Downloads · 30 days
0
GiantAILab/YingMusic-Singer
YingMusic-Singer is a machine learning model from GiantAILab. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as cc-by-nc-4.0.
Downloads · 30 days
0
Access
Public
Updated Feb 9, 2026
Repo size
6.5 GB
Likes
9
Public
Click a slice to open those files.
.pt4 GB · 85%
From the Hugging Face model README
github:YingMusic-Singer
YingMusic-Singer is a unified framework for Zero-shot Singing Voice Synthesis (SVS) and Editing, driven by Annotation-free Melody Guidance. Addressing the scalability challenges of real-world applications, our system eliminates the reliance on costly phoneme-level alignment and manual melody annotations. It enables arbitrary lyrics to be synthesized or edited with any reference melody in a zero-shot manner. Our approach leverages a Diffusion Transformer (DiT) based generative model, incorporating a pre-trained melody extraction module to derive MIDI information directly from reference audio. By introducing a structured guidance mechanism and employing Flow-GRPO reinforcement learning, we achieve superior pronunciation clarity, melodic accuracy, and musicality without requiring fine-grained alignment.