Skip to content
Snippets Groups Projects
Commit 2eb76659 authored by Jan Ebert's avatar Jan Ebert
Browse files

Integrate script updates into README

parent 78951613
No related branches found
No related tags found
No related merge requests found
...@@ -262,9 +262,10 @@ use ...@@ -262,9 +262,10 @@ use
Instead of Instead of
```shell ```shell
srun python -u -m torchrun_jsc \ srun env -u CUDA_VISIBLE_DEVICES python -u -m torchrun_jsc \
--nproc_per_node="$DEVICES_PER_NODE" \ --nproc_per_node=gpu \
--nnodes="$SLURM_JOB_NUM_NODES" \ --nnodes="$SLURM_JOB_NUM_NODES" \
--rdzv_id="$SLURM_JOB_ID" \
--rdzv_endpoint="$MASTER_ADDR":"$MASTER_PORT" \ --rdzv_endpoint="$MASTER_ADDR":"$MASTER_PORT" \
--rdzv_backend=c10d \ --rdzv_backend=c10d \
"$curr_dir"/main.py "$@" "$curr_dir"/main.py "$@"
...@@ -273,7 +274,7 @@ srun python -u -m torchrun_jsc \ ...@@ -273,7 +274,7 @@ srun python -u -m torchrun_jsc \
use use
```shell ```shell
srun python -u "$curr_dir"/main.py "$@" srun env -u CUDA_VISIBLE_DEVICES python -u "$curr_dir"/main.py "$@"
``` ```
Additionally, if using PyTorch Lightning, you may encounter issues Additionally, if using PyTorch Lightning, you may encounter issues
...@@ -301,6 +302,7 @@ def patch_lightning_slurm_master_addr(): ...@@ -301,6 +302,7 @@ def patch_lightning_slurm_master_addr():
'juwelsbooster', 'juwelsbooster',
'juwels', 'juwels',
'jurecadc', 'jurecadc',
'jusuf',
]: ]:
return return
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment