code/README.md

ACE-Step 1.5

Pushing the Boundaries of Open-Source Music Generation

Project | Hugging Face | ModelScope | Space Demo | Discord | Technical Report

Table of Contents

• ✨ Features
• 📦 Installation
• 🚀 Usage
• 🔨 Train
• 🏗️ Architecture
• 🦁 Model Zoo

📝 Abstract

🚀 We present ACE-Step v1.5, a highly efficient open-source music foundation model that brings commercial-grade generation to consumer hardware. On commonly used evaluation metrics, ACE-Step v1.5 achieves quality beyond most commercial music models while remaining extremely fast—under 2 seconds per full song on an A100 and under 10 seconds on an RTX 3090. The model runs locally with less than 4GB of VRAM, and supports lightweight personalization: users can train a LoRA from just a few songs to capture their own style.

🌉 At its core lies a novel hybrid architecture where the Language Model (LM) functions as an omni-capable planner: it transforms simple user queries into comprehensive song blueprints—scaling from short loops to 10-minute compositions—while synthesizing metadata, lyrics, and captions via Chain-of-Thought to guide the Diffusion Transformer (DiT). ⚡ Uniquely, this alignment is achieved through intrinsic reinforcement learning relying solely on the model's internal mechanisms, thereby eliminating the biases inherent in external reward models or human preferences. 🎚️

🔮 Beyond standard synthesis, ACE-Step v1.5 unifies precise stylistic control with versatile editing capabilities—such as cover generation, repainting, and vocal-to-BGM conversion—while maintaining strict adherence to prompts across 50+ languages. This paves the way for powerful tools that seamlessly integrate into the creative workflows of music artists, producers, and content creators. 🎸

✨ Features

⚡ Performance

• ✅ Ultra-Fast Generation — Under 2s per full song on A100, under 10s on RTX 3090 (0.5s to 10s on A100 depending on think mode & diffusion steps)
• ✅ Flexible Duration — Supports 10 seconds to 10 minutes (600s) audio generation
• ✅ Batch Generation — Generate up to 8 songs simultaneously

🎵 Generation Quality

• ✅ Commercial-Grade Output — Quality beyond most commercial music models (between Suno v4.5 and Suno v5)
• ✅ Rich Style Support — 1000+ instruments and styles with fine-grained timbre description
• ✅ Multi-Language Lyrics — Supports 50+ languages with lyrics prompt for structure & style control

🎛️ Versatility & Control

Feature	Description
✅ Reference Audio Input	Use reference audio to guide generation style
✅ Cover Generation	Create covers from existing audio
✅ Repaint & Edit	Selective local audio editing and regeneration
✅ Track Separation	Separate audio into individual stems
✅ Multi-Track Generation	Add layers like Suno Studio's "Add Layer" feature
✅ Vocal2BGM	Auto-generate accompaniment for vocal tracks
✅ Metadata Control	Control duration, BPM, key/scale, time signature
✅ Simple Mode	Generate full songs from simple descriptions
✅ Query Rewriting	Auto LM expansion of tags and lyrics
✅ Audio Understanding	Extract BPM, key/scale, time signature & caption from audio
✅ LRC Generation	Auto-generate lyric timestamps for generated music
✅ LoRA Training	One-click annotation & training in Gradio. 8 songs, 1 hour on 3090 (12GB VRAM)
✅ Quality Scoring	Automatic quality assessment for generated audio

📦 Installation

Requirements: Python 3.11, CUDA GPU recommended (works on CPU/MPS but slower)

1. Install uv (Package Manager)

# macOS / Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows (PowerShell)
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

2. Clone & Install

git clone https://github.com/ACE-Step/ACE-Step-1.5.git
cd ACE-Step-1.5
uv sync

3. Launch

🖥️ Gradio Web UI (Recommended)

uv run acestep

Open http://localhost:7860 in your browser. Models will be downloaded automatically on first run.

🌐 REST API Server

uv run acestep-api

API runs at http://localhost:8001. See API Documentation for endpoints.

Command Line Options

Gradio UI (acestep):

Option	Default	Description
--port	7860	Server port
--server-name	127.0.0.1	Server address (use 0.0.0.0 for network access)
--share	false	Create public Gradio link
--language	en	UI language: en, zh, ja
--init_service	false	Auto-initialize models on startup
--config_path	auto	DiT model (e.g., acestep-v15-turbo, acestep-v15-turbo-shift3)
--lm_model_path	auto	LM model (e.g., acestep-5Hz-lm-0.6B, acestep-5Hz-lm-1.7B)
--offload_to_cpu	auto	CPU offload (auto-enabled if VRAM < 16GB)

Examples:

# Public access with Chinese UI
uv run acestep --server-name 0.0.0.0 --share --language zh

# Pre-initialize models on startup
uv run acestep --init_service true --config_path acestep-v15-turbo

Development

# Add dependencies
uv add package-name
uv add --dev package-name

# Update all dependencies
uv sync --upgrade

🚀 Usage

We provide multiple ways to use ACE-Step:

Method	Description	Documentation
🖥️ Gradio Web UI	Interactive web interface for music generation	Gradio Guide
🐍 Python API	Programmatic access for integration	Inference API
🌐 REST API	HTTP-based async API for services	REST API

📚 Documentation available in: English | 中文 | 日本語

🔨 Train

See the LoRA Training tab in Gradio UI for one-click training, or check Gradio Guide - LoRA Training for details.

🏗️ Architecture

🦁 Model Zoo

DiT Models

DiT Model	Pre-Training	SFT	RL	CFG	Step	Refer audio	Text2Music	Cover	Repaint	Extract	Lego	Complete	Quality	Diversity	Fine-Tunability	Hugging Face
acestep-v15-base	✅	❌	❌	✅	50	✅	✅	✅	✅	✅	✅	✅	Medium	High	Easy	Link
acestep-v15-sft	✅	✅	❌	✅	50	✅	✅	✅	✅	❌	❌	❌	High	Medium	Easy	Link
acestep-v15-turbo	✅	✅	❌	❌	8	✅	✅	✅	✅	❌	❌	❌	Very High	Medium	Medium	Link
acestep-v15-turbo-rl	✅	✅	✅	❌	8	✅	✅	✅	✅	❌	❌	❌	Very High	Medium	Medium	To be released

LM Models

LM Model	Pretrain from	Pre-Training	SFT	RL	CoT metas	Query rewrite	Audio Understanding	Composition Capability	Copy Melody	Hugging Face
acestep-5Hz-lm-0.6B	Qwen3-0.6B	✅	✅	✅	✅	✅	Medium	Medium	Weak	✅
acestep-5Hz-lm-1.7B	Qwen3-1.7B	✅	✅	✅	✅	✅	Medium	Medium	Medium	✅
acestep-5Hz-lm-4B	Qwen3-4B	✅	✅	✅	✅	✅	Strong	Strong	Strong	To be released

📜 License & Disclaimer

This project is licensed under MIT

ACE-Step enables original music generation across diverse genres, with applications in creative production, education, and entertainment. While designed to support positive and artistic use cases, we acknowledge potential risks such as unintentional copyright infringement due to stylistic similarity, inappropriate blending of cultural elements, and misuse for generating harmful content. To ensure responsible use, we encourage users to verify the originality of generated works, clearly disclose AI involvement, and obtain appropriate permissions when adapting protected styles or materials. By using ACE-Step, you agree to uphold these principles and respect artistic integrity, cultural diversity, and legal compliance. The authors are not responsible for any misuse of the model, including but not limited to copyright violations, cultural insensitivity, or the generation of harmful content.

🔔 Important Notice
The only official website for the ACE-Step project is our GitHub Pages site.
We do not operate any other websites.
🚫 Fake domains include but are not limited to: ac**p.com, a**p.org, a***c.org
⚠️ Please be cautious. Do not visit, trust, or make payments on any of those sites.

🙏 Acknowledgements

This project is co-led by ACE Studio and StepFun.

📖 Citation

If you find this project useful for your research, please consider citing:

@misc{gong2026acestep,
    title={ACE-Step 1.5: Pushing the Boundaries of Open-Source Music Generation},
    author={Junmin Gong, Yulin Song, Wenxiao Zhao, Sen Wang, Shengyuan Xu, Jing Guo}, 
    howpublished={\url{https://github.com/ace-step/ACE-Step-1.5}},
    year={2026},
    note={GitHub repository}
}