when i try to train gpt on HPC with slurm, I encounter this error. pytorch 11.3, cuda 11.4, nccl 2.8(I'm not sure whether nccl has been used) File "/gs/home/momo ...
2023-03-23 10:33:03,882 launcher.py[line:45] INFO ProxyPool Version: 2.4.0 2023-03-23 10:33:03,883 launcher.py[line:50] INFO ProxyPool configure HOST: 0.0.0.0 2023-03-23 10:33:03,884 ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results