Merge branch 'main' into huggingface_space

README.md

@@ -1,69 +1,217 @@
-# ACE-Step-1.5
+<h1 align="center">ACE-Step 1.5</h1>
+<h1 align="center">Pushing the Boundaries of Open-Source Music Generation</h1>
+<p align="center">
+    <a href="https://ace-step-v1.5.github.io">Project</a> |
+    <a href="https://huggingface.co/collections/ACE-Step/ace-step-15">Hugging Face</a> |
+    <a href="https://modelscope.cn/models/ACE-Step/ACE-Step-v1-5">ModelScope</a> |
+    <a href="https://huggingface.co/spaces/ACE-Step/ACE-Step-1.5">Space Demo</a> |
+    <a href="https://discord.gg/PeWDxrkdj7">Discord</a> |
+    <a href="https://arxiv.org/abs/2506.00045">Technical Report</a>
+</p>
 
-## Installation
+<p align="center">
+    <img src="./assets/orgnization_logos.png" width="100%" alt="StepFun Logo">
+</p>
 
-This project uses [uv](https://github.com/astral-sh/uv) for dependency management.
+## Table of Contents
 
-### Install uv
+- [✨ Features](#-features)
+- [📦 Installation](#-installation)
+- [🚀 Usage](#-usage)
+- [🔨 Train](#-train)
+- [🏗️ Architecture](#️-architecture)
+- [🦁 Model Zoo](#-model-zoo)
+
+## 📝 Abstract
+We present ACE-Step v1.5, a highly efficient foundation model that democratizes commercial-grade music production on consumer hardware. Optimized for local deployment (<4GB VRAM), the model accelerates generation by over 100× compared to traditional pure LM architectures, producing superior high-fidelity audio in seconds characterized by coherent semantics and exceptional melodies. 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.
+
+
+## ✨ Features
+
+<p align="center">
+    <img src="./assets/application_map.png" width="100%" alt="ACE-Step Framework">
+</p>
+
+### ⚡ Performance
+- ✅ **Ultra-Fast Generation** — 0.5s to 10s generation time 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 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)
 
 ```bash
+# 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"
-
-# macOS/Linux
-curl -LsSf https://astral.sh/uv/install.sh | sh
 ```
 
-### Install Project Dependencies
+### 2. Clone & Install
 
 ```bash
-# Sync all dependencies
+git clone https://github.com/ACE-Step/ACE-Step-1.5.git
+cd ACE-Step-1.5
 uv sync
 ```
 
-### Run the Project
+### 3. Launch
+
+#### 🖥️ Gradio Web UI (Recommended)
 
 ```bash
-# Simplest way - run directly with uv
 uv run acestep
+```
 
-# Run with parameters
-uv run acestep --port 7860 --server-name 0.0.0.0 --share
+Open http://localhost:7860 in your browser. Models will be downloaded automatically on first run.
 
-# Or use the full module path
-uv run python -m acestep.acestep_v15_pipeline
+#### 🌐 REST API Server
 
-# Just Run profiling
-uv run profile_inference.py
+```bash
+uv run acestep-api
+```
 
-# Or activate the virtual environment first
-source .venv/bin/activate  # Linux/macOS
-# or
-.venv\Scripts\activate  # Windows
+API runs at http://localhost:8001. See [API Documentation](./docs/en/API.md) for endpoints.
 
-acestep
-```
+### Command Line Options
 
-Available parameters:
-- `--port`: Server port (default: 7860)
-- `--server-name`: Server address (default: 127.0.0.1, use 0.0.0.0 to listen on all interfaces)
-- `--share`: Create a public share link
-- `--debug`: Enable debug mode
+**Gradio UI (`acestep`):**
 
-## Development
+| 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) |
 
-Add new dependencies:
+**Examples:**
 
 ```bash
-# Add runtime dependencies
-uv add package-name
+# Public access with Chinese UI
+uv run acestep --server-name 0.0.0.0 --share --language zh
 
-# Add development dependencies
-uv add --dev package-name
+# Pre-initialize models on startup
+uv run acestep --init_service true --config_path acestep-v15-turbo
 ```
 
-Update dependencies:
+### Development
 
 ```bash
+# 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](./docs/en/GRADIO_GUIDE.md) |
+| 🐍 **Python API** | Programmatic access for integration | [Inference API](./docs/en/INFERENCE.md) |
+| 🌐 **REST API** | HTTP-based async API for services | [REST API](./docs/en/API.md) |
+
+**📚 Documentation available in:** [English](./docs/en/) | [中文](./docs/zh/) | [日本語](./docs/ja/)
+
+
+## 🔨 Train
+
+See the **LoRA Training** tab in Gradio UI for one-click training, or check [Gradio Guide - LoRA Training](./docs/en/GRADIO_GUIDE.md#lora-training) for details.
+
+## 🏗️ Architecture
+
+<p align="center">
+    <img src="./assets/ACE-Step_framework.png" width="100%" alt="ACE-Step Framework">
+</p>
+
+## 🦁 Model Zoo
+
+<p align="center">
+    <img src="./assets/model_zoo.png" width="100%" alt="Model Zoo">
+</p>
+
+### 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](https://huggingface.co/ACE-Step/acestep-v15-base) |
+| `acestep-v15-sft` | ✅ | ✅ | ❌ | ✅ | 50 | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ | High | Medium | Easy | [Link](https://huggingface.co/ACE-Step/acestep-v15-sft) |
+| `acestep-v15-turbo` | ✅ | ✅ | ❌ | ❌ | 8 | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ | Very High | Medium | Medium | [Link](https://huggingface.co/ACE-Step/Ace-Step1.5) |
+| `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](./LICENSE)
+
+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:
+
+```BibTeX
+@misc{gong2026acestep,
+	title={ACE-Step 1.5: Pushing the Boundaries of Open-Source Music Generation},
+	author={Junmin Gong, Song Yulin, Wenxiao Zhao, Sen Wang, Shengyuan Xu, Jing Guo}, 
+	howpublished={\url{https://github.com/ace-step/ACE-Step-1.5}},
+	year={2026},
+	note={GitHub repository}
+}
+```

acestep/acestep_v15_pipeline.py

@@ -134,7 +134,7 @@ def main():
     # Service initialization arguments
     parser.add_argument("--init_service", type=lambda x: x.lower() in ['true', '1', 'yes'], default=False, help="Initialize service on startup (default: False)")
     parser.add_argument("--checkpoint", type=str, default=None, help="Checkpoint file path (optional, for display purposes)")
-    parser.add_argument("--config_path", type=str, default=None, help="Main model path (e.g., 'acestep-v15-turbo-rl')")
+    parser.add_argument("--config_path", type=str, default=None, help="Main model path (e.g., 'acestep-v15-turbo')")
     parser.add_argument("--device", type=str, default="auto", choices=["auto", "cuda", "cpu"], help="Processing device (default: auto)")
     parser.add_argument("--init_llm", type=lambda x: x.lower() in ['true', '1', 'yes'], default=True, help="Initialize 5Hz LM (default: True)")
     parser.add_argument("--lm_model_path", type=str, default=None, help="5Hz LM model path (e.g., 'acestep-5Hz-lm-0.6B')")

acestep/api_server.py

@@ -608,7 +608,7 @@ def create_app() -> FastAPI:
         app.state.handler3 = handler3
         app.state._initialized2 = False
         app.state._initialized3 = False
-        app.state._config_path = os.getenv("ACESTEP_CONFIG_PATH", "acestep-v15-turbo-rl")
+        app.state._config_path = os.getenv("ACESTEP_CONFIG_PATH", "acestep-v15-turbo")
         app.state._config_path2 = config_path2
         app.state._config_path3 = config_path3
 
@@ -661,7 +661,7 @@ def create_app() -> FastAPI:
                     raise RuntimeError(app.state._init_error)
 
                 project_root = _get_project_root()
-                config_path = os.getenv("ACESTEP_CONFIG_PATH", "acestep-v15-turbo-rl")
+                config_path = os.getenv("ACESTEP_CONFIG_PATH", "acestep-v15-turbo")
                 device = os.getenv("ACESTEP_DEVICE", "auto")
 
                 use_flash_attention = _env_bool("ACESTEP_USE_FLASH_ATTENTION", True)
@@ -868,7 +868,7 @@ def create_app() -> FastAPI:
 
                         project_root = _get_project_root()
                         checkpoint_dir = os.path.join(project_root, "checkpoints")
-                        lm_model_path = (req.lm_model_path or os.getenv("ACESTEP_LM_MODEL_PATH") or "acestep-5Hz-lm-0.6B-v3").strip()
+                        lm_model_path = (req.lm_model_path or os.getenv("ACESTEP_LM_MODEL_PATH") or "acestep-5Hz-lm-0.6B").strip()
                         backend = (req.lm_backend or os.getenv("ACESTEP_LM_BACKEND") or "vllm").strip().lower()
                         if backend not in {"vllm", "pt"}:
                             backend = "vllm"
@@ -1195,7 +1195,7 @@ def create_app() -> FastAPI:
                     return s
 
                 # Get model information
-                lm_model_name = os.getenv("ACESTEP_LM_MODEL_PATH", "acestep-5Hz-lm-0.6B-v3")
+                lm_model_name = os.getenv("ACESTEP_LM_MODEL_PATH", "acestep-5Hz-lm-0.6B")
                 # Use selected_model_name (set at the beginning of _run_one_job)
                 dit_model_name = selected_model_name
                 

docs/en/API.md

@@ -84,7 +84,7 @@ Suitable for passing only text parameters, or referencing audio file paths that
 
 | Parameter Name | Type | Default | Description |
 | :--- | :--- | :--- | :--- |
-| `model` | string | null | Select which DiT model to use (e.g., `"acestep-v15-turbo"`, `"acestep-v15-turbo-rl"`). Use `/v1/models` to list available models. If not specified, uses the default model. |
+| `model` | string | null | Select which DiT model to use (e.g., `"acestep-v15-turbo"`, `"acestep-v15-turbo-shift3"`). Use `/v1/models` to list available models. If not specified, uses the default model. |
 
 **thinking Semantics (Important)**:
 
@@ -148,7 +148,7 @@ These parameters control 5Hz LM sampling, used for metadata auto-completion and
 
 | Parameter Name | Type | Default | Description |
 | :--- | :--- | :--- | :--- |
-| `lm_model_path` | string | null | 5Hz LM checkpoint dir name (e.g. `acestep-5Hz-lm-0.6B-v3`) |
+| `lm_model_path` | string | null | 5Hz LM checkpoint dir name (e.g. `acestep-5Hz-lm-0.6B`) |
 | `lm_backend` | string | `"vllm"` | `vllm` or `pt` |
 | `lm_temperature` | float | `0.85` | Sampling temperature |
 | `lm_cfg_scale` | float | `2.5` | CFG scale (>1 enables CFG) |
@@ -258,7 +258,7 @@ curl -X POST http://localhost:8001/v1/music/generate \
   -H 'Content-Type: application/json' \
   -d '{
     "caption": "electronic dance music",
-    "model": "acestep-v15-turbo-rl",
+    "model": "acestep-v15-turbo",
     "thinking": true
   }'
 ```
@@ -382,8 +382,8 @@ The response contains basic task information, queue status, and final results.
     "keyscale": "C Major",
     "timesignature": "4",
     "genres": null,
-    "lm_model": "acestep-5Hz-lm-0.6B-v3",
-    "dit_model": "acestep-v15-turbo-rl"
+    "lm_model": "acestep-5Hz-lm-0.6B",
+    "dit_model": "acestep-v15-turbo"
   },
   "error": null
 }
@@ -441,15 +441,15 @@ Returns a list of available DiT models loaded on the server.
 {
   "models": [
     {
-      "name": "acestep-v15-turbo-rl",
+      "name": "acestep-v15-turbo",
       "is_default": true
     },
     {
-      "name": "acestep-v15-turbo",
+      "name": "acestep-v15-turbo-shift3",
       "is_default": false
     }
   ],
-  "default_model": "acestep-v15-turbo-rl"
+  "default_model": "acestep-v15-turbo"
 }
 ```
 
@@ -514,14 +514,14 @@ The API server can be configured using environment variables:
 | :--- | :--- | :--- |
 | `ACESTEP_API_HOST` | `127.0.0.1` | Server bind host |
 | `ACESTEP_API_PORT` | `8001` | Server bind port |
-| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo-rl` | Primary DiT model path |
+| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo` | Primary DiT model path |
 | `ACESTEP_CONFIG_PATH2` | (empty) | Secondary DiT model path (optional) |
 | `ACESTEP_CONFIG_PATH3` | (empty) | Third DiT model path (optional) |
 | `ACESTEP_DEVICE` | `auto` | Device for model loading |
 | `ACESTEP_USE_FLASH_ATTENTION` | `true` | Enable flash attention |
 | `ACESTEP_OFFLOAD_TO_CPU` | `false` | Offload models to CPU when idle |
 | `ACESTEP_OFFLOAD_DIT_TO_CPU` | `false` | Offload DiT specifically to CPU |
-| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B-v3` | Default 5Hz LM model |
+| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B` | Default 5Hz LM model |
 | `ACESTEP_LM_BACKEND` | `vllm` | LM backend (vllm or pt) |
 | `ACESTEP_LM_DEVICE` | (same as ACESTEP_DEVICE) | Device for LM |
 | `ACESTEP_LM_OFFLOAD_TO_CPU` | `false` | Offload LM to CPU |

docs/en/GRADIO_GUIDE.md

@@ -29,7 +29,7 @@ This guide provides comprehensive documentation for using the ACE-Step Gradio we
 python app.py
 
 # With pre-initialization
-python app.py --config acestep-v15-turbo-rl --init-llm
+python app.py --config acestep-v15-turbo --init-llm
 
 # With specific port
 python app.py --port 7860
@@ -55,14 +55,14 @@ The Gradio interface consists of several main sections:
 | Setting | Description |
 |---------|-------------|
 | **Checkpoint File** | Select a trained model checkpoint (if available) |
-| **Main Model Path** | Choose the DiT model configuration (e.g., `acestep-v15-turbo`, `acestep-v15-turbo-rl`) |
+| **Main Model Path** | Choose the DiT model configuration (e.g., `acestep-v15-turbo`, `acestep-v15-turbo-shift3`) |
 | **Device** | Processing device: `auto` (recommended), `cuda`, or `cpu` |
 
 ### 5Hz LM Configuration
 
 | Setting | Description |
 |---------|-------------|
-| **5Hz LM Model Path** | Select the language model (e.g., `acestep-5Hz-lm-0.6B`, `acestep-5Hz-lm-0.6B-v3`) |
+| **5Hz LM Model Path** | Select the language model (e.g., `acestep-5Hz-lm-0.6B`, `acestep-5Hz-lm-1.7B`) |
 | **5Hz LM Backend** | `vllm` (faster, recommended) or `pt` (PyTorch, more compatible) |
 | **Initialize 5Hz LM** | Check to load the LM during initialization (required for thinking mode) |
 
@@ -477,7 +477,7 @@ After training, export the final adapter:
 
 ### For Faster Generation
 
-1. **Use turbo model** - Select `acestep-v15-turbo` or `acestep-v15-turbo-rl`
+1. **Use turbo model** - Select `acestep-v15-turbo` or `acestep-v15-turbo-shift3`
 2. **Keep inference steps at 8** - Default is optimal for turbo
 3. **Reduce batch size** - Lower batch size if you need quick results
 4. **Disable AutoGen** - Manual control over batch generation

docs/en/INFERENCE.md

@@ -35,13 +35,13 @@ llm_handler = LLMHandler()
 # Initialize services
 dit_handler.initialize_service(
     project_root="/path/to/project",
-    config_path="acestep-v15-turbo-rl",
+    config_path="acestep-v15-turbo",
     device="cuda"
 )
 
 llm_handler.initialize(
     checkpoint_dir="/path/to/checkpoints",
-    lm_model_path="acestep-5Hz-lm-0.6B-v3",
+    lm_model_path="acestep-5Hz-lm-0.6B",
     backend="vllm",
     device="cuda"
 )

docs/ja/API.md

@@ -84,7 +84,7 @@ APIはほとんどのパラメータで **snake_case** と **camelCase** の両
 
 | パラメータ名 | 型 | デフォルト | 説明 |
 | :--- | :--- | :--- | :--- |
-| `model` | string | null | 使用するDiTモデルを選択（例：`"acestep-v15-turbo"`、`"acestep-v15-turbo-rl"`）。`/v1/models` で利用可能なモデルを一覧表示。指定しない場合はデフォルトモデルを使用。|
+| `model` | string | null | 使用するDiTモデルを選択（例：`"acestep-v15-turbo"`、`"acestep-v15-turbo-shift3"`）。`/v1/models` で利用可能なモデルを一覧表示。指定しない場合はデフォルトモデルを使用。|
 
 **thinkingのセマンティクス（重要）**：
 
@@ -148,7 +148,7 @@ APIはほとんどのパラメータで **snake_case** と **camelCase** の両
 
 | パラメータ名 | 型 | デフォルト | 説明 |
 | :--- | :--- | :--- | :--- |
-| `lm_model_path` | string | null | 5Hz LMチェックポイントディレクトリ名（例：`acestep-5Hz-lm-0.6B-v3`）|
+| `lm_model_path` | string | null | 5Hz LMチェックポイントディレクトリ名（例：`acestep-5Hz-lm-0.6B`）|
 | `lm_backend` | string | `"vllm"` | `vllm` または `pt` |
 | `lm_temperature` | float | `0.85` | サンプリング温度 |
 | `lm_cfg_scale` | float | `2.5` | CFGスケール（>1でCFGを有効化）|
@@ -258,7 +258,7 @@ curl -X POST http://localhost:8001/v1/music/generate \
   -H 'Content-Type: application/json' \
   -d '{
     "caption": "エレクトロニックダンスミュージック",
-    "model": "acestep-v15-turbo-rl",
+    "model": "acestep-v15-turbo",
     "thinking": true
   }'
 ```
@@ -382,8 +382,8 @@ curl -X POST http://localhost:8001/v1/music/generate \
     "keyscale": "C Major",
     "timesignature": "4",
     "genres": null,
-    "lm_model": "acestep-5Hz-lm-0.6B-v3",
-    "dit_model": "acestep-v15-turbo-rl"
+    "lm_model": "acestep-5Hz-lm-0.6B",
+    "dit_model": "acestep-v15-turbo"
   },
   "error": null
 }
@@ -441,15 +441,15 @@ curl -X POST http://localhost:8001/v1/music/random \
 {
   "models": [
     {
-      "name": "acestep-v15-turbo-rl",
+      "name": "acestep-v15-turbo",
       "is_default": true
     },
     {
-      "name": "acestep-v15-turbo",
+      "name": "acestep-v15-turbo-shift3",
       "is_default": false
     }
   ],
-  "default_model": "acestep-v15-turbo-rl"
+  "default_model": "acestep-v15-turbo"
 }
 ```
 
@@ -514,14 +514,14 @@ APIサーバーは環境変数で設定できます：
 | :--- | :--- | :--- |
 | `ACESTEP_API_HOST` | `127.0.0.1` | サーバーバインドホスト |
 | `ACESTEP_API_PORT` | `8001` | サーバーバインドポート |
-| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo-rl` | プライマリDiTモデルパス |
+| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo` | プライマリDiTモデルパス |
 | `ACESTEP_CONFIG_PATH2` | （空）| セカンダリDiTモデルパス（オプション）|
 | `ACESTEP_CONFIG_PATH3` | （空）| 3番目のDiTモデルパス（オプション）|
 | `ACESTEP_DEVICE` | `auto` | モデルロードデバイス |
 | `ACESTEP_USE_FLASH_ATTENTION` | `true` | flash attentionを有効化 |
 | `ACESTEP_OFFLOAD_TO_CPU` | `false` | アイドル時にモデルをCPUにオフロード |
 | `ACESTEP_OFFLOAD_DIT_TO_CPU` | `false` | DiTを特にCPUにオフロード |
-| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B-v3` | デフォルト5Hz LMモデル |
+| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B` | デフォルト5Hz LMモデル |
 | `ACESTEP_LM_BACKEND` | `vllm` | LMバックエンド（vllmまたはpt）|
 | `ACESTEP_LM_DEVICE` | （ACESTEP_DEVICEと同じ）| LMデバイス |
 | `ACESTEP_LM_OFFLOAD_TO_CPU` | `false` | LMをCPUにオフロード |

docs/ja/GRADIO_GUIDE.md

@@ -29,7 +29,7 @@
 python app.py
 
 # 事前初期化付き
-python app.py --config acestep-v15-turbo-rl --init-llm
+python app.py --config acestep-v15-turbo --init-llm
 
 # 特定のポートで
 python app.py --port 7860
@@ -55,14 +55,14 @@ Gradioインターフェースは以下の主要セクションで構成され
 | 設定 | 説明 |
 |---------|-------------|
 | **チェックポイントファイル** | トレーニング済みモデルチェックポイントを選択（利用可能な場合）|
-| **メインモデルパス** | DiTモデル設定を選択（例：`acestep-v15-turbo`、`acestep-v15-turbo-rl`）|
+| **メインモデルパス** | DiTモデル設定を選択（例：`acestep-v15-turbo`、`acestep-v15-turbo-shift3`）|
 | **デバイス** | 処理デバイス：`auto`（推奨）、`cuda`、または `cpu` |
 
 ### 5Hz LM設定
 
 | 設定 | 説明 |
 |---------|-------------|
-| **5Hz LMモデルパス** | 言語モデルを選択（例：`acestep-5Hz-lm-0.6B`、`acestep-5Hz-lm-0.6B-v3`）|
+| **5Hz LMモデルパス** | 言語モデルを選択（例：`acestep-5Hz-lm-0.6B`、`acestep-5Hz-lm-1.7B`）|
 | **5Hz LMバックエンド** | `vllm`（より高速、推奨）または `pt`（PyTorch、互換性が高い）|
 | **5Hz LMを初期化** | 初期化時にLMを読み込むためにチェック（thinkingモードに必要）|
 
@@ -477,7 +477,7 @@ LoRAトレーニングタブはカスタムLoRAアダプターを作成するた
 
 ### より高速な生成のために
 
-1. **turboモデルを使用** - `acestep-v15-turbo` または `acestep-v15-turbo-rl` を選択
+1. **turboモデルを使用** - `acestep-v15-turbo` または `acestep-v15-turbo-shift3` を選択
 2. **推論ステップを8に保つ** - turboに最適なデフォルト
 3. **バッチサイズを減らす** - 迅速な結果が必要な場合はバッチサイズを下げる
 4. **AutoGenを無効化** - バッチ生成の手動制御

docs/ja/INFERENCE.md

@@ -35,13 +35,13 @@ llm_handler = LLMHandler()
 # サービスの初期化
 dit_handler.initialize_service(
     project_root="/path/to/project",
-    config_path="acestep-v15-turbo-rl",
+    config_path="acestep-v15-turbo",
     device="cuda"
 )
 
 llm_handler.initialize(
     checkpoint_dir="/path/to/checkpoints",
-    lm_model_path="acestep-5Hz-lm-0.6B-v3",
+    lm_model_path="acestep-5Hz-lm-0.6B",
     backend="vllm",
     device="cuda"
 )

docs/zh/API.md

@@ -84,7 +84,7 @@ API 支持大多数参数的 **snake_case** 和 **camelCase** 命名。例如：
 
 | 参数名 | 类型 | 默认值 | 说明 |
 | :--- | :--- | :--- | :--- |
-| `model` | string | null | 选择使用哪个 DiT 模型（例如 `"acestep-v15-turbo"`、`"acestep-v15-turbo-rl"`）。使用 `/v1/models` 列出可用模型。如果未指定，使用默认模型。|
+| `model` | string | null | 选择使用哪个 DiT 模型（例如 `"acestep-v15-turbo"`、`"acestep-v15-turbo-shift3"`）。使用 `/v1/models` 列出可用模型。如果未指定，使用默认模型。|
 
 **thinking 语义（重要）**：
 
@@ -148,7 +148,7 @@ API 支持大多数参数的 **snake_case** 和 **camelCase** 命名。例如：
 
 | 参数名 | 类型 | 默认值 | 说明 |
 | :--- | :--- | :--- | :--- |
-| `lm_model_path` | string | null | 5Hz LM 检查点目录名（例如 `acestep-5Hz-lm-0.6B-v3`）|
+| `lm_model_path` | string | null | 5Hz LM 检查点目录名（例如 `acestep-5Hz-lm-0.6B`）|
 | `lm_backend` | string | `"vllm"` | `vllm` 或 `pt` |
 | `lm_temperature` | float | `0.85` | 采样温度 |
 | `lm_cfg_scale` | float | `2.5` | CFG 比例（>1 启用 CFG）|
@@ -258,7 +258,7 @@ curl -X POST http://localhost:8001/v1/music/generate \
   -H 'Content-Type: application/json' \
   -d '{
     "caption": "电子舞曲",
-    "model": "acestep-v15-turbo-rl",
+    "model": "acestep-v15-turbo",
     "thinking": true
   }'
 ```
@@ -382,8 +382,8 @@ curl -X POST http://localhost:8001/v1/music/generate \
     "keyscale": "C Major",
     "timesignature": "4",
     "genres": null,
-    "lm_model": "acestep-5Hz-lm-0.6B-v3",
-    "dit_model": "acestep-v15-turbo-rl"
+    "lm_model": "acestep-5Hz-lm-0.6B",
+    "dit_model": "acestep-v15-turbo"
   },
   "error": null
 }
@@ -441,15 +441,15 @@ curl -X POST http://localhost:8001/v1/music/random \
 {
   "models": [
     {
-      "name": "acestep-v15-turbo-rl",
+      "name": "acestep-v15-turbo",
       "is_default": true
     },
     {
-      "name": "acestep-v15-turbo",
+      "name": "acestep-v15-turbo-shift3",
       "is_default": false
     }
   ],
-  "default_model": "acestep-v15-turbo-rl"
+  "default_model": "acestep-v15-turbo"
 }
 ```
 
@@ -514,14 +514,14 @@ API 服务器可以通过环境变量进行配置：
 | :--- | :--- | :--- |
 | `ACESTEP_API_HOST` | `127.0.0.1` | 服务器绑定主机 |
 | `ACESTEP_API_PORT` | `8001` | 服务器绑定端口 |
-| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo-rl` | 主 DiT 模型路径 |
+| `ACESTEP_CONFIG_PATH` | `acestep-v15-turbo` | 主 DiT 模型路径 |
 | `ACESTEP_CONFIG_PATH2` | （空）| 辅助 DiT 模型路径（可选）|
 | `ACESTEP_CONFIG_PATH3` | （空）| 第三个 DiT 模型路径（可选）|
 | `ACESTEP_DEVICE` | `auto` | 模型加载设备 |
 | `ACESTEP_USE_FLASH_ATTENTION` | `true` | 启用 flash attention |
 | `ACESTEP_OFFLOAD_TO_CPU` | `false` | 空闲时将模型卸载到 CPU |
 | `ACESTEP_OFFLOAD_DIT_TO_CPU` | `false` | 专门将 DiT 卸载到 CPU |
-| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B-v3` | 默认 5Hz LM 模型 |
+| `ACESTEP_LM_MODEL_PATH` | `acestep-5Hz-lm-0.6B` | 默认 5Hz LM 模型 |
 | `ACESTEP_LM_BACKEND` | `vllm` | LM 后端（vllm 或 pt）|
 | `ACESTEP_LM_DEVICE` | （与 ACESTEP_DEVICE 相同）| LM 设备 |
 | `ACESTEP_LM_OFFLOAD_TO_CPU` | `false` | 将 LM 卸载到 CPU |

docs/zh/GRADIO_GUIDE.md

@@ -29,7 +29,7 @@
 python app.py
 
 # 预初始化
-python app.py --config acestep-v15-turbo-rl --init-llm
+python app.py --config acestep-v15-turbo --init-llm
 
 # 指定端口
 python app.py --port 7860
@@ -55,14 +55,14 @@ Gradio 界面包含以下主要部分：
 | 设置 | 说明 |
 |---------|-------------|
 | **检查点文件** | 选择已训练的模型检查点（如果可用）|
-| **主模型路径** | 选择 DiT 模型配置（例如 `acestep-v15-turbo`、`acestep-v15-turbo-rl`）|
+| **主模型路径** | 选择 DiT 模型配置（例如 `acestep-v15-turbo`、`acestep-v15-turbo-shift3`）|
 | **设备** | 处理设备：`auto`（推荐）、`cuda` 或 `cpu` |
 
 ### 5Hz LM 配置
 
 | 设置 | 说明 |
 |---------|-------------|
-| **5Hz LM 模型路径** | 选择语言模型（例如 `acestep-5Hz-lm-0.6B`、`acestep-5Hz-lm-0.6B-v3`）|
+| **5Hz LM 模型路径** | 选择语言模型（例如 `acestep-5Hz-lm-0.6B`、`acestep-5Hz-lm-1.7B`）|
 | **5Hz LM 后端** | `vllm`（更快，推荐）或 `pt`（PyTorch，兼容性更好）|
 | **初始化 5Hz LM** | 勾选以在初始化期间加载 LM（thinking 模式必需）|
 
@@ -477,7 +477,7 @@ LoRA 训练选项卡提供创建自定义 LoRA 适配器的工具。
 
 ### 加快生成速度
 
-1. **使用 turbo 模型** - 选择 `acestep-v15-turbo` 或 `acestep-v15-turbo-rl`
+1. **使用 turbo 模型** - 选择 `acestep-v15-turbo` 或 `acestep-v15-turbo-shift3`
 2. **保持推理步数为 8** - 这是 turbo 的最佳默认值
 3. **减少批量大小** - 如果需要快速结果，降低批量大小
 4. **禁用 AutoGen** - 手动控制批次生成

docs/zh/INFERENCE.md

@@ -35,13 +35,13 @@ llm_handler = LLMHandler()
 # 初始化服务
 dit_handler.initialize_service(
     project_root="/path/to/project",
-    config_path="acestep-v15-turbo-rl",
+    config_path="acestep-v15-turbo",
     device="cuda"
 )
 
 llm_handler.initialize(
     checkpoint_dir="/path/to/checkpoints",
-    lm_model_path="acestep-5Hz-lm-0.6B-v3",
+    lm_model_path="acestep-5Hz-lm-0.6B",
     backend="vllm",
     device="cuda"
 )

generate_examples.py

@@ -39,8 +39,8 @@ def generate_examples(num_examples=50, output_dir="examples/text2music", start_i
         logger.error("No 5Hz LM models found in checkpoints directory")
         return
     
-    # Prefer acestep-5Hz-lm-0.6B-v3 if available
-    lm_model = "acestep-5Hz-lm-0.6B-v3" if "acestep-5Hz-lm-0.6B-v3" in available_models else available_models[0]
+    # Prefer acestep-5Hz-lm-0.6B if available
+    lm_model = "acestep-5Hz-lm-0.6B" if "acestep-5Hz-lm-0.6B" in available_models else available_models[0]
     logger.info(f"Using LM model: {lm_model}")
     
     # Initialize LM

profile_inference.py

@@ -40,8 +40,8 @@ if project_root not in sys.path:
 def load_env_config():
     """从 .env 文件加载配置"""
     env_config = {
-        'ACESTEP_CONFIG_PATH': 'acestep-v15-turbo-rl',
-        'ACESTEP_LM_MODEL_PATH': 'acestep-5Hz-lm-0.6B-v3',
+        'ACESTEP_CONFIG_PATH': 'acestep-v15-turbo',
+        'ACESTEP_LM_MODEL_PATH': 'acestep-5Hz-lm-0.6B',
         'ACESTEP_DEVICE': 'auto',
         'ACESTEP_LM_BACKEND': 'vllm',
     }

skills/acemusic/SKILL.md

@@ -250,7 +250,7 @@ project_root/
     "bpm": 120,
     "keyscale": "C Major",
     "duration": 60.0,
-    "dit_model": "acestep-v15-turbo-rl"
+    "dit_model": "acestep-v15-turbo"
   }
 }
 ```