torch.fb.rendezvous.zeus is imported in the util/misc.py, although there are no references to it; so I was able to run the code in non-distributed mode by simply commenting it out.
However, after trying to run the code in distributed mode for quite some time, I realized that zeus may be the culprit.
Since torch.fb.rendezvous.zeus doesn't seem to be open-source, could you please help clarify its functionality or just share a copy of the env variables, etc that it may be configuring?
|
import torch.fb.rendezvous.zeus |
ModuleNotFoundError: No module named 'torch.fb'
torch.fb.rendezvous.zeusis imported in the util/misc.py, although there are no references to it; so I was able to run the code in non-distributed mode by simply commenting it out.However, after trying to run the code in distributed mode for quite some time, I realized that
zeusmay be the culprit.Since
torch.fb.rendezvous.zeusdoesn't seem to be open-source, could you please help clarify its functionality or just share a copy of the env variables, etc that it may be configuring?mae_st/util/misc.py
Line 23 in dc072aa