ML-Purdue
/
aqacf-ml-main


  
1

	
2

	
3

	
4

	
5

	
6

	
7

	
8

	
9

	
10

	
11

	
12

	
13

	
14

	
15

	
16

	
17

	
18

	
19

	
20

	
21

	
22

	
23

	
24

	
25

	
            #!/usr/bin/bash
#SBATCH --job-name=epa_ddp
#SBATCH --nodes=1
#SBATCH --cpus-per-task=8
#SBATCH --mem=24G
#SBATCH --account=csso-e
#SBATCH --gres=gpu:2
#SBATCH --time=24:00:00
#SBATCH --output=epa_ddp_%j.out
#SBATCH --error=epa_ddp_%j.err

module load cuda/12.1.1 cudnn/cuda-12.1_8.9 anaconda
conda activate /scratch/gilbreth/rai53/fire

export MLFLOW_TRACKING_USERNAME=$MLFLOW_USERNAME
export MLFLOW_TRACKING_PASSWORD=$MLFLOW_TOKEN
export OMP_NUM_THREADS=8
NCCL_DEBUG=INFO

torchrun --standalone --nnodes=1 --nproc_per_node=gpu \
    epa_seq2seq_train_ddp.py \
    config.json \
    dropout_config.json \
    data.zarr \
    output_directory/