CUDA error

[mrinal-shekhar]

Hi I have tried to install equibind and create an environment for both CUDA enabled and CPU machines using environment.yml and environment_cpu.yml. I am trying to run a test example using the following command after activating the equibind environment python multiligand_inference.py -o ./test_output/ -r ./test_input/protein.pdb -l ./test_input/ligand.sdf However, i am getting the same error for both CUDA and CPU-enabled installations:

device = cpu Entering batch ending in index 8/18 Traceback (most recent call last): File "multiligand_inference.py", line 275, in main() File "multiligand_inference.py", line 272, in main write_while_inferring(lig_loader, model, args) File "multiligand_inference.py", line 216, in write_while_inferring lig_graphs = lig_graphs.to(args.device) File "/Users/mshekhar/miniconda3/envs/equibind/lib/python3.7/site-packages/dgl/heterograph.py", line 5448, in to ret._graph = self._graph.copy_to(utils.to_dgl_context(device)) File "/Users/mshekhar/miniconda3/envs/equibind/lib/python3.7/site-packages/dgl/utils/internal.py", line 533, in to_dgl_context device_id = F.device_id(ctx) File "/Users/mshekhar/miniconda3/envs/equibind/lib/python3.7/site-packages/dgl/backend/pytorch/tensor.py", line 90, in device_id return 0 if ctx.type == 'cpu' else th.cuda.current_device() File "/Users/mshekhar/miniconda3/envs/equibind/lib/python3.7/site-packages/torch/cuda/init.py", line 479, in current_device _lazy_init() File "/Users/mshekhar/miniconda3/envs/equibind/lib/python3.7/site-packages/torch/cuda/init.py", line 208, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabled

Please help Regards Mrinal

[pykao]

Hi @mrinal-shekhar,

I got the same error when I was trying to call python multiligand_inference.py

DGL backend not selected or invalid.  Assuming PyTorch for now.
Setting the default backend to "pytorch". You can change it in the ~/.dgl/config.json file or export the DGLBACKEND environment variable.  Valid options are: pytorch, mxnet, tensorflow (all lowercase)
[2022-07-20 11:40:09.969125] [ Using Seed :  1  ]
device = cpu
Entering batch ending in index 8/667
Traceback (most recent call last):
  File "multiligand_inference.py", line 275, in <module>
    main()
  File "multiligand_inference.py", line 272, in main
    write_while_inferring(lig_loader, model, args)
  File "multiligand_inference.py", line 216, in write_while_inferring
    lig_graphs = lig_graphs.to(args.device)
  File "/home/ken/anaconda3/envs/equibind/lib/python3.7/site-packages/dgl/heterograph.py", line 5448, in to
    ret._graph = self._graph.copy_to(utils.to_dgl_context(device))
  File "/home/ken/anaconda3/envs/equibind/lib/python3.7/site-packages/dgl/utils/internal.py", line 533, in to_dgl_context
    device_id = F.device_id(ctx)
  File "/home/ken/anaconda3/envs/equibind/lib/python3.7/site-packages/dgl/backend/pytorch/tensor.py", line 90, in device_id
    return 0 if ctx.type == 'cpu' else th.cuda.current_device()
  File "/home/ken/anaconda3/envs/equibind/lib/python3.7/site-packages/torch/cuda/__init__.py", line 479, in current_device
    _lazy_init()
  File "/home/ken/anaconda3/envs/equibind/lib/python3.7/site-packages/torch/cuda/__init__.py", line 208, in _lazy_init
    raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled

[niksubramanian]

Could it be that you're using an incompatible CUDA version?

If you've used the environment.yml file that was provided, you'll need to have CUDA10.2 installed. I haven't tested with different versions of CUDA (e.g. 11.6), but I assume you'll need to install different versions of cudatoolkit and torch to get things to work.

[pykao]

@niksubramanian

I am using CUDA11.7. I will try to modify the environment.yml file to make it work with CUDA11.7.

Thanks, Ken