sbatch_base_simulations.sh

#!/bin/bash

# Default resources are 1 core with 2.8GB of memory per core.

# job name:
#SBATCH -J train_dat

# priority
##SBATCH --account=bibs-frankmj-condo
#SBATCH --account=carney-frankmj-condo

# output file
#SBATCH --output /users/afengler/batch_job_out/tpl_1_%A_%a.out

# Request runtime, memory, cores
#SBATCH --time=48:00:00
#SBATCH --mem=32G
#SBATCH -c 10
#SBATCH -N 1
##SBATCH --array=1-300  # DO THIS FOR TRAINING DATA GENERATION
#SBATCH --array=1-1

# --------------------------------------------------------------------------------------
# Sequentially run different kind of models
declare -a dgps=( "ornstein_pos" ) # "race_model" "lca" ) #"ddm_sdv_analytic" "ddm_sdv_red_analytic" ) #( "ddm" "full_ddm" "angle" "weibull_cdf" "ornstein" "levy" )  #( "ddm_mic2_angle" "ddm_par2_angle" ) # ( "ddm_seq2_angle" )
n_samples=( 1024 4096 )   # ( 128 256 512 1024 2048 4096 8192 50000 100000 200000 400000 )
n_choices=( 3 ) #( 4 5 6 )
n_parameter_sets=1000  # cnn 20000 but 150 array   # mlp 10000 but 300 array # KRISHN: 10
n_bins=( 0 ) # KRISHN: n_bins=0
binned=0 # KRISHN: binned=0
machine="ccv" #"ccv" "home" "x7"
datatype="parameter_recovery" #"parameter_recovery" #"parameter_recovery" #"parameter_recovery_hierarchical" "parameter_recovery" "cnn_train" # KRISHN: 'parameter_recovery'
nsubjects=1 #10
mode="test" # "test" #"test"  #'mlp' 'cnn' # KRISHN: 'test'
maxt=20 # 20 for mlp # KRISHN: 20
# outer -------------------------------------
for bins in "${n_bins[@]}"
do
    for n in "${n_samples[@]}"
    do
    # inner ----------------------------tmux---------
        for dgp in "${dgps[@]}"
        do
            if [[ "$dgp" = "lca" ]] || [[ "$dgp" = "race_model" ]];
            then
                for n_c in "${n_choices[@]}"
                    do
                       echo "$dgp"
                       echo $n_c
                       python -u dataset_generator.py --machine $machine \
                                                      --dgplist $dgp \
                                                      --datatype $datatype \
                                                      --nsubjects $nsubjects \
                                                      --nreps 1 \
                                                      --binned $binned \
                                                      --nbins $bins \
                                                      --maxt $maxt \
                                                      --nchoices $n_c \
                                                      --nsamples $n \
                                                      --mode $mode \
                                                      --nparamsets $n_parameter_sets \
                                                      --save 1 \
                                                      --deltat 0.001 \
                                                      --fileid $SLURM_ARRAY_TASK_ID 
                done
            else
                 echo "$dgp"
                 #echo $n_c
                 python -u dataset_generator.py --machine $machine \
                                                --dgplist $dgp \
                                                --nsubjects $nsubjects \
                                                --datatype $datatype \
                                                --nreps 1 \
                                                --binned $binned \
                                                --nbins $bins \
                                                --maxt $maxt \
                                                --nchoices ${n_choices[0]} \
                                                --nsamples $n \
                                                --mode $mode \
                                                --nparamsets $n_parameter_sets \
                                                --save 1  \
                                                --deltat 0.001 \
                                                --fileid $SLURM_ARRAY_TASK_ID
                
            fi
        done
                # normal call to function
    done
done
#---------------------------------------------------------------------------------------