Error while training and use of pre-trained weights

[Shuhul24]

While training occworld, using the command for training occworld model, after training successfully for 1 epoch, I end up with the following error:

Traceback (most recent call last):
  File "/scratch/p24cs0005/OccWorld/train.py", line 362, in <module>
    main(0, args)
  File "/scratch/p24cs0005/OccWorld/train.py", line 327, in main
    val_miou, _ = CalMeanIou_sem._after_epoch()
  File "/scratch/p24cs0005/OccWorld/utils/metric_util.py", line 96, in _after_epoch
    dist.all_reduce(self.total_seen)
  File "/csehome/p24cs0005/miniconda3/envs/occworld/lib/python3.9/site-packages/torch/distributed/distributed_c10d.py", line 1451, in wrapper
    return func(*args, **kwargs)
  File "/csehome/p24cs0005/miniconda3/envs/occworld/lib/python3.9/site-packages/torch/distributed/distributed_c10d.py", line 1699, in all_reduce
    default_pg = _get_default_group()
  File "/csehome/p24cs0005/miniconda3/envs/occworld/lib/python3.9/site-packages/torch/distributed/distributed_c10d.py", line 707, in _get_default_group
    raise RuntimeError(
RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.

Why does this error occur as it happens after being trained for 1 epoch and evaluation?

Also, in the github repo, there is a pretrained weight link which leads to a file latest.pth. Is it for occworld model or the VQVAE model?

I have this doubt because I tried training the VQVAE, and the weights in out/vqvae also consists of a latest.pth file. Also, if on putting the latest.pth in out/occworld, can I evaluate my results by running the command for evaluation?