Sleepychord/CogLTX

RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.

Opened this issue · 5 comments

when i run the run_20news.py,occur this mistake
图片

soultion: pytorch/pytorch#22538

does it mean we should add a dist.init_process_group('gloo', init_method='file:///tmp/somefile', rank=0, world_size=1) inside main_loop.py? But what should we pass in for init_method?
and the wierd thing is I am training on 1 gpu 1 machine, why a parallel or distributed problem occur?

soultion: pytorch/pytorch#22538

does it mean we should add a dist.init_process_group('gloo', init_method='file:///tmp/somefile', rank=0, world_size=1) inside main_loop.py? But what should we pass in for init_method? and the wierd thing is I am training on 1 gpu 1 machine, why a parallel or distributed problem occur?

Have you solved this problem?

when i run the run_20news.py,occur this mistake 图片

Have you solved this problem?

zhci commented

請問有人解決這個問題了嘛