unsloth-tts - Unsloth-optimized TTS fine-tuning for voice cloning
Fine-tunes text-to-speech models using Unsloth's optimized framework for voice cloning and speech synthesis.
Tags
Updated: 2026-02-15Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Load TTS models
- Prepare audio/text data
- Apply LoRA adapters
- Preprocess audio data
- Tokenize text with tags
- Train models with LoRA
- Save LoRA adapters
- Annotate transcripts with emotions
Inputs
- Audio/text data
- Annotated transcripts
- TTS models
- Unsloth library
- Audio files
- Text with special tags
Outputs
- Fine-tuned TTS models
- LoRA adapters
- Processed audio data
- Tokenized text data
- Trained voice clones
Requirements
- unsloth library
- librosa/soundfile libraries
- datasets library
- GPU with VRAM
- Orpheus-TTS compatibility