Tool Details & Introduction
ACE-Step 1.5 is an advanced open source, locally deployable AI music base model. It uses a hybrid architecture of "Language Model Planner + Diffusion Transformer (DiT)" to convert text prompt words into a complete musical composition (up to 10 minutes) containing high-fidelity human voices and rich accompaniment.
Core functions and features
- Fully Open Source and Runs Natively: Open source under the MIT license and can be deployed natively on consumer hardware with mainstream NVIDIA, AMD, Intel or Apple Silicon GPUs.
- LoRA fine-tuning support: Users can use their own audio samples or the voices of specific singers for LoRA training to customize their own singing voices and music styles.
- Multi-functional task support: In addition to basic text-to-music generation, it also supports audio inpainting, vocal-background sound separation and song cover generation.