training_loop.sh

#!/usr/bin/env bash
# Expected to be called with X_DO_LAUNCH=1
MODELS=( "EvaluationVGG" )
BASE_DIR=${1:?First argument needs to be path to folder containing experiment configurations}
EXPERIMENTS_DIR=( $(ls "$BASE_DIR") )
CHALLENGE="cifar"
NUM_CLASSES=10
NUMBER_OF_EPOCHS=30

for n in "${EXPERIMENTS_DIR[@]}"; do
    exp_name="${n##*/}"
    
    echo "$exp_name"
    for model in "${MODELS[@]}"; do
        for evaluationtype in "buggy" "corrected"; do
            # echo "RUNNING EXP ${exp_name} - ${model} for evaluation type $evaluationtype."
            # envfile_path="${BASE_DIR}/${exp_name}/${evaluationtype}.env"
            # echo "Content of envfile: "
            # echo "------"
            # cat "$envfile_path"
            # echo "======"

            NUMBER_OF_EPOCHS=${NUMBER_OF_EPOCHS} X_DO_LAUNCH=${X_DO_LAUNCH:-0} ./run_launcher.sh ${exp_name} ${evaluationtype} $model
            # echo docker run --name "${exp_name}_${evaluationtype}_${model}" --mount source=pip-cache,target=/pip-cache --mount source=results,target=/results --mount source=build-vol,target=/builds --gpus all --env-file "$envfile_path" --env MODEL_NAME="$model" --env CHALLENGE="$CHALLENGE" --env NUM_CLASSES="$NUM_CLASSES" -it emiliorivera/ml-frameworks:eval100_client
        done
    done
done