LGM — Large Gaussian Model for Image-to-3D

This is a capstone project model for the Hugging Face ML for 3D Course.

Model Description

LGM (Large Gaussian Model) is a generative 3D model that converts single images to 3D Gaussian Splatting representations. It uses a multi-view diffusion approach to generate novel views of an object from a single input image, then reconstructs a 3D Gaussian Splatting representation.

This model was duplicated from dylanebert/LGM-full as part of the ML for 3D Course capstone project.

Usage

from diffusers import DiffusionPipeline

pipe = DiffusionPipeline.from_pretrained("ISLAM-PO/LGM")
image = pipe("path/to/image.png")

Capstone Project

• Model: https://huggingface.co/ISLAM-PO/LGM
• Space (Demo): https://huggingface.co/spaces/ISLAM-PO/image-to-3d-demo
• Course: ML for 3D Course (https://huggingface.co/learn/ml-for-3d-course)

License

MIT License. See dylanebert/LGM-full for original license.