mirror of
https://github.com/wassname/pytorch-lightning.git
synced 2026-09-11 12:31:23 +08:00
fixed nccl init
This commit is contained in:
@@ -386,7 +386,8 @@ class Trainer(TrainerIO):
|
||||
root_node = os.environ['SLURM_NODELIST'].split(' ')[0]
|
||||
os.environ['MASTER_ADDR'] = root_node
|
||||
os.environ['MASTER_PORT'] = f'{port}'
|
||||
dist.init_process_group("nccl", rank=self.proc_rank, world_size=self.world_size)
|
||||
# dist.init_process_group("nccl", rank=self.proc_rank, world_size=self.world_size)
|
||||
dist.init_process_group("nccl")
|
||||
|
||||
def __run_pretrain_routine(self, model):
|
||||
"""
|
||||
|
||||
Reference in New Issue
Block a user