README.md

---
title: StableVITON Virtual Try-On
emoji: 👕
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
license: mit
app_port: 7860
version: 1.0.1
---

# StableVITON Virtual Try-On Backend

AI-powered virtual try-on service using StableVITON and FastAPI.

## Features

- Virtual try-on inference via REST API
- Support for JPEG/PNG images
- Automatic image preprocessing and validation
- GPU-optimized inference (CUDA support)
- Single-request processing to prevent OOM
- Comprehensive error handling

## API Documentation

### POST /tryon

Perform virtual try-on inference.

**Request:**
- `person_image`: Full-body photo (multipart/form-data)
- `garment_image`: Garment image (multipart/form-data)

**Response:**
```json
{
  "success": true,
  "result_image": "data:image/png;base64,...",
  "processing_time": 23.4,
  "model_version": "stablevton-v1"
}
```

### GET /health

Health check endpoint.

## Local Development

```bash
# Install dependencies
pip install -r requirements.txt

# Run server
python main.py
```

## Deployment

This project is designed to run on Hugging Face Spaces with Docker SDK.

1. Create a Space with Docker SDK
2. Select GPU hardware (T4 small is sufficient)
3. Push the code to the Space