phenotypic.sdk_.slurm.with_default_gpu_request#
- phenotypic.sdk_.slurm.with_default_gpu_request(slurm_args: dict[str, Any]) dict[str, Any][source]#
slurm_args as a non-staged SLURM run submits them for a GPU pipeline.
AutonomousSLURMStrategyaddsslurm_gpus_per_node=1unless the profile already names it. The run preflight tests the same profile, so the two share this definition (Phase E review E2).