⚡ RLD UVR-Opt: Ultimate Vocal Remover GUI (Optimized)
High-performance, memory-optimized distribution of Ultimate Vocal Remover GUI (UVR v5.6).
Focused on accelerated ONNX Runtime inference, massive VRAM reduction, and instant CLI startup.
🚀 Up to ~4x Faster Inference | ~53% Lower VRAM Usage
Prerequisites & Installation \cdot Usage & Quick Start \cdot Key Features \cdot Verified Benchmarks \cdot Roadmap \cdot Credits
📦 Installation & Prerequisites
1. System Requirements & Platform Support
Platform Support: RLD UVR-Opt is actively tested and optimized exclusively for Windows 10 / 11 (64-bit).
- Python: Python 3.11 or higher (64-bit) is required.
- Compute Device:
- NVIDIA GPU (CUDA): NVIDIA GeForce RTX 20 Series GPU or higher with at least 6GB VRAM is the minimum requirement for GPU conversions (8GB+ VRAM recommended for optimal performance).
- CPU: Multi-core modern 64-bit x86 processor (fully supported via multi-threaded CPU inference).
- (Note: Hardware focus is strictly on CUDA and CPU).
2. Model & Architecture Scope
ONNX Model Priority:
RLD UVR-Opt's core acceleration pipeline is built around the MDX-Net family based on the ONNX runtime (.onnx).
While legacy PyTorch checkpoint formats may load, non-.onnxmodels are currently untested / experimental and may not function properly. For maximum performance and stability, use.onnxmodels.
3. ⚡ One-Click Automated Installation (Windows)
RLD UVR-Opt provides an automated setup script that handles your entire environment setup from start to finish:
Step 1: Clone or Download the Repository
git clone https://github.com/revlaydust/rld-uvr-opt.git
cd rld-uvr-opt
(Or Download the ZIP and extract it to a folder).
Step 2: Run the Automated Installer
Simply double-click install\install_packages.bat in File Explorer, or run in terminal:
.\install\install_packages.bat
What the One-Click Installer automates for you:
- 🔍 Hardware & Driver Detection: Automatically identifies your GPU architecture (NVIDIA RTX 20/30/40/50 series) and selects PyTorch CUDA 12.8 or optimized CPU fallback.
- 🐍 Python Environment Management: Automatically bootstraps
uvand provisions an isolated Python 3.11 virtual environment (venv). - 📦 Fast Dependency Installation: Pulls and configures all required packages, including PyTorch and ONNX Runtime GPU/CPU.
- 🎬 Automated Media Binaries: Detects existing system installations of FFmpeg and rubberband-cli. If either is missing, it automatically downloads and extracts portable binaries into the application folder—zero manual downloading required!
- 🧪 Smoke Test Verification: Runs an immediate runtime sanity check on tensor operations and ONNX execution providers before launching.
🛠️ Manual Installation via Astral uv (Click to Expand for Advanced Users)
If you prefer to configure your environment manually instead of running install\install_packages.bat, use Astral uv for lightning-fast setup:
-
Install Astral uv (if not already installed):
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" -
Create Python 3.11 Virtual Environment via uv:
uv venv venv --python 3.11 .\venv\Scripts\activate -
Install PyTorch & ONNX Runtime with uv pip:
- For NVIDIA GPU (CUDA 12.8):
uv pip install torch==2.11.0 torchvision==0.26.0 torchaudio==2.11.0 --index-url https://download.pytorch.org/whl/cu128 uv pip install onnxruntime-gpu==1.22.0 - For CPU Only Fallback:
uv pip install torch==2.11.0 torchvision==0.26.0 torchaudio==2.11.0 --index-url https://download.pytorch.org/whl/cpu uv pip install onnxruntime==1.22.0
- For NVIDIA GPU (CUDA 12.8):
-
Install Base Dependencies with uv pip:
uv pip install -r .\install\requirements.txt -
Place External Media Binaries:
- Place
ffmpeg.exein the application root folder (or ensure it is in your systemPATH). - Place
rubberband.exeandsndfile.dllin the application root folder andlib_v5/.
- Place
🚀 Usage / Quick Start
Get from zero to audio separation in seconds:
A. Launch the GUI
Simply run the included launcher batch file or start via Python:
.\run_uvr.bat
Or via terminal:
python UVR.py
B. UVR Command-Line Interface (UVR-CLI)
RLD UVR-Opt provides a dedicated, lightweight CLI designed for automation and headless environments:
# 1. View all available commands instantly
uvr_cli --help
# 2. Run high-performance separation with MDX-Net Inst_HQ_4
uvr_cli run --audio "path/to/song.wav" --model "UVR-MDX-NET-Inst_HQ_4.onnx" --output "output_stems"
# 3. High-throughput separation with custom overlap and segment size
uvr_cli run -a "song.wav" -m "UVR-MDX-NET-Inst_HQ_4.onnx" --overlap 0.50 --segment-size 4000 -o "output_stems"
✨ Key Features
- ⚡ Accelerated ONNX Inference Loop: Re-engineered tensor pipelines and pre-allocated contiguous buffers eliminate redundant Python-level copies during chunk processing.
- 📉 Zero-Disk In-Memory FP16: Dynamic model FP16 precision downcasting happens entirely in system RAM. UVR no longer writes or depends on persistent
.fp16.onnxfiles on disk. - 🧹 Aggressive Memory Hygiene: Immediate session eviction on model switch releases Python references and clears CUDA allocations (
torch.cuda.empty_cache()), eliminating VRAM bloat across sequential jobs. - ⚡ Sub-Second UVR-CLI (PEP 562): Proxy-based lazy module loading cuts CLI startup latency from ~9 seconds down to ~0.20 seconds.
- 🎯 Truthful Benchmark Telemetry: Pure model inference time is isolated from module loading and audio I/O for accurate, reproducible performance measurements.
- 🔒 100% Bit-Exact Quality: Mathematically identical STFT processing and model execution ensuring zero quality degradation compared to original UVR v5.6.
📊 Verified Benchmarks
All metrics were gathered using the built-in isolated benchmark suite on identical test audio (48.34s stereo WAV, model: UVR-MDX-NET-Inst_HQ_4.onnx, Ultra Quality FP32).
Test Hardware: NVIDIA GeForce RTX 5060 Ti | AMD Ryzen 5 5600 6-Core Processor | 32 GB RAM | Windows 11
1. GPU Inference Performance (CUDA)
| Configuration | Metric | UVR v5.6 (Legacy) | RLD UVR-Opt | Improvement |
|---|---|---|---|---|
| Overlap 0.75 (Segment: 4000) |
Inference Time Processing Speed Peak Device VRAM PyTorch VRAM Alloc |
13.52 s 3.58x Realtime 8,210 MB 6,360 MB |
3.85 s 12.54x Realtime 3,941 MB 177.5 MB |
~3.5x Faster +250% Throughput -52.0% VRAM Usage (4.2 GB Saved) -97.2% PyTorch Memory Bloat |
| Overlap 0.50 (Segment: 4000) |
Inference Time Processing Speed Peak Device VRAM |
7.99 s 6.05x Realtime 8,214 MB |
2.06 s 23.44x Realtime 3,893 MB |
~3.87x Faster +287% Throughput -52.6% VRAM Usage (4.3 GB Saved) |
2. CPU Inference Performance
| Configuration | Metric | UVR v5.6 (Legacy) | RLD UVR-Opt | Improvement |
|---|---|---|---|---|
| Overlap 0.50 (Segment: 4000) |
Inference Time Processing Speed Peak System RAM |
79.84 s 0.61x Realtime 8,928 MB |
32.17 s 1.50x Realtime 3,210 MB |
~2.48x Faster +146% Throughput -64.0% RAM Usage (5.7 GB Saved) |
Note
Raw per-round session logs and metric datasets are preserved in
benchmark_results/(summary.csvandsummary_legacy.csv).
🗺️ Roadmap
- Accelerated ONNX inference pipeline and buffer optimizations.
- In-memory
FP16downcasting (e.g.,FP32\rightarrowFP16) and cache lifecycle management. - In-memory
BF16/FP8(E4M3/E5M2) downcasting. - Windows ML (via ONNX Runtime): Planned modern hardware acceleration backend for Windows as the official replacement for legacy DirectML.
- TensorRT Execution Provider (via ONNX Runtime): Ultra-high performance Execution Provider for NVIDIA GPUs. (Note: CUDA EP will remain as the primary, most stable, and balanced execution provider).
👥 Credits & Acknowledgments
RLD UVR-Opt Maintainer
- RevLayDust — Core architecture optimizations, ONNX pipeline acceleration, CLI engine, and memory lifecycle management.
Core UVR Developers
Special Thanks & Upstream Contributors
- ZFTurbo — Created & trained weights for MDX23C models.
- DilanBoskan — Essential contributions at the start of the UVR project.
- Bas Curtiz — Designed the official UVR logo, icon, banner, and splash screen.
- tsurumeso — Developed the original VR Architecture code.
- Kuielab & Woosung Choi — Developed the original MDX-Net AI code.
- Adefossez & Demucs — Core developer of Facebook's Demucs Music Source Separation.
- KimberleyJSN — Advised and aided implementation of training scripts for MDX-Net and Demucs.
- Hv — Helped implement chunk processing into MDX-Net AI code.
- Audio Separation & CC Karaoke Discord Communities — For continuous testing, feedback, and support.
📄 License & References
The Ultimate Vocal Remover GUI and RLD UVR-Opt codebase is open-source and released under the MIT License.
Copyright (c) 2022 Ultimate Vocal Remover
Copyright (c) 2026 RevLayDust
Academic References
- [1] Takahashi et al., "Multi-scale Multi-band DenseNets for Audio Source Separation", arXiv:1706.09588