For Windows users

[jtydhr88]

I got it working on my Windows.
It might require Visual Studio Build Tools. However, I am not sure because my local already installed previously. If it needs, you can find from Visual Studio Build Tools.

Env

1. WIndows 10
2. Conda - Python 3.11
3. 3090
4. Cuda 12.1

Setup Steps

1. conda create -n unique3d-py311 python=3.11
2. conda activate unique3d-py311
3. pip install torch torchvision torchaudio xformers --index-url https://download.pytorch.org/whl/cu121
4. (Download the trition whl from https://huggingface.co/madbuda/triton-windows-builds, because I use python 311, so download whl build for py311 and install manually) pip install triton-2.1.0-cp311-cp311-win_amd64.whl
5. pip install Ninja
6. (Due the latest diffusers is not compatible to the code of repo, we MUST install this for diffusers manually) pip install diffusers==0.27.2
7. pip install grpcio werkzeug tensorboard-data-server
8. remove torch>=2.0.1, diffusers>=0.26.3, xformers, onnxruntime_gpu from requirement.txt
   
9. pip install -r requirements.txt
10. pip uninstall onnxruntime
11. pip uninstall onnxruntime-gpu
12. (We MUST uninstall onnxruntime and onnxruntime-gpu first, then install this given version manually) pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

Runtime

1. you need to create the output folder tmp\gradio under the running driver by yourself, for me, it is F:\tmp\gradio

2. You will see the error of onnxruntime
   
   but it would not effect the generation of the mesh (maybe?)

(Update) In order to fix the TensorRT error, you need to download TensorRT bundle for windows from https://github.com/NVIDIA/TensorRT, and configure TensorRT-10.0.1.6\lib in your PATH env variable. meanwhile also need to configure CUDA and cuDNN, please refer to github page for more details.

Quick test

Good luck~

[SoftologyPro]

I had a go getting this working under Windows too. For my setup I am supporting CUDA 11.8 and Python 3.10.x because that is what I use for all the AI/ML scripts in Visions of Chaos.

This is my install.bat file. Save to an empty directory (in my case D:\Tests\Unique3D) and run it command line to install Unique3D automatically.

@echo off

echo *** Deleting Unique3D directory if it exists
if exist Unique3D\. rd /S /Q Unique3D

echo *** Cloning Unique3D repository
git clone https://huggingface.co/spaces/Wuvin/Unique3D
cd Unique3D

echo *** Creating venv
python -m venv venv
call venv\scripts\activate.bat

python -m pip install --upgrade pip
pip install torch torchvision torchaudio xformers --index-url https://download.pytorch.org/whl/cu118
pip install https://softology.pro/wheels/triton-2.1.0-cp310-cp310-win_amd64.whl
pip install ninja
pip install diffusers==0.27.2
pip install grpcio werkzeug tensorboard-data-server

echo *** Removing linux specific packages from requirements.txt
type requirements.txt | findstr /v pytorch3d > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v ort-nightly > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v onnxruntime > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v torch-scatter > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v diffusers > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v xformers > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v torch > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt

echo *** Installing requirments.txt
pip install -r requirements.txt

echo *** Installing other required packages
pip uninstall -y xformers
pip install --no-cache-dir --ignore-installed --force-reinstall --no-warn-conflicts xformers==0.0.25 --index-url https://download.pytorch.org/whl/cu118
pip uninstall -y torch
pip uninstall -y torch
pip install --no-cache-dir --ignore-installed --force-reinstall --no-warn-conflicts torch==2.2.1+cu118 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

pip uninstall onnxruntime
pip uninstall onnxruntime-gpu
pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/
pip install ort-nightly-gpu

pip install https://files.pythonhosted.org/packages/b2/3c/6205090507ea96e6e56d0deda8d0fc4c507026ef3772e55b637a5d0b7c61/spaces-0.28.3-py3-none-any.whl

call venv\scripts\deactivate.bat
cd ..

echo *** Finished Unique3D install
echo.

Once the install finishes I try running
python gradio_app\gradio_local.py --port 7860
which gives this error

Traceback (most recent call last):
  File "D:\Tests\Unique3D\Unique3D\gradio_app\gradio_local.py", line 20, in <module>
    from gradio_app.gradio_3dgen import create_ui as create_3d_ui
  File "D:\Tests\Unique3D\Unique3D\.\gradio_app\gradio_3dgen.py", line 1, in <module>
    import spaces
ModuleNotFoundError: No module named 'spaces'

After pip install spaces and rerunning it I get

python gradio_app\gradio_local.py --port 7860
Traceback (most recent call last):
  File "D:\Tests\Unique3D\Unique3D\gradio_app\gradio_local.py", line 20, in <module>
    from gradio_app.gradio_3dgen import create_ui as create_3d_ui
  File "D:\Tests\Unique3D\Unique3D\.\gradio_app\gradio_3dgen.py", line 1, in <module>
    import spaces
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\__init__.py", line 10, in <module>
    from .zero.decorator import GPU
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\zero\decorator.py", line 21, in <module>
    from .wrappers import regular_function_wrapper
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\zero\wrappers.py", line 15, in <module>
    from multiprocessing.context import ForkProcess
ImportError: cannot import name 'ForkProcess' from 'multiprocessing.context' (D:\Python\lib\multiprocessing\context.py)

ForkProcess seems to be Linux only so no go for me.
If we can get a single click install.bat working that would really help Windows users (and me so I can add support for Unique3D into Visions of Chaos).

[IIIIIIIllllllllIIIII]

bump

[Aiunii]

Hey!  Welcome to 4D digital art creation universe—— Aiuni. 🤖First, join our community by:  https://discord.gg/aiuni 🦄️Then you will get the invitation code at “getting-started” channel. Start your creative journey and have fun!🌟

…

From: ***@***.***> Date: Sat, Jun 15, 2024, 4:23 AM Subject: Re: [AiuniAI/Unique3D] For Windows users (Issue #15) To: ***@***.***> Cc: ***@***.***> bump — Reply to this email directly, view it on GitHub<#15 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BI3NO5RD3VCI4QZO2WZX2ODZHNGKXAVCNFSM6AAAAABJAOPC56VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRYG4YTEMJZGI>. You are receiving this because you are subscribed to this thread.[image: https://github.com/notifications/beacon/BI3NO5QIEWXNSVVCHOWHMPLZHNGKXA5CNFSM6AAAAABJAOPC56WGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTUBIPWAA.gif]Message ID: ***@***.***>

[DamienCz]

for autodownload
(unique3d-py311) E:\Unique3D>pip show torch
Name: torch
Version: 2.3.0+cu121
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Home-page: https://pytorch.org/
Author: PyTorch Team
Author-email: packages@pytorch.org
License: BSD-3
Location: C:\Users\WS01\anaconda3\envs\unique3d-py311\Lib\site-packages
Requires: filelock, fsspec, jinja2, mkl, networkx, sympy, typing-extensions
Required-by: torchaudio, torchvision, xformers

(unique3d-py311) E:\Unique3D>
then ....
Failed to build pytorch3d
ERROR: Could not build wheels for pytorch3d, which is required to install pyproject.toml-based projects

how to deal with this problem...

[jtydhr88]

See #29

[jayc0207]

python app\gradio_local.py --port 7860

Warning! extra parameter in cli is not verified, may cause erros.
Loading pipeline components...: 100%|███████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:00<00:00, 416.27it/s]
You have disabled the safety checker for <class 'custum_3d_diffusion.custum_pipeline.unifield_pipeline_img2mvimg.StableDiffusionImage2MVCustomPipeline'> by passing safety_checker=None. Ensure that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered results in services or applications open to the public. Both the diffusers team and Hugging Face strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling it only for use-cases that involve analyzing network behavior or auditing its results. For more information, please have a look at huggingface/diffusers#254 .
Warning! extra parameter in cli is not verified, may cause erros.
C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\huggingface_hub\file_download.py:1150: FutureWarning: resume_download is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use force_download=True.
warnings.warn(
Traceback (most recent call last):
File "D:\Unique3D\Unique3D\app\gradio_local.py", line 20, in
from app.gradio_3dgen import create_ui as create_3d_ui
File "D:\Unique3D\Unique3D\app\gradio_3dgen.py", line 7, in
from app.custom_models.normal_prediction import predict_normals
File "D:\Unique3D\Unique3D\app\custom_models\normal_prediction.py", line 10, in
trainer, pipeline = load_pipeline(training_config, checkpoint_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Unique3D\Unique3D\app\custom_models\utils.py", line 65, in load_pipeline
configurable_unet.unet.to(device, dtype=weight_dtype)
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 1173, in to
return self._apply(convert)
^^^^^^^^^^^^^^^^^^^^
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
[Previous line repeated 7 more times]
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 804, in _apply
param_applied = fn(param)
^^^^^^^^^
File "C:\Users\jayis\anaconda3\envs\unique3d-py311\Lib\site-packages\torch\nn\modules\module.py", line 1159, in convert
return t.to(
^^^^^
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 26.00 MiB. GPU

How to solve this problem ?
windows 11
nvidia rt3070 Ti 8G
32 G RAM
i9 CPU

[turugi-ni]

I had a go getting this working under Windows too. For my setup I am supporting CUDA 11.8 and Python 3.10.x because that is what I use for all the AI/ML scripts in Visions of Chaos.

This is my install.bat file. Save to an empty directory (in my case D:\Tests\Unique3D) and run it command line to install Unique3D automatically.

@echo off

echo *** Deleting Unique3D directory if it exists
if exist Unique3D\. rd /S /Q Unique3D

echo *** Cloning Unique3D repository
git clone https://huggingface.co/spaces/Wuvin/Unique3D
cd Unique3D

echo *** Creating venv
python -m venv venv
call venv\scripts\activate.bat

python -m pip install --upgrade pip
pip install torch torchvision torchaudio xformers --index-url https://download.pytorch.org/whl/cu118
pip install https://softology.pro/wheels/triton-2.1.0-cp310-cp310-win_amd64.whl
pip install ninja
pip install diffusers==0.27.2
pip install grpcio werkzeug tensorboard-data-server

echo *** Removing linux specific packages from requirements.txt
type requirements.txt | findstr /v pytorch3d > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v ort-nightly > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v onnxruntime > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v torch-scatter > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v diffusers > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v xformers > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt
type requirements.txt | findstr /v torch > stripped.txt
del requirements.txt
ren stripped.txt requirements.txt

echo *** Installing requirments.txt
pip install -r requirements.txt

echo *** Installing other required packages
pip uninstall -y xformers
pip install --no-cache-dir --ignore-installed --force-reinstall --no-warn-conflicts xformers==0.0.25 --index-url https://download.pytorch.org/whl/cu118
pip uninstall -y torch
pip uninstall -y torch
pip install --no-cache-dir --ignore-installed --force-reinstall --no-warn-conflicts torch==2.2.1+cu118 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

pip uninstall onnxruntime
pip uninstall onnxruntime-gpu
pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/
pip install ort-nightly-gpu

pip install https://files.pythonhosted.org/packages/b2/3c/6205090507ea96e6e56d0deda8d0fc4c507026ef3772e55b637a5d0b7c61/spaces-0.28.3-py3-none-any.whl

call venv\scripts\deactivate.bat
cd ..

echo *** Finished Unique3D install
echo.

Once the install finishes I try running python gradio_app\gradio_local.py --port 7860 which gives this error

Traceback (most recent call last):
  File "D:\Tests\Unique3D\Unique3D\gradio_app\gradio_local.py", line 20, in <module>
    from gradio_app.gradio_3dgen import create_ui as create_3d_ui
  File "D:\Tests\Unique3D\Unique3D\.\gradio_app\gradio_3dgen.py", line 1, in <module>
    import spaces
ModuleNotFoundError: No module named 'spaces'

After pip install spaces and rerunning it I get

python gradio_app\gradio_local.py --port 7860
Traceback (most recent call last):
  File "D:\Tests\Unique3D\Unique3D\gradio_app\gradio_local.py", line 20, in <module>
    from gradio_app.gradio_3dgen import create_ui as create_3d_ui
  File "D:\Tests\Unique3D\Unique3D\.\gradio_app\gradio_3dgen.py", line 1, in <module>
    import spaces
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\__init__.py", line 10, in <module>
    from .zero.decorator import GPU
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\zero\decorator.py", line 21, in <module>
    from .wrappers import regular_function_wrapper
  File "D:\Tests\Unique3D\Unique3D\venv\lib\site-packages\spaces\zero\wrappers.py", line 15, in <module>
    from multiprocessing.context import ForkProcess
ImportError: cannot import name 'ForkProcess' from 'multiprocessing.context' (D:\Python\lib\multiprocessing\context.py)

ForkProcess seems to be Linux only so no go for me. If we can get a single click install.bat working that would really help Windows users (and me so I can add support for Unique3D into Visions of Chaos).

Windows does not support ForkProcess, open venv\lib\site-packages\spaces\zero\wrappers.py in a text editor tool, Change from multiprocessing.context import ForkProcess to from multiprocessing.context import SpawnProcess in line 15 and Try changing line 50 Process = from multiprocessing.get_context('fork').Process to Process = from multiprocessing.get_context('spawn').Process, Still, we encounter new errors. Unique3D has too many dependencies of dependencies and more dependencies of dependencies I gave up installing on Windows.