search

IBM / Multi-GNN

Multi-GNN architectures for Anti-Money Laundering.
Apache License 2.0
36 stars 15 forks source link

Initial setup on Mac M1 #7

Open Stephen37927 opened 2 months ago

Stephen37927 commented 2 months ago

after running conda env create -f env.yml on my mac M1 in China it reports such an error:


Channels:
 - pytorch
 - nvidia
 - pyg
 - conda-forge
 - defaults
Platform: osx-arm64
Collecting package metadata (repodata.json): done
Solving environment: failed

PackagesNotFoundError: The following packages are not available from current channels:

  - pytorch::torchvision
  - pyg::pytorch-sparse
  - pyg::pytorch-scatter
  - pytorch::pytorch-cuda==11.8
  - pytorch::pytorch
  - pyg::pyg
  - cudatoolkit=11.8*
  - datatable

Current channels:

  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/osx-arm64
  - https://conda.anaconda.org/nvidia/osx-arm64
  - https://conda.anaconda.org/pyg/osx-arm64
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/osx-arm64
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/osx-arm64
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r/osx-arm64
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2/osx-arm64
  - https://conda.anaconda.org/pyg
  - https://conda.anaconda.org/pyg
  - https://conda.anaconda.org/pyg
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch

To search for alternate channels that may provide the conda package you're
looking for, navigate to

    https://anaconda.org

and use the search bar at the top of the page.

I 'm not familiar with torch but I know it may has something to do with cuda. I tried removing two packages regarding cuda but it still reports the same error. Anybody kindly offer guidance? I'm just trying to adopting it to my course project and the data prepared is small, unnecessary referring to GPU, I guess.

Linnore commented 1 week ago

@Stephen37927 There are some known issues of using pytorch-geometric on Apple M1/2/3. For me, installing the latest pytorch-geometric's cpu version suffices to run this repo.

pip install torch_geometric pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-2.3.0+cpu.html

Other libraries used by this repo can be pip installed. E.g., datatable, wandb. You may need to make wandb offline to run the codes due to connection stability in CN, btw.