trellis-3d-generation

Convert 2D images to 3D models using Microsoft TRELLIS structured latents.

Updated Jun 4, 2026
One-click install
npx skills add https://github.com/kngender5/hermes --skill trellis-3d-generation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: trellis-3d-generation
Source: https://github.com/kngender5/hermes/tree/main/skills/creative/trellis-3d-generation
Command: npx skills add https://github.com/kngender5/hermes --skill trellis-3d-generation

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch, torchvision, diffusers, transformers, accelerate, safetensors, rembg, onnxruntime, trimesh, open3d, spconv-cu120, imageio, imageio-ffmpeg, scipy, ninja, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill addresses the challenge of converting 2D images into 3D models using Microsoft TRELLIS, offering a solution for image-to-3D conversion, 3D reconstruction from photos, and TRELLIS pipeline integration.

Core Features & Use Cases

  • Image-to-3D Conversion: Transform single or multi-view images into 3D models.
  • 3D Reconstruction: Generate 3D models from photo sets.
  • TRELLIS Pipeline Integration: Use with Microsoft TRELLIS for structured 3D latents.
  • Use Case: For a product designer who needs to create a 3D model from a product photo, this skill streamlines the process from image input to 3D model output.

Quick Start

Run the following command to generate a 3D model from an image: python generate_3d_model.py input_image.jpg output_model.glb

Frequently Asked Questions about trellis-3d-generation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert a 2D image to a 3D model?

To convert a 2D image to a 3D model, use the provided script to process an input image and generate a 3D output file. You can run `python generate_3d_model.py input_image.jpg output_model.glb` to execute the image-to-3D transformation using Microsoft TRELLIS.

Can I use Microsoft TRELLIS for 3D reconstruction from multiple photos?

Yes, you can use Microsoft TRELLIS for 3D reconstruction from multiple photos. This Skill supports transforming single or multi-view images into 3D models by leveraging structured 3D latents to build the final geometry from your photo sets.

What dependencies do I need for image-to-3D conversion with Microsoft TRELLIS?

Image-to-3D conversion with Microsoft TRELLIS requires dependencies including PyTorch, Diffusers, Transformers, and Spconv. You also need image processing libraries like Rembg and OnnxRuntime, plus 3D mesh handling tools like Trimesh and Open3D.

Does this image-to-3D pipeline support exporting to GLB format?

Yes, the image-to-3D pipeline supports exporting to GLB format. The quick start command explicitly demonstrates outputting a 3D model to a `.glb` file, utilizing Trimesh and Open3D to handle the generated 3D structure.

What's the best way to generate a 3D model from a product photo?

The best way to generate a 3D model from a product photo is to use this Skill's TRELLIS pipeline integration. It streamlines the process from image input directly to a 3D model output, specifically addressing the needs of product designers.

Why do I need background removal tools like Rembg for 3D reconstruction?

Background removal tools like Rembg are required for 3D reconstruction to isolate the target object from its environment. Processing a clean subject allows Microsoft TRELLIS to accurately calculate structured 3D latents and generate precise geometry.