leeoisaboy/cosyvoice-lora-finetune-framework
🎙️ CosyVoice LoRA 微调框架：LLM+Flow 联合训练，实现无 Prompt 语音合成
This framework helps content creators, voice artists, or anyone needing custom voice generation to fine-tune an AI voice model using a small amount of audio data (10-50 samples). It takes sample audio of a specific speaker and produces an AI voice that can generate speech in that speaker's unique tone and style, without needing a reference audio each time. This is ideal for scenarios like creating narrated content or preserving a distinctive voice for various applications.
Use this if you need to create a high-quality, custom AI voice that sounds exactly like a specific person and can generate speech without requiring a reference audio during use.
Not ideal if you only need generic voice synthesis or if you prefer to always provide a reference audio for each generation rather than training a persistent custom voice.
Stars
7
Forks
—
Language
Python
License
—
Category
Last pushed
Jan 05, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/leeoisaboy/cosyvoice-lora-finetune-framework"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
BoltzmannEntropy/MimikaStudio
MimikaStudio - A local-first application for macOS (Apple Silicon) + Agentic MCP Support
aahl/qwen-asr2api
🎤 Qwen 3 ASR to OpenAI API, 免费STT语音识别模型
gabriele-mastrapasqua/qwen3-tts
Pure C inference engine for Qwen3-TTS text-to-speech. No Python, no PyTorch — just C and BLAS....
zhao-kun/VibeVoiceFusion
VibeVoiceFusion is a full-stack, multi-speaker voice generation web system featuring LoRA...
shijincai/VibeVoice
Archive of the official Microsoft VibeVoice repository (7B & 1.5B). Backup of the deleted source...