diff --git a/nemo_run/run/torchx_backend/schedulers/slurm.py b/nemo_run/run/torchx_backend/schedulers/slurm.py index f0d4f6f0..b65e0504 100644 --- a/nemo_run/run/torchx_backend/schedulers/slurm.py +++ b/nemo_run/run/torchx_backend/schedulers/slurm.py @@ -360,6 +360,7 @@ def _check_finished(self) -> None: if self._ls_term else os.path.join(self._remote_dir, f"log*{extension}") ) + ls_term = ls_term.replace("%j", self._app_id) ls_output = self._scheduler.tunnel.run( f"ls -1 {ls_term} 2> /dev/null", warn=True,