The range of milivolt (mV) unit in PTB Diagnostic ECG Database

I am working on PTB Diagnostic ECG Database and I came out with a controversy for the signals I read. I am using Python wfdb library to read the records. I realized that the milivolt (mV) range of the signals are different if i use wfdb library or if I plot it from the website (physionet ATM). I used the first 10 seconds of the record "patient015/s0152lre" as an example and here is the plot if I use wfdb:

import wfdb
import matplotlib.pyplot as plt
record = wfdb.rdrecord('s0152lre', 
                        channel_names=['i', 'ii',
                                       'iii', 'avr',
                                       'avl', 'avf',
                                       'v1','v2',
                                       'v3','v4',
                                       'v5', 'v6',],
                        sampfrom=0, sampto=10000, 
                        pn_dir='ptbdb/patient015/')
wfdb.plot_wfdb(record=record, title='Patient 1',figsize=(20,20))

mine

For example, here you can see that Lead i starts from about -4.50 mV or lead aVL starts from about -3.50 mV. However, if I want to plot the same graph from the website (physionet ATM), the result is different. The plot from the website:

chart

For the plot from the website, It looks like each lead has an offset value to start from 0 mV. For example, the first peak in lead i goes to -3 mV in my plot but it goes a bit further than +1mV in the website plot. Which way is the better to plot the raw signal? I am not professional for analysis of ECG graphs but as far as I understand normal ECG signals are in the range of +- 1.5mV. Do i need to give an offset value for each signal so that the signal can be within this range?

MIT-LCP / physionet

The range of milivolt (mV) unit in PTB Diagnostic ECG Database #131