generate_cnn_experiments.py

import os
import sys
generated_name = str(sys.argv[1])
log_path = str(sys.argv[2])
save_path = str(sys.argv[3])
network_type = str(sys.argv[4])
dataset = str(sys.argv[5])
exp_name = str(sys.argv[6])
old_savename = "None"
base_call = "python cnn.py --network_type " + network_type + " --dataset " + dataset +" --learning_rate 0.00025 --n_inference_steps 100"
output_file = open(generated_name, "w")
seeds = 5
for s in range(seeds):
    lpath = log_path + "/"+str(exp_name) + "/" + str(s)
    spath = save_path + "/" + str(exp_name) + "/" + str(s)
    if old_savename != "None":
        old_savepath = save_path + "/" + str(old_savename) + "/" + str(s)
    else:
        old_savepath = "None"
    final_call = base_call + " --logdir " + str(lpath) + " --savedir " + str(spath) + " " +" --old_savedir " + str(old_savepath)
    print(final_call)
    print(final_call, file=output_file)