint fixs 2021-12-09 13:37:50 +00:00
Edresson cad82a9296 Remove the data from the set_d_vectors_from_file function 2021-12-09 13:37:50 +00:00
Edresson ec31dacbb7 Remove unusable speaker manager function 2021-12-09 13:37:50 +00:00
Edresson 86b2536491 Turn more clear the VITS loss function 2021-12-09 13:37:50 +00:00
Edresson 5fc127bb7a Remove the unusable fine-tuning model 2021-12-09 13:37:50 +00:00
WeberJulian 390096fe0f fix d-vector 2021-12-09 13:36:48 +00:00
WeberJulian 868cf6424f Fix small issues 2021-12-09 13:36:48 +00:00
WeberJulian e04577575e Fix use_speaker_embedding logic 2021-12-09 13:36:48 +00:00
WeberJulian 61251bd86c Fix phonemes 2021-12-09 13:36:48 +00:00
WeberJulian 686c7381e2 fix phonemes per language 2021-12-09 13:36:48 +00:00
WeberJulian 215a74b32e fix linter 2021-12-09 13:36:48 +00:00
WeberJulian 3e9ca4b95d make style 2021-12-09 13:36:48 +00:00
WeberJulian 88d6399e12 fix test sentence synthesis 2021-12-09 13:35:43 +00:00
WeberJulian 20ac31dc71 fix f0_cache_path in dataset 2021-12-09 13:35:12 +00:00
WeberJulian 6ed55ba57e fix test vits 2021-12-09 13:35:12 +00:00
WeberJulian 21b49c3acd fix collate_fn 2021-12-09 13:34:33 +00:00
Julian WEBER ec83ffbd7a PitchExtractor 2021-12-09 13:34:33 +00:00
Julian WEBER 3440c54bbe get_aux_input 2021-12-09 13:34:33 +00:00
Julian WEBER 5c89803968 Merge dataset 2021-12-09 13:33:35 +00:00
Edresson 3ac428340d Add audio resample in the speaker consistency loss 2021-12-09 13:32:25 +00:00
Edresson 39aff6685e Add freeze vocoder generator and flow-based decoder option 2021-12-09 13:31:04 +00:00
WeberJulian de41165af4 freeze vits parts 2021-12-09 13:31:04 +00:00
WeberJulian 9d2c445e3d get_speaker_weighted_sampler 2021-12-09 13:31:04 +00:00
Edresson 56480360cf Update the VITS model docs 2021-12-09 13:29:58 +00:00
Edresson cd7639ca70 Add voice conversion fine tuning mode 2021-12-09 13:29:58 +00:00
Edresson 3cd889a9d4 Add support to use the speaker encoder as loss function in VITS model 2021-12-09 13:29:58 +00:00
Edresson a3901032f4 Add H/ASP original checkpoint support 2021-12-09 13:28:16 +00:00
Edresson 256197b6aa Fix the optimizer parameters bug in multilingual and multispeaker training 2021-12-09 13:27:21 +00:00
Edresson f4abb19515 Fix bug after merge 2021-12-09 13:26:33 +00:00
Edresson d7042ecfd8 Fix d-vector multispeaker training bug 2021-12-09 13:26:33 +00:00