Production-grade fine-tuning & LoRA toolkit for Chatterbox-Flash zero-shot TTS models. Combines parallel block diffusion and FlashInfer acceleration with smart placeholder vocabulary extension supporting languages. Features offline feature preprocessing, Silero VAD silence trimming, zero-padding leakage prevention, and fast voice cloning for custom
text-to-speech deep-learning tts speech-synthesis lora audio-processing chatterbox fine-tuning peft voice-training voice-cloning turkish-tts zero-shot-tts flashinfer block-diffusion chatterbox-tts resemble-ai chatterbox-flash
-
Updated
Aug 23, 2026 - Python