Skip to content

Commit 253d77f

Browse files
committed
Adding an extra slice to the Pathways cluster to swap in when there is a slice failure
1 parent b52dfd8 commit 253d77f

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

axlearn/cloud/gcp/pathways_utils.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -593,7 +593,7 @@ def __call__(self) -> Sequence[Nested[Any]]:
593593
),
594594
dict(
595595
name=_PATHWAYS_WORKER_REPLICATED_JOB_NAME,
596-
replicas=cfg.accelerator.num_replicas,
596+
replicas=cfg.accelerator.num_replicas + 1,
597597
template=self._build_pathways_worker_job(),
598598
),
599599
]

0 commit comments

Comments
 (0)