wassname/pytorch-lightning
1
0
Fork 0
mirror of https://github.com/wassname/pytorch-lightning.git synced 2026-09-09 11:32:07 +08:00
Code Issues Packages Projects Releases Wiki Activity

using slurm flag to fine node nb

Browse Source
This commit is contained in:
William Falcon
2019-07-08 14:22:09 -04:00
parent d2a717d31e
commit 7c0e3715dd
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pytorch_lightning/models/trainer.py
+2 -1
View File
@@ -358,7 +358,8 @@ class Trainer(TrainerIO):
# the first gpu in the world becomes the host
# this is based on its global rank
# saves the ip to disk
# it communicates its ip by saving an ip_table to the slurm cluster logging dir
# every other process waits for this ip to appear before continuing
ip_table_name = f'.ip_meta_' + os.environ['SLURM_JOB_ID']
ip_file = os.path.join(ip_file_dir, ip_table_name)
os.makedirs(ip_file_dir, exist_ok=True)