v0.24.2

[eginhard]

What's Changed

Highlights

• Prebuilt wheels now also provided for Mac and Windows
• Hindi fully supported for XTTS fine-tuning

Features

• Add multilingual phoneme cleaner by @eginhard in Add multilingual phoneme cleaner #44
• Support hindi in XTTS tokenizer by @akshatrocky in feat(xtts): support hindi in tokenizer #64
• Normalize unicode characters in text cleaners by @shavit in Add normalizer type C to text cleaners #85

Fixes

• Improve CLI handling of boolean arguments by @eginhard in Improve CLI handling of boolean arguments #38
• Use the specified device for pretrained speaker encoder by @ChristianRomberg in Use the specified device for pretrained speaker encoder #45
• XTTS streaming fixes by @eginhard in Fix XTTS streaming for transformers update #46 and by @gravityrail in Fix missing param on transformers 4.4 #59
• Numpy 2 support by @eginhard in build: add numpy2 support #47 Numpy 2 support (gruut, soxr, spacy) #56 Enable Numpy 2 support after Spacy 3.8 release #81
• VC-related refactors and fixes by @eginhard in VC-related refactors and fixes #49
• Update to coqui-tts-trainer 0.1.4 by @eginhard in Update to coqui-tts-trainer 0.1.4 #51
• Fix typo in xtts.py by @abrahammathews2000 in Update xtts.py #53
• build: move umap-learn into optional notebook dependencies by @eginhard in build: move umap-learn into optional notebook dependencies #50
• fix(xtts): load tokenizer file based on config as last resort by @eginhard in fix(xtts): load tokenizer file based on config as last resort #57
• Skip audio files that can't be decoded by @eginhard in Skip audio files that can't be decoded #66
• Fix wrong references in documentation by @hykilpikonna in Fix wrong references in documentation #74
• fix(build): restrict spacy version to unbreak installation by @KoljaB in Potential fix for #90, #91 #92
• ci(release): switch to cibuildwheel by @eginhard in ci(release): switch to cibuildwheel #95

New Contributors

• @ChristianRomberg made their first contribution in Use the specified device for pretrained speaker encoder #45
• @abrahammathews2000 made their first contribution in Update xtts.py #53
• @gravityrail made their first contribution in Fix missing param on transformers 4.4 #59
• @akshatrocky made their first contribution in feat(xtts): support hindi in tokenizer #64
• @hykilpikonna made their first contribution in Fix wrong references in documentation #74
• @shavit made their first contribution in Add normalizer type C to text cleaners #85
• @KoljaB made their first contribution in Potential fix for #90, #91 #92

Full Changelog: v0.24.1...v0.24.2

---

This discussion was created from the release v0.24.2.