Open Chengyuelan opened 3 years ago
I got the same problem and no solution:(
Inspired by issue #352 , I solved my problem.
In AdelaiDet/adet/config/defaults.py
,
change
_C.MODEL.BASIS_MODULE.NORM = "SyncBN"
to
_C.MODEL.BASIS_MODULE.NORM = "BN"
I hope it can help you.
我也是遇到了这个问题,方法很有效,谢谢!
I encountered this problem when I used my own data set to train BlendMask. I would like to ask, how can I solve it?
File "/home/excellence/anaconda3/envs/BlendMask/lib/python3.8/site-packages/torch/nn/modules/batchnorm.py", line 532, in forward world_size = torch.distributed.get_world_size(process_group) File "/home/excellence/anaconda3/envs/BlendMask/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 711, in getworld size return _get_group_size(group) File "/home/excellence/anaconda3/envs/BlendMask/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 263, in _get_group _size default_pg = _get_default_group() File "/home/excellence/anaconda3/envs/BlendMask/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 347, in _get_defau lt_group raise RuntimeError("Default process group has not been initialized, " RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.