mirrors/mmengine
1
0
Fork 0
mirror of https://github.com/open-mmlab/mmengine.git synced 2025-06-03 21:54:44 +08:00
Code Issues Projects Releases Wiki Activity

[Enhancement] Handle the case for Multi-Instance GPUs when using cuda_visible_devices (#1164)

Browse Source
This commit is contained in:
Adrian Joshua Strutt 2023-05-28 15:30:36 +02:00 committed by GitHub
parent 691500dce6
commit 0ff89f7c36
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
mmengine/logging/logger.py
View File

@ -376,7 +376,12 @@ def _get_device_id():
cuda_visible_devices = list(range(num_device)) cuda_visible_devices = list(range(num_device))
else: else:
cuda_visible_devices = cuda_visible_devices.split(',') cuda_visible_devices = cuda_visible_devices.split(',')
if cuda_visible_devices[local_rank].isdigit():
return int(cuda_visible_devices[local_rank]) return int(cuda_visible_devices[local_rank])
else:
# handle case for Multi-Instance GPUs
# see #1148 for details
return cuda_visible_devices[local_rank]
def _get_host_info() -> str: def _get_host_info() -> str: