ComfyUI-IF_Trellis

ComfyUI-IF_Trellis
★ 449

3D网格单视图重建多视角重建纹理导出
从单视图或多视角图片生成带纹理的3D网格，快速产出可导出的高质量三维素材并集成ComfyUI流程。
💡 从一张或多张图片快速生成可导出的3D模型用于游戏和可视化
🍴 40 Forks💻 Python🔄 2025-03-09
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/9671236b7e59
📦 requirements.txt
--extra-index-url
https://download.pytorch.org/whl/cu121
--find-links
https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.4.0_cu121.html
torch==2.4.0
torchvision==0.19.0
pillow==10.4.0
imageio==2.36.1
imageio-ffmpeg==0.5.1
tqdm==4.67.1
easydict==1.13
opencv-python-headless==4.10.0.84
scipy==1.14.1
rembg==2.0.60
onnxruntime==1.20.1
trimesh==4.5.3
xatlas==0.0.9
pyvista==0.44.2
pymeshfix==0.17.0
igraph==0.11.8
git+https://github.com/EasternJournalist/utils3d.git@9a4eb15e4021b67b12c460c7057d642626897ec8
xformers==0.0.27.post2
kaolin==0.17.0
spconv-cu120==2.3.6
transformers==4.46.3
gradio_litmodel3d==0.0.1
https://github.com/Dao-AILab/flash-attention/releases/download/v2.7.0.post2/flash_attn-2.7.0.post2+cu12torch2.4cxx11abiFALSE-cp310-cp310-linux_x86_64.whl
https://huggingface.co/spaces/JeffreyXiang/TRELLIS/resolve/main/wheels/diff_gaussian_rasterization-0.0.0-cp310-cp310-linux_x86_64.whl?download=true
https://huggingface.co/spaces/JeffreyXiang/TRELLIS/resolve/main/wheels/nvdiffrast-0.3.3-cp310-cp310-linux_x86_64.whl?download=true
SageAttention
teaser (1)
Watch this guide if you need extra help
blender_oWFuSjA6yq
thorium_jtbGswydSr
📄 README

ComfyUI-IF_Trellis

arXiv

Project Page

ComfyUI TRELLIS is a large 3D asset generation in various formats, such as Radiance Fields, 3D Gaussians, and meshes. The cornerstone of TRELLIS is a unified Structured LATent (SLAT) representation that allows decoding to different output formats and Rectified Flow Transformers tailored for SLAT as the powerful backbones.

Prerequisites

  • System: The original code is currently tested only on Linux. For windows setup, you may refer to #3
  • (This comfyui node is following this suggeted installation steps It works but you need to follow the steps as described in part one and two of the guide).
  • Windows users need to use the Win_requirements.txt.
  • Linux Users (Not tested Yet) use the linux_requirements.txt but I am still testing if it works with comfy in Linux I just added the original repo requirements.

  • Hardware: This repo improve memmory mangement with the PR made by Amorano you will need NVIDIA GPU with at least 8GB.
  • Software:
  • The CUDA Toolkit is needed to compile certain submodules. The code has been tested with CUDA versions 11.8 and 12.2. This repo use CUDA 12.4.
  • Conda is recommended for managing dependencies.
  • Python version 3.8 or higher is required.

    • Give unrestricted script access to powershell so venv can work:

  • Open an administrator powershell window
  • Type Set-ExecutionPolicy Unrestricted and answer A
  • Close admin powershell window

    • Installation Steps

  • Clone the repo:

    • “`

    cd ComfyUI/Custom_nodes

    git clone –recurse-submodules https://github.com/if-ai/ComfyUI-IF_Trellis.git

    “`

    MUST HAVE --recurse-submodules

    ONLY tested on windows but it should work easier in Linux without any issues or needing such specific stuffs.

    *NOT tested or compatible with PORTABLE comfy embeded python env*

    watch quick overview of setting the env if needed

    [](https://www.youtube.com/watch?v=-vEpuYL9I3g)

    You need to set up the environment first

    follow this guide for the first part

    Set the VSCode Cpp Envirronment as in the guide

    Installing Triton and Sage Attention Flash Attention

    Setting up ComfyUI with the Xformers, flash attention, Sage-attention(Optional Recommended for Hunyuan and other Video models)

    * You can Also Try This Other Guide *

    [](https://www.youtube.com/watch?v=FjNfDsX-jR0)

    How to Run Micromamba in ComfyUI with Triton, SAGE Attention, Flash Attention, and xFormers for ComfyUI 3D

    📦 Second part of the installation

    Activate youur comfy environment

    (gen) PS D:\ComfyUI\custom_nodes\ComfyUI-IF_Trellis> micromamba activate gen

    If you haven’t set your vars or for some reason it can’t compile some of this specially nvdiffrast

    it doesn’t hurt if you do it again now.

    cmd.exe /c "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvarsall.bat" x64 "&&" powershell

    You will see some message like this:

    ** Visual Studio 2019 Developer Command Prompt v16.11.41

    ** Copyright (c) 2021 Microsoft Corporation

    [vcvarsall.bat] Environment initialized for: ‘x64’

    Windows PowerShell

    Copyright (C) Microsoft Corporation. All rights reserved.

    pip install -r win_requirements.txt


    pip install git+https://github.com/EasternJournalist/utils3d.git@9a4eb15e4021b67b12c460c7057d642626897ec8

New-Item -ItemType Directory -Force -Path C:\tmp\extensions

git clone --recurse-submodules https://github.com/JeffreyXiang/diffoctreerast.git C:\tmp\extensions\diffoctreerast

pip install C:\tmp\extensions\diffoctreerast

git clone https://github.com/autonomousvision/mip-splatting.git C:\tmp\extensions\mip-splatting

pip install C:\tmp\extensions\mip-splatting\submodules\diff-gaussian-rasterization\

pip install kaolin -f https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.4.0_cu121.html

git clone https://github.com/NVlabs/nvdiffrast.git C:\tmp\extensions\nvdiffrast

pip install C:\tmp\extensions\nvdiffrast

    🌟 Features

  • High Quality: It produces diverse 3D assets at high quality with intricate shape and texture details.
  • Versatility: It takes text or image prompts and can generate various final 3D representations including but not limited to *Radiance Fields*, *3D Gaussians*, and *meshes*, accommodating diverse downstream requirements.
  • Flexible Editing: It allows for easy editings of generated 3D assets, such as generating variants of the same object or local editing of the 3D asset.

    • 🚧 TODO List

  • [x] Release comfyUI-IF_Trellis
  • [x] 3D Viewport
  • [x] OPT mode
  • [x] Multiviews
  • [ ] Sage attn
  • [ ] Installation
  • [ ] colab notebook
  • [ ] MarchingCubes mode
  • [ ] Exposing Texture paranmeters