Spaces:
Build error
Build error
| title: LGM Tiny | |
| emoji: 🦀 | |
| colorFrom: red | |
| colorTo: green | |
| sdk: gradio | |
| sdk_version: 6.15.1 | |
| app_file: app.py | |
| pinned: false | |
| license: mit | |
| # 🐙 WasabiOctopus / LGM Tiny | |
| A clean and lightweight Hugging Face Space for **single-image to 3D Gaussian asset generation** using [`WasabiOctopus/LGM`](https://huggingface.co/WasabiOctopus/LGM). | |
| ## ✨ What it does | |
| Upload a single object image, and the Space generates a `.ply` 3D Gaussian asset that can be previewed directly in the browser. | |
| ## 🚀 Best input images | |
| For better results, use images with: | |
| - One centered object | |
| - Clean background | |
| - Clear silhouette | |
| - Good lighting | |
| - Minimal occlusion | |
| Transparent PNG images are supported. | |
| ## ⚠️ Limitations | |
| This is a research-oriented 3D generation demo. It may struggle with: | |
| - Thin structures | |
| - Transparent objects | |
| - Hair, fur, wires, or complex topology | |
| - Strong reflections | |
| - Ambiguous back-side geometry | |
| - Production-ready topology or rigging requirements | |
| ## 🔗 Links | |
| - Model: https://huggingface.co/WasabiOctopus/LGM | |
| - Original paper: https://arxiv.org/abs/2402.05054 | |
| - Original LGM repository: https://github.com/3DTopia/LGM | |
| - Upstream full pipeline: https://huggingface.co/dylanebert/LGM-full |