Why this code take more than 5 second if execute first time and when repeat takes less then 0.2s

I begin from a question:

Why first time execution is taking longer time then usual. Should I be worried about malware or similar?

I have code snippets below. I have pickled numpy ndarray in file. I tried to read all and generate total ndarray.

I registered a strange behavior.

Execution time takes more than 5s and only when I turn off Linux machine and I open terminal and i run first time this program.
Next program executions takes less then 0.2s.

Why first time execution is taking longer time then usual.
First I was thinking that it is a extra time for pycache

I am running this code under venv and venv is activated via main.sh

read_pickle.py:

#!/home/****/bin/python3
# coding=utf-8

from os import listdir
from os.path import isfile, join
import multiprocessing
import numpy as np
import pandas as pd
import time as t
from utilsC0 import read_pickle_from_file
from enums import *


def process(file):
    directory = "some localization" # external hdd
    path = directory + f"/{file}"
    arr = read_pickle_from_file(path)
    return arr


def f2():
    directory = "some localization" # external hdd
    onlyfiles = [f for f in listdir(directory) if isfile(join(directory, f))]

    with multiprocessing.Pool() as p: 
        result = p.map(process, onlyfiles)

    result = np.vstack(result)

    print("End f2")


if __name__ == "__main__":
    t1 = t.time()
    f2()
    t2 = t.time()
    print(f"{t2 - t1}")

And simple main.sh:

#!/usr/bin/env bash

source /home/***/bin/activate
python3 src/read_pickle.py