image-to-rigged-3d / README.md
JSCPPProgrammer's picture
Image -> 3D (TRELLIS.2) -> rigged GLB/FBX (TokenRig) pipeline
3face62 verified
|
Raw
History Blame Contribute Delete
1.19 kB

A newer version of the Gradio SDK is available: 6.20.0

Upgrade
metadata
title: Image to Rigged 3D Model
emoji: 🦴
colorFrom: purple
colorTo: indigo
sdk: gradio
sdk_version: 6.12.0
python_version: '3.12'
app_file: app.py
startup_duration_timeout: 1h
pinned: false
license: mit
short_description: Image to 3D to auto-rigged GLB/FBX for Blender
models:
  - microsoft/TRELLIS.2-4B
  - VAST-AI/SkinTokens
  - Qwen/Qwen3-0.6B

Image β†’ 3D β†’ Rigged Model

Turn a single image into a fully rigged 3D model ready for Blender:

  1. Image β†’ 3D: TRELLIS.2-4B (Microsoft) generates a high-fidelity textured PBR mesh from your image.
  2. 3D β†’ Rigged: SkinTokens / TokenRig (VAST AI Research, successor to UniRig) automatically predicts a skeleton hierarchy and per-vertex skinning weights.
  3. Export: Download the rigged model as GLB or FBX and import it straight into Blender (File β†’ Import β†’ glTF 2.0 / FBX) to pose and animate.

Built on the official microsoft/TRELLIS.2 and VAST-AI/SkinTokens Spaces. Both models are MIT licensed.