github
/
ColossalAI
mirror of https://github.com/hpcaitech/ColossalAI
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

[cli] hotfix launch command for multi-nodes (#4165)

pull/4167/head
Hongxin Liu 2023-07-04 17:54:40 +08:00 committed by GitHub
parent 2ac24040eb
commit 1908caad38
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
colossalai/cli/launcher/run.py
View File

@ -164,9 +164,7 @@ def get_launch_command(
]
else:
# extra launch args for torch distributed launcher with torch >= 1.9
default_torchrun_rdzv_args = dict(rdzv_backend="c10d",
rdzv_endpoint=f"{master_addr}:{master_port}",
rdzv_id="colossalai-default-job")
default_torchrun_rdzv_args = dict(master_addr=master_addr, master_port=master_port)
# update rdzv arguments
for key in default_torchrun_rdzv_args.keys():