run_oracle_server_lkval.sh

#!/bin/bash
worker_id=$1
world_size=$2
master_ip=$3
world_size_trainers=$4
batch_size=$5
lk_val=$6
git pull
source activate pytorch_latest_p37
python oracle_cacher.py --prefetch --lookahead-value $lk_val --arch-mlp-bot="13-512-256-64-16" --data-generation=dataset --data-set=kaggle --raw-data-file=/home/saurabh/Work/dlrm_simulator/criteo_kaggle_data/train.txt  --processed-data-file=/home/saurabh/Work/dlrm_simulator/criteo_kaggle_data/kaggleAdDisplayChallenge_processed.npz  --mini-batch-size=$batch_size --dataset-multiprocessing --worker-id $worker_id --world-size $world_size --cache --master-ip $master_ip --world-size-trainers $world_size_trainers --processed-csv  /home/ubuntu/kaggle_criteo_weekly.txt --cache-size 6000000 2>&1 | tee out