seed-star-v1.5

Seed-VC seed-uvit-whisper-base Finetune

Introduction

This model is a fine-tuned version of Plachta/Seed-VC's seed-uvit-whisper-base with 168 hours of clean singing audios in korean.
It demonstrates significant improvements in naturalness and voice quality.

---

🎯 Reference Audio

<audio controls preload="none" style="width:100%;"> <source src="https://huggingface.co/WeirdhostRnD/seed-star-v1.5/resolve/main/sample/refrence.wav" type="audio/mpeg" /> Your browser does not support the audio element. </audio>

---

🎧 Audio Comparison

<table> <tr> <th>Model</th> <th>===================Converted Singing Audio (25 steps)===================</th> </tr> <tr> <td><strong>Original</strong></td> <td> <audio controls preload="none" style="width:100%;"> <source src="https://huggingface.co/WeirdhostRnD/seed-star-v1.5/resolve/main/sample/original.mp3" type="audio/mpeg" /> Your browser does not support the audio element. </audio> </td> </tr> <tr> <td><strong>Base</strong></td> <td> <audio controls preload="none" style="width:100%;"> <source src="https://huggingface.co/WeirdhostRnD/seed-star-v1.5/resolve/main/sample/base.mp3" type="audio/mpeg" /> Your browser does not support the audio element. </audio> </td> </tr> <tr> <td><strong>Finetune</strong></td> <td> <audio controls preload="none" style="width:100%;"> <source src="https://huggingface.co/WeirdhostRnD/seed-star-v1.5/resolve/main/sample/finetune.mp3" type="audio/mpeg" /> Your browser does not support the audio element. </audio> </td> </tr> </table>