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
4,659 Commits 80 Branches 98 Tags 314 MiB
Commit Graph

416 Commits

Author SHA1 Message Date
Eren Gölge 72d85e53c9
Update model file extension (#1422) 
* Update model file ext to ```.pth```

* Update docs

* Rename more

* Find model files
 2022-03-22 17:55:00 +01:00
Eren Gölge 0870a4faa2
Make style (#1405) 2022-03-16 12:13:55 +01:00
Edresson Casanova f81892483d
REBASED: Transform Speaker Encoder in a Generic Encoder and Implement Emotion Encoder training support (#1349) 
* Rename Speaker encoder module to encoder

* Add a generic emotion dataset formatter

* Transform the Speaker Encoder dataset to a generic dataset and create emotion encoder config

* Add class map in emotion config

* Add Base encoder config

* Add evaluation encoder script

* Fix the bug in plot_embeddings

* Enable Weight decay for encoder training

* Add argumnet to disable storage

* Add Perfect Sampler and remove storage

* Add evaluation during encoder training

* Fix lint checks

* Remove useless config parameter

* Active evaluation in speaker encoder test and use multispeaker dataset for this test

* Unit tests fixs

* Remove useless tests for speedup the aux_tests

* Use get_optimizer in Encoder

* Add BaseEncoder Class

* Fix the unitests

* Add Perfect Batch Sampler unit test

* Add compute encoder accuracy in a function
 2022-03-11 14:43:40 +01:00
Edresson Casanova 36e9ea2f97
Open bible dataset formatter (#1365) 
* Add support for voice conversion inference

* Cache d_vectors_by_speaker for fast inference using a bigger speakers.json

* Rebase bug fix

* Use the average d-vector for inference

* Fix the bug in find unique chars script

* Add OpenBible formatter

Co-authored-by: Eren Gölge <erogol@hotmail.com>
 2022-03-11 10:43:31 +01:00
Edresson Casanova dbe9da7f15
Add Voice conversion inference support (#1337) 
* Add support for voice conversion inference

* Cache d_vectors_by_speaker for fast inference using a bigger speakers.json

* Rebase bug fix

* Use the average d-vector for inference
 2022-03-10 14:57:12 +01:00
Edresson Casanova f381e29b91
REBASED: Add support for the speaker encoder training using torch spectrograms (#1348) 
* Add support for the speaker encoder training using torch spectrograms

* Remove useless function in speaker encoder dataset class
 2022-03-10 14:54:51 +01:00
Eren Gölge 1425a023fe Make style and lint 2022-03-02 13:25:35 +01:00
Eren Gölge 1e414b3a09 Make stlye 2022-02-25 11:31:56 +01:00
Eren Gölge bf540f4323 Update imports for trainer 2022-02-25 11:31:56 +01:00
Eren Gölge 424d04e4f6 Make stlye 2022-02-25 11:31:56 +01:00
Eren Gölge be3a03126a Update imports for trainer 2022-02-25 11:28:14 +01:00
Eren Gölge 8622226f3f Make style 2022-02-25 11:26:59 +01:00
Eren Gölge 27db089d6c Change TrainingArgs -> TrainerArgs 2022-02-25 11:26:59 +01:00
Eren Gölge 1f0c8179da Make style 2022-02-25 11:26:59 +01:00
Eren Gölge bc2243bac4 Fix tests 2022-02-25 11:25:00 +01:00
Eren Gölge 2fe16de8e3 Make lint 2022-02-25 11:25:00 +01:00
Eren Gölge 8e248913d6 Update train_tts for the new API 2022-02-25 11:24:13 +01:00
Eren Gölge 730f7c0df4 Add file_ext args to resample.py 2022-02-25 11:15:46 +01:00
Eren Gölge 4d99fee3e2 Update spec extractor 2022-02-25 11:12:44 +01:00
Eren Gölge 38a0b3b6c7 Update train_tts.py 2022-02-25 11:11:35 +01:00
Eren Gölge fbad17e084 Update imports for symbols -> characters 2022-02-25 10:48:02 +01:00
Eren Gölge a51b031bff
Merge branch 'dev' into dev-fix-glowtts-infer 2022-02-21 12:01:40 +03:00
Edresson Casanova 28a7464975
Fix the bug in split dataset function (#1251) 
* Fix the bug in split_dataset

* Make eval_split_size configurable

* Change test_loader to use load_tts_samples function

* Change eval_split_portion to eval_split_size and permits to set the absolute number of samples in eval

* Fix samplers unit test

* Add data unit test on GitHub workflow
 2022-02-21 11:59:36 +03:00
Edresson Casanova bc5db13d06 Fix the bug in extract tts spectrogram script 2022-02-19 19:24:00 +00:00
Eren Gölge 127118c637
Update TTS.tts formatters (#1228) 
* Return Dict from tts formatters

* Make style
 2022-02-11 23:03:43 +01:00
Edresson Casanova 0860d73cf8
Remove Tensorflow requeriment (#1225) 
* Remove TF modules

* Remove TF unit tests

* Remove TF vocoder modules

* Remove TF convert scripts

* Remove TF requirement

* Remove the Docs TF instructions

* Remove TF inference support
 2022-02-10 16:14:54 +01:00
WeberJulian c7f5e005e1 Compute embedding for new audios only 2022-01-06 15:41:38 +01:00
WeberJulian e1accb6e28
Fix train_tts.py and uncomment code (#1051) 
* Fix SE loading and language embedding logic

* remove trailing white space

* Uncomment resmapling code for SCL
 2022-01-03 17:44:57 +01:00
Eren Gölge 56378b12f7 Fix speaker encoder init 2021-12-21 12:26:25 +00:00
Eren Gölge 4c50f6f4df Add functions to get and check and argument in config and config.model_args 2021-12-20 11:54:10 +00:00
Eren Gölge 704dddcffa Make style 2021-12-20 11:54:10 +00:00
WeberJulian a564eb9f54 Add support for multi-lingual models in CLI 2021-12-20 11:54:10 +00:00
WeberJulian 6b03943526 Move multilingual logic out of the trainer 2021-12-20 11:54:10 +00:00
Edresson 4196a42de7 Get the number speaker from the Speaker Manager property 2021-12-20 11:54:10 +00:00
Edresson f394d60695 Fix the bug in multispeaker vits 2021-12-20 11:54:10 +00:00
Edresson 45d0b04179 Lint fixs 2021-12-20 11:54:10 +00:00
Edresson 85418ffeaa Fix the bug in extract tts spectrograms 2021-12-20 11:54:10 +00:00
Edresson 34749f8727 Remove the call to get_speaker_manager 2021-12-20 11:54:10 +00:00
Edresson 352aa69eca Create a module for the VAD script 2021-12-20 11:54:10 +00:00
WeberJulian da6c1e858c Fix small issues 2021-12-20 11:54:10 +00:00
WeberJulian 23d789c072 Fix continue path 2021-12-20 11:54:10 +00:00
WeberJulian 120332d53f Fix phonemes 2021-12-20 11:54:10 +00:00
WeberJulian 846bf16f02 fix imports for load_meta_data 2021-12-20 11:54:10 +00:00
WeberJulian e995a63bd6 fix linter 2021-12-20 11:54:10 +00:00
WeberJulian 1472b6df49 make style 2021-12-20 11:54:10 +00:00
Edresson 10ff90d6d2 Add remove silence VAD script 2021-12-20 11:54:10 +00:00
Edresson eeb8ac07d9 Add voice conversion fine tuning mode 2021-12-20 11:54:10 +00:00
Reuben Morais 859ac1a54c Include usage instructions in README 2021-12-17 11:37:19 +01:00
Eren Gölge babdd84f91 Fix GST inference 
commit d3e477875a7e46a101fcf95a1794442823750fe2
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Wed Nov 3 10:16:12 2021 +0000

    Read .wav for GST conditioning from CL

commit 074e6d0874d3b34fb6a4991fc17d66dccd413fbb
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Fri Oct 29 14:43:47 2021 +0100

    Fix GST during inference in Tacotron2

commit fdece14585ab5a36eed1061a9a838d8e48aa6882
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Wed Nov 3 10:16:12 2021 +0000

    Read .wav for GST conditioning from CL

commit cd29e21b8d0a541ee298d2bf5f67223ad60be38f
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Fri Oct 29 14:43:47 2021 +0100

    Fix GST during inference in Tacotron2

commit 908ce39370eadcc9fa8510cdb26c9ead87305427
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Fri Oct 29 12:49:37 2021 +0100

    Make trim_db value negative

commit 1008a2e0f72fa7ca7f0307424f570386f2f16d42
Author: George Rousssos <25833833+george-roussos@users.noreply.github.com>
Date:   Fri Oct 29 12:22:24 2021 +0100

    Set find_endpoint db threshold in config.json
 2021-12-07 13:28:49 +00:00
Eren Gölge ce45d9e1af Make style and lint 2021-12-01 10:42:52 +00:00
Eren Gölge dc3dd55dd9 Add collect_env_info.py 2021-11-08 08:59:08 +01:00
Eren Gölge a409e0f8f8 Update train_tts for multi-speaker 2021-10-21 16:29:06 +00:00
Eren Gölge ba2b8c827f Update `train_tts.py` and `train_vocoder.py` 2021-09-30 14:47:56 +00:00
Eren Gölge 2e9b6b4f90 Refactor Speaker Encoder training 2021-09-30 14:47:56 +00:00
Eren Gölge 043dca61b4 Rename `load_meta_data` as `load_tts_data` 2021-09-30 14:47:56 +00:00
Eren Gölge 3c740d4893 Style extract_tts_spectrogram.py 2021-09-10 08:21:21 +00:00
Eren Gölge 807f1d3817 Fix `extract_tts_spectrograms.py` model init 2021-09-09 08:59:55 +00:00
Eren Gölge 91a70e80b2 Refactor TTSDataset 
Return a dict by `collate`
Refactor batch handling in `collate`
A couple of bug fixes
 2021-09-06 15:16:58 +00:00
Eren Gölge 545a00fc04 Use absolute paths of the attention masks 2021-09-06 15:16:58 +00:00
Eren Gölge 0f19f8c911 Fix `compute_attention_masks.py` 2021-09-06 15:16:58 +00:00
Eren Gölge 18da8f5dbd Update pylint 2.10.2 and fix lint issues 2021-08-30 08:10:35 +00:00
Eren Gölge f186856e5d Add option to sort input sequnce by audio len 2021-08-30 08:10:35 +00:00
Eren Gölge 5911eec3b1 Small trainer refactoring 
1. Use a single Gradscaler for all the optimizers
2. Save terminal logs to a file. In DDP mode, each worker creates `trainer_N_log.txt`.
3. Fixes to allow only the main worker (rank==0) writing to Tensorboard
4. Pass parameters owned by the target optimizer to the grad_clip_norm
 2021-08-26 17:08:58 +00:00
Eren Gölge ecf5f17dca Fix distribute.py and ddp training 2021-08-12 22:22:32 +00:00
Eren Gölge 6af03ac476 Fix `num_char` init in Tacotron models 2021-08-09 21:46:15 +00:00
Ayush Chaurasia 936a47504d Update Logger API, recipes 2021-08-09 18:34:00 +00:00
Ayush Chaurasia f63cf46c55 Unified logger API 2021-08-09 18:34:00 +00:00
Ayush Chaurasia f606741dc4 Add artifacts logging , wandb args 2021-08-09 18:31:16 +00:00
Agrin Hilmkil ced4cfdbbf Allow saving / loading checkpoints from cloud paths (#683) 
* Allow saving / loading checkpoints from cloud paths

Allows saving and loading checkpoints directly from cloud paths like
Amazon S3 (s3://) and Google Cloud Storage (gs://) by using fsspec.

Note: The user will have to install the relevant dependency for each
protocol. Otherwise fsspec will fail and specify which dependency is
missing.

* Append suffix _fsspec to save/load function names

* Add a lower bound to the fsspec dependency

Skips the 0 major version.

* Add missing changes from refactor

* Use fsspec for remaining artifacts

* Add test case with path requiring fsspec

* Avoid writing logs to file unless output_path is local

* Document the possibility of using paths supported by fsspec

* Fix style and lint

* Add missing lint fixes

* Add type annotations to new functions

* Use Coqpit method for converting config to dict

* Fix type annotation in semi-new function

* Add return type for load_fsspec

* Fix bug where fs not always created

* Restore the experiment removal functionality
 2021-08-09 18:02:36 +00:00
Eren Gölge 4b7b88dd3d Add fullband-melgan DE vocoder 2021-07-26 15:38:30 +02:00
Edresson Casanova d5adc35fdf
Add docstring to compute_embeddings script 2021-07-21 07:16:10 -03:00
Edresson b1620d1f3f remove ignore generate eval flag 2021-07-15 03:34:28 -03:00
Edresson d906fea08c lint fix and eval as argparse in extract tts spectrograms 2021-07-13 02:15:31 -03:00
Edresson 2e5baffa9c Merge fix and eval split as argparse 2021-07-13 01:47:32 -03:00
Eren Gölge 93a74cbb71
Merge pull request #628 from Aloento/patch-2 
Change to _get_preprocessor_by_name
 2021-07-11 22:17:50 +02:00
Edresson 4eac1c4651 bug fix on train_encoder and unit tests 2021-07-11 12:00:39 -03:00
Aloento 6e3e6d5756
Change to _get_preprocessor_by_name 2021-07-08 09:53:13 +02:00
Eren Gölge a4c658f5ef Fix for using the `Synthesizer` out of the model 2021-07-02 10:43:38 +02:00
Eren Gölge b3c073c99b Allow runing full path scripts with `distribute.py` 2021-06-28 17:03:47 +02:00
Eren Gölge a7617d8ab6 Add 🐍 python 3.9 to CI 2021-06-28 17:03:47 +02:00
Eren Gölge 9790eddada Fix wrong argument name 🛠️ 2021-06-28 17:03:47 +02:00
Eren Gölge 45947acb60 Update `TTS.bin` scripts for the new API 2021-06-28 17:03:47 +02:00
Eren Gölge c7aad884cd Implement unified trainer 2021-06-28 17:03:19 +02:00
Eren Gölge c754a0e17d `TrainerAbstract` and related updates for `TrainerTTS` 2021-06-28 17:03:19 +02:00
Eren Gölge 00c82c516d rename to 2021-06-28 17:03:19 +02:00
Eren Gölge 03494ad642 adjust `distribute.py` for the `train_tts.py` 2021-06-28 17:03:19 +02:00
Eren Gölge d6b2b6add6 make style and linter fixes 2021-06-28 17:03:19 +02:00
Eren Gölge 802d461389 Compute d_vectors and speaker_ids separately in TTSDataset 2021-06-28 17:03:19 +02:00
Eren Gölge db6a97d1a2 rename external speaker embedding arguments as `d_vectors` 2021-06-28 17:03:19 +02:00
Eren Gölge ef4ea9e527 update imports for `formatters` 2021-06-28 17:03:19 +02:00
Eren Gölge 421194880d linter fixes 2021-06-28 17:03:19 +02:00
Eren Gölge 8e52a69230 delete separate tts training scripts and pre-commit configuration 2021-06-28 17:03:19 +02:00
Eren Gölge d96ebcd6d3 make style 2021-06-28 17:03:19 +02:00
Eren Gölge b500338faa make style 2021-06-28 17:03:19 +02:00
Eren Gölge 469d2e620a update extract_tts_spectrogram for `cond_input` API of the models 2021-06-28 17:03:19 +02:00
Eren Gölge 5ab28fa618 update `extract_tts_spec...` using `SpeakerManager` 2021-06-28 17:03:19 +02:00
Eren Gölge c392fa4288 update `extract_tts_spectrograms` for the new model API 2021-06-28 17:03:19 +02:00
Eren Gölge 8f47f95998 correct import of `load_meta_data` 
remove redundant import
 2021-06-28 17:03:19 +02:00
Eren Gölge d25f017b42 update `setup_model.py` imports 2021-06-28 17:03:19 +02:00
Eren Gölge e298b8e364 update trainer.py for better logging handling, restoring models and 
rename init_ functions with get_
 2021-06-28 17:03:19 +02:00