OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Materialanything

Materialanything is a diffusion-based model and Python code repository for generating physically based rendering materials for 3D meshes, with documented inputs including texture-less, albedo-only, generated, and scanned objects.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Materialanything — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Materialanything when you need a self-hosted research-code workflow for generating PBR materials on 3D meshes or rendering material maps with Blender.

Who it is for

  • 3D artists and technical creators preparing material maps for meshes.
  • Graphics and computer-vision researchers exploring diffusion-based material generation.
  • Developers incorporating self-hosted material generation or rendering into 3D workflows.

Documented uses

  • Generate albedo, roughness, metallic, and bump maps for a texture-less mesh.
  • Process objects with pre-existing albedo or textures.
  • Render material maps for an object in view or UV space.
  • Prepare large-scale material rendering using the supplied distributed rendering script.
Before you choose

Requirements & limits

The repository is public and lists an MIT license. The documented inference workflow is self-hosted.

  • Inference setup requires PyTorch and PyTorch3D, repository dependencies, and downloaded estimator and refiner models.
  • Blender 3.2.2 is recommended for the supplied scripts; newer versions may have node-name compatibility issues.
  • The repository includes a separate dataset and rendering scripts in addition to the material-generation model.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up Materialanything inference

Install the documented inference dependencies, prepare Blender, and download the estimator and refiner models.

Before you start

  • Linux environment
  • Python
  • PyTorch and PyTorch3D
  • Blender 3.2.2 recommended

Steps

  1. Install PyTorch and PyTorch3D.

  2. Install repository dependencies with the command below.

    Documented command
    pip install -r requirements.txt

  3. Use Blender 3.2.2 and ensure its binary path is on PATH.

  4. Download the models with the command below.

    Documented command
    bash ./download_models.sh

Expected outcome

The inference dependencies and the material estimator and refiner model files are prepared.

Guide limitations

  • The repository notes that newer Blender versions may have compatibility issues because some node names have changed.
  • This setup guide does not cover running inference.
Documentation 1 ↗
Source-based practical guide

Render material maps for a 3D object

Use the supplied Blender script to render an object's material maps in view or UV space.

Before you start

  • Blender with the repository scripts available
  • A 3D object file such as my_object.glb
  • Blender binary path added to PATH

Steps

  1. Choose VIEW or UV as the render space.

  2. Run rendering_scripts/blender_script_material.py with the object path, output directory, and selected render space.

Expected outcome

The script is configured to render material maps for the specified object in the chosen space.

Guide limitations

  • The documented commands illustrate rendering; they do not establish execution results.
  • Blender 3.2.2 is recommended, and newer versions may encounter node-name compatibility issues.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
3DTopia/MaterialAnything · Official homepage title · retained in its source language
How we review sourcesSuggest a correction