README.md

---
title: Zero123++ Inference
emoji: 🏛️
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
---

# Zero123++ Inference API

This Space takes one input image and generates a Zero123++ 6-view grid.

It then crops views 1, 3, 5, and 6 into separate 256x256 white-background images for the LGM model.

## Endpoints

### Health check
GET `/health`

### Generate views
POST `/generate`

Form-data:
- `file`: input image
- `steps`: optional, default 75

Returns:
- `multiview_grid.png`
- `view_1_front_right.png`
- `view_3_right.png`
- `view_5_back_right.png`
- `view_6_back_left.png`