diff --git a/nemo_run/core/execution/templates/dgxc.sh.j2 b/nemo_run/core/execution/templates/dgxc.sh.j2 index 75bdede2..df7ab9f9 100644 --- a/nemo_run/core/execution/templates/dgxc.sh.j2 +++ b/nemo_run/core/execution/templates/dgxc.sh.j2 @@ -8,6 +8,7 @@ export TORCHX_MAX_RETRIES={{max_retries}} {%- for env_var in env_vars %} {{env_var}} {%- endfor %} +export GROUP_RANK=$(echo $HOSTNAME | grep -oE '[0-9]+$') {%- if ft_enabled %} {{ fault_tolerance.ft_launcher_setup(fault_tol_cfg_path, fault_tol_finished_flag_file, fault_tol_job_results_file) }}