README.md

---
license: other
title: ace-step-v1-5
sdk: docker
emoji: 🎼
colorFrom: gray
colorTo: gray
pinned: true
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/65d93c08e4d11d592e21dfc5/NEe28KukXKoZChdv6EmP8.webp
short_description: Music generation with ACE-STEP v1.5
---
# 🎵 ACE-Step v1.5 — Gradio Interface

This Space runs the official [ACE-Step v1.5](https://github.com/ACE-Step/ACE-Step-1.5) music generation model with its included Gradio interface.

## 🚀 What is ACE-Step v1.5?
- **Commercial-grade music generation** outperforming most closed-source alternatives
- **Ultra-fast**: < 2 seconds per song on A100, < 10 seconds on RTX 3090
- **Flexible duration**: 10 seconds to 10 minutes
- **Multi-language lyrics**: 50+ languages supported
- **Cover generation, repainting, editing, LoRA training** and more

## 💾 Storage
- **Persistent storage** is mounted at `/data` for models, cache, and outputs.
- **Bucket**: [Falln87/ace-step-storage](https://huggingface.co/buckets/Falln87/ace-step-storage) for additional file storage.

## ⚙️ GPU Configuration
For best performance, configure this space to use a **GPU** in the space settings.

## 📝 First Run
The first time you run this space, it will download the required models. This may take several minutes depending on your hardware. The models are cached in `/data` for subsequent runs.

## 🔗 Links
- [ACE-Step v1.5 GitHub](https://github.com/ACE-Step/ACE-Step-1.5)
- [ACE-Step Hugging Face](https://huggingface.co/ACE-Step/Ace-Step1.5)
- [Official Space Demo](https://huggingface.co/spaces/ACE-Step/Ace-Step-v1.5)