search

txie-93 / cdvae

An SE(3)-invariant autoencoder for generating the periodic structure of materials [ICLR 2022]
MIT License
230 stars 91 forks source link

Running `mp-20` is making my Dell-G7 on Windows non-responsive, can't run basic usage #17

Open sgbaird opened 2 years ago

sgbaird commented 2 years ago

I end up getting some combination of my computer becoming non-responsive/needing to force quit VS Code and the following error message:

[WinError 1455] The paging file is too small for this operation to complete. Error loading "C:\Users\sterg\Miniconda3\envs\cdvae\lib\site-packages\torch\lib\caffe2_detectron_ops_gpu.dll" or one of its dependencies.
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\site-packages\torch\__init__.py", line 123, in <module>
    raise err
  File "C:\Users\sterg\Documents\GitHub\sparks-baird\cdvae\cdvae\run.py", line 6, in <module>
    import torch
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\runpy.py", line 97, in _run_module_code
    _run_code(code, mod_globals, init_globals,
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\runpy.py", line 265, in run_path
    return _run_module_code(code, init_globals, run_name,
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\site-packages\multiprocess\spawn.py", line 287, in _fixup_main_from_path
    main_content = runpy.run_path(main_path,
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\site-packages\multiprocess\spawn.py", line 236, in prepare
    _fixup_main_from_path(data['init_main_from_path'])
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\site-packages\multiprocess\spawn.py", line 125, in _main
    prepare(preparation_data)
  File "C:\Users\sterg\miniconda3\envs\cdvae\Lib\site-packages\multiprocess\spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "<string>", line 1, in <module> (Current frame)

Related: https://discuss.pytorch.org/t/winerror-1455-the-paging-file-is-too-small-for-this-operation-to-complete/131233

Which discusses a somewhat involved / hacky workaround described in https://stackoverflow.com/a/69489193/13697228 and that this will probably no longer be an issue with Cuda 11.7.

Incidentally, despite running PyTorch for a bunch of different projects, this is the first time I'm seeing this one.

sgbaird commented 2 years ago

Spent a while trying to get this up and running. Will probably come back to it later, including trying to run on WSL.

sgbaird commented 2 years ago

It seems to be running OK got further on Windows Subsystem for Linux 2 (WSL2) https://github.com/txie-93/cdvae/issues/2#issuecomment-1153050037, but not on Windows. Maybe the workaround above would help for Windows (no WSL2).