v0.5.5

What's Changed

• fix(packaging): add missing python-multipart to all extra in uv.lock by @ykhrustalev in #960
• fix: correct LFM2.5-Audio sampling params and generation order by @ykhrustalev in #959
• Add Irodori-TTS MeanFlow distillation support by @yoshphys in #962
• fix: correct ISTFT geometry, output alignment and gradients by @pentaoa in #958
• Use the model's own default voice when none is given by @Lazarus-931 in #965
• Fix stream_generate with transformers tokenizers (eos_token_ids is an int) by @Lazarus-931 in #964
• Bump version to 0.5.5 by @lucasnewman in #968

New Contributors

• @ykhrustalev made their first contribution in #960
• @pentaoa made their first contribution in #958

Full Changelog: v0.5.4...v0.5.5