Thank you for your code! I can run it successfully with one GPU. But when I ran the code with multiple GPUs, I got the following error:
Traceback (most recent call last):
File "/opt/conda/lib/python3.8/site-packages/torch/multiprocessing/spawn.py", line 59, in _wrap
fn(i, *args)
File "/3dod2/Workspace/irem_code_release/train.py", line 690, in main_single
sync_model(model)
File "/3dod2/Workspace/irem_code_release/train.py", line 305, in sync_model
for model in model:
TypeError: 'EBM' object is not iterable
I am fairly new to the python world. I got an error that the torch-sparse module was not found.
Diving into torch (PyTorch) and the likes, well ..... you need to do a lot of versioning admin (hell) yourself. May be I got it running now . Using WSL2 Ubuntu 22 after three days of trying and learning.
tnx