naref 0.0.4

A quantum & classical reservoir handler for Temporal Series Prediction

NAREF

Neutral Atom Renewable Energy Forecasting

This file showcases quantum and classical reservoirs in a simple example.

import naref

from naref.classical.hyper import HyperC
from naref.quantum.hyper import HyperQ
from naref.quantum import QRC
from naref.classical import CRC
from naref import data

qrc = QRC(
    sample_len=8,
    nb_atoms=9,
    N_samples=1024,
    inp_duration=1000,
    reset_rate=0,
    geometry="grid_lattice_centred",
    atom_distance=15,
    input_type="mackey",
    test_len=data.real_test_len,
    train_len=data.real_train_len,
    verbose=True)

qrc.build_model(data.mackey[:250], data.mackey[250:280])
qrc.show_test_prediction()
qrc.show_train_prediction()

qrc.get_error_train()
qrc.get_error_test()

Building Quantum Reservoir...
10 training steps finished, training time : 24.890s
20 training steps finished, training time : 24.411s
30 training steps finished, training time : 24.441s
40 training steps finished, training time : 25.471s
50 training steps finished, training time : 25.757s
60 training steps finished, training time : 25.194s
70 training steps finished, training time : 24.324s
80 training steps finished, training time : 24.378s
90 training steps finished, training time : 31.824s
100 training steps finished, training time : 29.980s
110 training steps finished, training time : 28.739s
120 training steps finished, training time : 27.935s
130 training steps finished, training time : 30.140s
140 training steps finished, training time : 31.482s
150 training steps finished, training time : 33.550s
160 training steps finished, training time : 33.109s
170 training steps finished, training time : 31.198s
180 training steps finished, training time : 32.346s
190 training steps finished, training time : 34.320s
200 training steps finished, training time : 37.271s
210 training steps finished, training time : 37.513s
220 training steps finished, training time : 33.800s
230 training steps finished, training time : 35.306s
240 training steps finished, training time : 34.594s
Training finished. Total training time: 731.384s
Train error (Mean absolute error regression loss):  5.165748453421183e-15
10 testing steps finished, testing time : 36.515s
20 testing steps finished, testing time : 34.390s
Test error (Mean absolute error regression loss):  0.014927562760331666

0.014927562760331666

crc = CRC(
    sample_len=8,
    nb_neurons=9)
crc.build_model(data.mackey[:250], data.mackey[250:280])
crc.show_test_prediction()
crc.show_train_prediction()
crc.get_error_test()

Building Classical Reservoir...

0.047504000683266046

crc.get_error_train()

0.05308248069948858

crc = CRC(
    sample_len=8,
    nb_neurons=15)
crc.build_model(data.mackey[:500], data.mackey[500:530])
crc.show_test_prediction()
crc.show_train_prediction()
crc.get_error_test()

Building Classical Reservoir...

0.009784391178625047

crc.get_error_train()

0.008250889994704327