1. RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 37434) of binary: /home/lvying/anaconda3/envs/LVravs_py310_tr113/bin/python
2. RuntimeError: context has already been set