Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

for windows user #9

Open
jtydhr88 opened this issue Nov 6, 2024 · 10 comments
Open

for windows user #9

jtydhr88 opened this issue Nov 6, 2024 · 10 comments

Comments

@jtydhr88
Copy link

jtydhr88 commented Nov 6, 2024

Hi,
I just run it on Windows 10, here is my step.
Env:
Windows 10
3090

Prerequisites:

  1. conda
  2. you also need to install c++ build tools from here
    img_1

Steps:

  1. conda create -n hunyuan3d-py39 python=3.9
  2. conda activate hunyuan3d-py39
  3. pip install torch==2.2.0 torchvision==0.17.0 --index-url https://download.pytorch.org/whl/cu118
  4. pip install diffusers transformers
  5. pip install rembg tqdm omegaconf matplotlib opencv-python imageio jaxtyping einops
  6. pip install SentencePiece accelerate trimesh PyMCubes xatlas libigl
  7. pip install git+https://github.com/facebookresearch/pytorch3d
  8. pip install git+https://github.com/NVlabs/nvdiffrast
  9. pip install open3d
  10. pip install Ninja
  11. pip install gradio
  12. git clone https://github.com/Tencent/Hunyuan3D-1
  13. cd Hunyuan3D-1
  14. pip install "huggingface_hub[cli]"
  15. create folder ./weights and ./weights/hunyuanDiT
  16. huggingface-cli download tencent/Hunyuan3D-1 --local-dir ./weights
  17. huggingface-cli download Tencent-Hunyuan/HunyuanDiT-v1.1-Diffusers-Distilled --local-dir ./weights/hunyuanDiT
  18. go to app.py and change 0.0.0.0 to 127.0.0.1 on line 52
  19. go to demos/example_list.txt and remove the first 3 lines of Chinese prompts.
  20. python app.py

quick test
image
image

good luck!

@h3clikejava
Copy link

3090多少显存也能跑?

@jtydhr88
Copy link
Author

jtydhr88 commented Nov 6, 2024

可以,就是慢点

@jtydhr88
Copy link
Author

jtydhr88 commented Nov 6, 2024

Update for python 3.11 and cuda 12.4, this repo also works well with them as tested.
Prerequisites:

  1. conda
  2. cuda 12.4
  3. you still need to install c++ build tools from here

Different Steps:

  1. conda create -n hunyuan3d-py311 python=3.11
  2. conda activate hunyuan3d-py311
  3. pip install torch torchvision xformers --index-url https://download.pytorch.org/whl/cu121
  4. (optional) you can download trition whl https://huggingface.co/madbuda/triton-windows-builds, as I am using python 3.11, so I will download 3.11 version
  5. (optional) pip install triton-2.1.0-cp311-cp311-win_amd64.whl

other steps are as same as above.

@wWzZb
Copy link

wWzZb commented Nov 7, 2024

image
安装SentencePiece accelerate trimesh PyMCubes xatlas libigl等库时报错,可以使用国内镜像,
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple SentencePiece accelerate trimesh PyMCubes xatlas libigl,
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple git+https://github.com/facebookresearch/pytorch3d

@jrc48
Copy link

jrc48 commented Nov 7, 2024

i'm stuck at installing pytorch3d, can anyone help?

@jtydhr88
Copy link
Author

jtydhr88 commented Nov 7, 2024

i'm stuck at installing pytorch3d, can anyone help?

please post your error

@jrc48
Copy link

jrc48 commented Nov 7, 2024

Collecting git+https://github.com/facebookresearch/pytorch3d
Cloning https://github.com/facebookresearch/pytorch3d to c:\users\jing\appdata\local\temp\pip-req-build-ezsxm3_q
Resolved https://github.com/facebookresearch/pytorch3d to commit 81d82980bc82fd605f27cca87f89ba08af94db3d
Preparing metadata (setup.py): started
Preparing metadata (setup.py): finished with status 'done'
Collecting iopath (from pytorch3d==0.7.8)
Using cached iopath-0.1.10-py3-none-any.whl
Requirement already satisfied: tqdm in c:\programdata\anaconda3\envs\hunyuan3d-py39\lib\site-packages (from iopath->pytorch3d==0.7.8) (4.67.0)
Requirement already satisfied: typing-extensions in c:\programdata\anaconda3\envs\hunyuan3d-py39\lib\site-packages (from iopath->pytorch3d==0.7.8) (4.9.0)
Collecting portalocker (from iopath->pytorch3d==0.7.8)
Using cached portalocker-2.10.1-py3-none-any.whl.metadata (8.5 kB)
Collecting pywin32>=226 (from portalocker->iopath->pytorch3d==0.7.8)
Using cached pywin32-308-cp39-cp39-win_amd64.whl.metadata (8.3 kB)
Requirement already satisfied: colorama in c:\programdata\anaconda3\envs\hunyuan3d-py39\lib\site-packages (from tqdm->iopath->pytorch3d==0.7.8) (0.4.6)
Using cached portalocker-2.10.1-py3-none-any.whl (18 kB)
Using cached pywin32-308-cp39-cp39-win_amd64.whl (6.6 MB)
Building wheels for collected packages: pytorch3d
Building wheel for pytorch3d (setup.py): started
Building wheel for pytorch3d (setup.py): finished with status 'error'
Running setup.py clean for pytorch3d
Failed to build pytorch3d

@HuiwenShi
Copy link
Collaborator

@SoftologyPro
Copy link

SoftologyPro commented Nov 7, 2024

I built a WHL for Pytorch3D to help people having build/install problems
pip install https://huggingface.co/datasets/Softology-Pro/VoC/resolve/main/pytorch3d-0.7.8-cp310-cp310-win_amd64.whl

@jrc48
Copy link

jrc48 commented Nov 8, 2024

still the same

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants