bugfixed: stuck when training with dist_train.sh (#728)

1f5b7872 · Shaoshuai Shi · GitHub · ec982888 · 1f5b7872
Unverified Commit 1f5b7872 authored Dec 27, 2021 by Shaoshuai Shi Committed by GitHub Dec 27, 2021
Show whitespace changes
Inline Side-by-side

Showing with 3 additions and 3 deletions

pcdet/utils/common_utils.py pcdet/utils/common_utils.py +3 -3

No files found.
--- a/pcdet/utils/common_utils.py
+++ b/pcdet/utils/common_utils.py
@@ -166,9 +166,9 @@ def init_dist_pytorch(tcp_port, local_rank, backend='nccl'):
    torch.cuda.set_device(local_rank % num_gpus)
    dist.init_process_group(
        backend=backend,
-        init_method='tcp://127.0.0.1:%d' % tcp_port,
-        rank=local_rank,
-        world_size=num_gpus
+        # init_method='tcp://127.0.0.1:%d' % tcp_port,
+        # rank=local_rank,
+        # world_size=num_gpus
    )
    rank = dist.get_rank()
    return num_gpus, rank