Troubleshooting OGBN-Arxiv Model: Error with Marius Preprocessing Command in Docker

[JhengLu]

Describe the bug Hello,

I encountered an issue after building the image and attempting to use the OGBN-Arxiv model. When running the following command within the Marius Docker container from the root directory ("/working_dir/marius"):

marius_preprocess --dataset ogbn_arxiv --output_directory datasets/ogbn_arxiv_example/

It shows:

Using existing dataset of OGBN_ARXIV
Traceback (most recent call last):
  File "/usr/local/bin/marius_preprocess", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.10/dist-packages/marius/tools/marius_preprocess.py", line 167, in main
    dataset.preprocess(
  File "/usr/local/lib/python3.10/dist-packages/marius/tools/preprocess/datasets/ogbn_arxiv.py", line 80, in preprocess
    converter = TorchEdgeListConverter(
TypeError: TorchEdgeListConverter.__init__() got an unexpected keyword argument 'columns'

Could you please assist me in resolving this issue?

Thank you!

Environment cmake: version: 3.27.7 cpu_info: num_cpus: 112 total_memory: 110GB cuda: version: '11.8' gpu_info:

• memory: 44GB name: NVIDIA A40 marius: bindings_installed: true install_path: /usr/local/lib/python3.10/dist-packages/marius version: 0.0.2 openmp: version: '201511' operating_system: platform: Linux-4.18.0-513.5.1.el8_9.x86_64-x86_64-with-glibc2.35 pybind: PYBIND11_BUILD_ABI: _cxxabi1011 PYBIND11_COMPILER_TYPE: _gcc PYBIND11_STDLIB: _libstdcpp python: deps: numpy_version: 1.26.2 omegaconf_version: 2.3.0 pandas_version: 2.1.3 pip_version: 22.0.2 pyspark_version: N/A pytest_version: N/A torch_version: !!python/object/new:torch.torch_version.TorchVersion
  • 2.0.1+cu118 tox_version: N/A version: 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] pytorch: install_path: /usr/local/lib/python3.10/dist-packages/torch version: !!python/object/new:torch.torch_version.TorchVersion
    • 2.0.1+cu118