v0.6.2

[erogol]

What's Changed

• Fix multilingual recipe by @Edresson in Fix multilingual recipe #1354
• Fix recipes as to the recent API changes. by @erogol in Fix recipes as to the recent API changes.  #1367
• Add docsqa to docs website by @nomagick in Add docsqa to docs website #1363
• REBASED: Add support for the speaker encoder training using torch spectrograms by @Edresson in REBASED: Add support for the speaker encoder training using torch spectrograms  #1348
• Add alphas to control language and speaker balancer by @Edresson in Add alphas to control language and speaker balancer #1216
• Add Voice conversion inference support by @Edresson in Add Voice conversion inference support #1337
• Update issue template by @erogol in Update issue template #1370
• Open bible dataset formatter by @Edresson in Open bible dataset formatter #1365
• REBASED: Transform Speaker Encoder in a Generic Encoder and Implement Emotion Encoder training support by @Edresson in REBASED: Transform Speaker Encoder in a Generic Encoder and Implement Emotion Encoder training support  #1349
• Fix typo workflow text by @WeberJulian in Fix typo workflow text #1403
• Add CITATION.cff by @erogol in Add CITATION.cff #1404
• Fix default phonemizer for ja and zh by @WeberJulian in Fix default phonemizer for ja and zh #1399
• Make Style by @erogol in Make Style #1405
• Fix [Bug] Unable to initialize a LibriTTS Dataset with the Python Trainer API? #1380 by @erogol in Fix #1380 #1409
• Hinge Gruut version to 2.2.3 by @erogol in Hinge Gruut version to 2.2.3 #1419
• Update CheckSpectrograms notebook by @erogol in Update CheckSpectrograms notebook #1418
• Fix [Bug] eval_split_size in recipe and config.json don't match up #1423 by @Edresson in Fix #1423 #1424
• Update model file extension by @erogol in Update model file extension #1422
• Fix model manager by @erogol in Fix model manager  #1436
• Add formatting tests by @WeberJulian in Add formatting tests #1437
• Update base model wrt 👟 by @erogol in Update base model wrt 👟 #1406
• Replace webrtcvad by silero-vad by @Edresson in Replace webrtcvad by silero-vad #1431
• Bug fix in freeze encoder by @Edresson in Bug fix in freeze encoder #1391
• Enforce phonemizer definition for synthesis by @WeberJulian in Enforce phonemizer definition for synthesis #1441
• Fix G2P backend of the released models by @WeberJulian in Fix G2P backend of the released models #1461
• Add EmbeddingManager and BaseIDManager by @Edresson in Add EmbeddingManager and BaseIDManager #1374
• Update requirements coqui_trainer -> trainer by @erogol in Update requirements coqui_trainer -> trainer #1478
• Update CONTRIBUTING.md, fix header by @Jackiexiao in Update CONTRIBUTING.md, fix header #1463
• Add African models by @WeberJulian in Add African models #1511
• Print Model's license when downloading by @erogol in Print Model's license when downloading  #1512
• Improve docsQA default questions by @nomagick in Improve docsQA default questions #1411
• patch print license by @WeberJulian in patch print license #1514
• v0.6.2 by @erogol in v0.6.2 #1353

New Contributors

• @nomagick made their first contribution in Add docsqa to docs website #1363
• @Jackiexiao made their first contribution in Update CONTRIBUTING.md, fix header #1463

Full Changelog: v0.6.1...v0.6.2

---

This discussion was created from the release v0.6.2.

[godspirit00]

Thanks for the great work!
Is there a tutorial for the voice conversion inference?
I tried tts --model_name "tts_models/en/vctk/vits" --reference_wav "J:/1646276861.wav" --speaker_idx p250 --out_path "D:/1.wav"
but got

Traceback (most recent call last):
  File "C:\Users\xxx\Envs\coqui\Scripts\tts-script.py", line 33, in <module>
    sys.exit(load_entry_point('TTS', 'console_scripts', 'tts')())
  File "d:\github\tts\TTS\bin\synthesize.py", line 305, in main
    wav = synthesizer.tts(
  File "d:\github\tts\TTS\utils\synthesizer.py", line 336, in tts
    outputs = transfer_voice(
  File "d:\github\tts\TTS\tts\utils\synthesis.py", line 273, in transfer_voice
    model_outputs = _func(reference_wav, speaker_id, d_vector, reference_speaker_id, reference_d_vector)
  File "C:\Users\xxx\Envs\coqui\lib\site-packages\torch\autograd\grad_mode.py", line 27, in decorate_context
    return func(*args, **kwargs)
  File "d:\github\tts\TTS\tts\models\vits.py", line 1022, in inference_voice_conversion
    wav, _, _ = self.voice_conversion(y, y_lengths, speaker_cond_src, speaker_cond_tgt)
  File "d:\github\tts\TTS\tts\models\vits.py", line 1039, in voice_conversion
    g_src = self.emb_g(speaker_cond_src).unsqueeze(-1)
  File "C:\Users\xxx\Envs\coqui\lib\site-packages\torch\nn\modules\module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "C:\Users\xxx\Envs\coqui\lib\site-packages\torch\nn\modules\sparse.py", line 158, in forward
    return F.embedding(
  File "C:\Users\xxx\Envs\coqui\lib\site-packages\torch\nn\functional.py", line 2183, in embedding
    return torch.embedding(weight, input, padding_idx, scale_grad_by_freq, sparse)
TypeError: embedding(): argument 'indices' (position 2) must be Tensor, not NoneType

Not sure if I got it right. Any help please? Thank you!

[erogol]

That model does not use reference audio. It just takes a speaker ID. Reference audio is only necessary with YourTTS.

But I guess we need better messaging to inform the user.