Audio Noise Removal Service - All-in-One

AI-powered audio noise removal service with Swagger API documentation. Built on ResembleAI's DeepFilterNet model.

🚀 Quick Start

# Pull the all-in-one image
docker pull neosun/noise-removal:v1.0-allinone

# Run the service
docker run -d \
  --name noise-removal \
  --gpus all \
  -p 5080:5080 \
  neosun/noise-removal:v1.0-allinone

# Wait 30 seconds for service to start
# Access Swagger UI: http://0.0.0.0:5080/docs/

That's it! No build required, no external dependencies, everything is included.

📚 API Documentation

• Swagger UI: http://0.0.0.0:5080/docs/
• API Spec: http://0.0.0.0:5080/swagger.json
• Health Check: http://0.0.0.0:5080/health

🎯 Features

• ✅ All-in-One Image: Pre-loaded DeepFilterNet model (9.25GB)
• ✅ Zero Configuration: Works out of the box
• ✅ Swagger Documentation: Interactive API docs
• ✅ GPU Accelerated: CUDA 12.4 support
• ✅ Async Processing: Support for long audio files
• ✅ Health Monitoring: Built-in health check

🛠️ Requirements

Hardware

• GPU: NVIDIA GPU with 4GB+ VRAM
• RAM: 8GB+ recommended

Software

• NVIDIA Driver: 525.60.13+
• Docker: 20.10+
• NVIDIA Container Toolkit: Required

📋 API Usage

Synchronous Processing

curl -X POST http://0.0.0.0:5080/api \
  -F "audio=@noisy_audio.wav" \
  -o clean_audio.wav

Python Example

import requests

url = "http://0.0.0.0:5080/api"
files = {"audio": open("noisy_audio.wav", "rb")}
response = requests.post(url, files=files)

with open("clean_audio.wav", "wb") as f:
    f.write(response.content)

🎨 Supported Formats

WAV, MP3, FLAC, OGG, M4A

📊 Performance

• Speed: ~10x faster than real-time on L40S GPU
• Memory: ~2GB VRAM per request
• Startup: <5 seconds

📄 License

MIT License

---

Docker Hub: https://hub.docker.com/r/neosun/noise-removal