LSTM Example

[1]:
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from scalecast.Forecaster import Forecaster

sns.set(rc={'figure.figsize':(15,8)})

Data preprocessing

[2]:
data = pd.read_csv('AirPassengers.csv',parse_dates=['Month'])
[3]:
data.head()
[3]:
Month #Passengers
0 1949-01-01 112
1 1949-02-01 118
2 1949-03-01 132
3 1949-04-01 129
4 1949-05-01 121
[4]:
data.shape
[4]:
(144, 2)
[5]:
data['Month'].min()
[5]:
Timestamp('1949-01-01 00:00:00')
[6]:
data['Month'].max()
[6]:
Timestamp('1960-12-01 00:00:00')

EDA

[7]:
f = Forecaster(y=data['#Passengers'],current_dates=data['Month'])
f.plot()
../../_images/Forecaster_examples_LSTM_9_0.png
[8]:
f
[8]:
Forecaster(
    DateStartActuals=1949-01-01T00:00:00.000000000
    DateEndActuals=1960-12-01T00:00:00.000000000
    Freq=MS
    ForecastLength=0
    Xvars=[]
    Differenced=0
    TestLength=1
    ValidationLength=1
    ValidationMetric=rmse
    ForecastsEvaluated=[]
    CILevel=0.95
    BootstrapSamples=100
    CurrentEstimator=None
)
[9]:
f.plot_acf(lags=26)
plt.show()
../../_images/Forecaster_examples_LSTM_11_0.png
[10]:
f.plot_pacf(lags=26)
plt.show()
../../_images/Forecaster_examples_LSTM_12_0.png
[11]:
f.seasonal_decompose().plot()
plt.show()
../../_images/Forecaster_examples_LSTM_13_0.png
[12]:
stat, pval, _, _, _, _ = f.adf_test(full_res=True)
print(stat)
print(pval)
0.8153688792060569
0.9918802434376411

LSTM Model

[13]:
f.set_test_length(12)
f.generate_future_dates(12)
f.set_estimator('lstm')

Attempt 1

[14]:
f.manual_forecast(call_me='lstm_default')
f.plot_test_set(ci=True)
4/4 [==============================] - 1s 2ms/step - loss: 0.3507
5/5 [==============================] - 1s 2ms/step - loss: 0.3386
../../_images/Forecaster_examples_LSTM_18_1.png

Attempt 2

[15]:
f.manual_forecast(call_me='lstm_24lags',lags=24)
f.plot_test_set(ci=True)
3/3 [==============================] - 1s 5ms/step - loss: 0.3823
4/4 [==============================] - 1s 7ms/step - loss: 0.3503
../../_images/Forecaster_examples_LSTM_20_1.png

Attempt 3

[16]:
f.manual_forecast(call_me='lstm_24lags_5epochs',lags=24,epochs=5,validation_split=.2,shuffle=True)
f.plot_test_set(ci=True)
Epoch 1/5
3/3 [==============================] - 2s 193ms/step - loss: 0.4549 - val_loss: 0.1873
Epoch 2/5
3/3 [==============================] - 0s 17ms/step - loss: 0.4461 - val_loss: 0.1813
Epoch 3/5
3/3 [==============================] - 0s 15ms/step - loss: 0.4374 - val_loss: 0.1752
Epoch 4/5
3/3 [==============================] - 0s 16ms/step - loss: 0.4287 - val_loss: 0.1691
Epoch 5/5
3/3 [==============================] - 0s 15ms/step - loss: 0.4199 - val_loss: 0.1631
Epoch 1/5
3/3 [==============================] - 2s 166ms/step - loss: 0.4308 - val_loss: 0.1668
Epoch 2/5
3/3 [==============================] - 0s 16ms/step - loss: 0.4225 - val_loss: 0.1609
Epoch 3/5
3/3 [==============================] - 0s 15ms/step - loss: 0.4142 - val_loss: 0.1551
Epoch 4/5
3/3 [==============================] - 0s 17ms/step - loss: 0.4059 - val_loss: 0.1493
Epoch 5/5
3/3 [==============================] - 0s 15ms/step - loss: 0.3975 - val_loss: 0.1434
../../_images/Forecaster_examples_LSTM_22_1.png

Attempt 4

[17]:
from tensorflow.keras.callbacks import EarlyStopping
f.manual_forecast(call_me='lstm_24lags_earlystop_3layers',
                  lags=24,
                  epochs=25,
                  validation_split=.2,
                  shuffle=True,
                  callbacks=EarlyStopping(monitor='val_loss',
                                          patience=5),
                  lstm_layer_sizes=(16,16,16),
                  dropout=(0,0,0))
f.plot_test_set(ci=True)
Epoch 1/25
3/3 [==============================] - 4s 432ms/step - loss: 0.4674 - val_loss: 0.1867
Epoch 2/25
3/3 [==============================] - 0s 27ms/step - loss: 0.4491 - val_loss: 0.1720
Epoch 3/25
3/3 [==============================] - 0s 26ms/step - loss: 0.4278 - val_loss: 0.1543
Epoch 4/25
3/3 [==============================] - 0s 26ms/step - loss: 0.4020 - val_loss: 0.1330
Epoch 5/25
3/3 [==============================] - 0s 28ms/step - loss: 0.3686 - val_loss: 0.1132
Epoch 6/25
3/3 [==============================] - 0s 26ms/step - loss: 0.3275 - val_loss: 0.1103
Epoch 7/25
3/3 [==============================] - 0s 26ms/step - loss: 0.2848 - val_loss: 0.1334
Epoch 8/25
3/3 [==============================] - 0s 34ms/step - loss: 0.2618 - val_loss: 0.1648
Epoch 9/25
3/3 [==============================] - 0s 31ms/step - loss: 0.2514 - val_loss: 0.1769
Epoch 10/25
3/3 [==============================] - 0s 27ms/step - loss: 0.2380 - val_loss: 0.1686
Epoch 11/25
3/3 [==============================] - 0s 26ms/step - loss: 0.2179 - val_loss: 0.1489
Epoch 1/25
3/3 [==============================] - 4s 424ms/step - loss: 0.4533 - val_loss: 0.1680
Epoch 2/25
3/3 [==============================] - 0s 33ms/step - loss: 0.4356 - val_loss: 0.1545
Epoch 3/25
3/3 [==============================] - 0s 35ms/step - loss: 0.4168 - val_loss: 0.1382
Epoch 4/25
3/3 [==============================] - 0s 30ms/step - loss: 0.3937 - val_loss: 0.1171
Epoch 5/25
3/3 [==============================] - 0s 32ms/step - loss: 0.3626 - val_loss: 0.0955
Epoch 6/25
3/3 [==============================] - 0s 36ms/step - loss: 0.3200 - val_loss: 0.0889
Epoch 7/25
3/3 [==============================] - 0s 33ms/step - loss: 0.2695 - val_loss: 0.1161
Epoch 8/25
3/3 [==============================] - 0s 33ms/step - loss: 0.2367 - val_loss: 0.1605
Epoch 9/25
3/3 [==============================] - 0s 32ms/step - loss: 0.2269 - val_loss: 0.1860
Epoch 10/25
3/3 [==============================] - 0s 31ms/step - loss: 0.2113 - val_loss: 0.1799
Epoch 11/25
3/3 [==============================] - 0s 39ms/step - loss: 0.1896 - val_loss: 0.1557
../../_images/Forecaster_examples_LSTM_24_1.png

Attempt 5

[18]:
f.manual_forecast(call_me='lstm_24lags_4layers_15epochs',
                  lags=36,
                  batch_size=32,
                  epochs=15,
                  validation_split=.2,
                  shuffle=True,
                  activation='tanh',
                  optimizer='Adam',
                  learning_rate=0.001,
                  lstm_layer_sizes=(72,)*4,
                  dropout=(0,)*4,
                  plot_loss=True)
f.plot_test_set(order_by='LevelTestSetMAPE',models='top_2',ci=True)
Epoch 1/15
3/3 [==============================] - 6s 633ms/step - loss: 0.4843 - val_loss: 0.2002
Epoch 2/15
3/3 [==============================] - 0s 58ms/step - loss: 0.3794 - val_loss: 0.1141
Epoch 3/15
3/3 [==============================] - 0s 61ms/step - loss: 0.2180 - val_loss: 0.1044
Epoch 4/15
3/3 [==============================] - 0s 67ms/step - loss: 0.1162 - val_loss: 0.0598
Epoch 5/15
3/3 [==============================] - 0s 68ms/step - loss: 0.1347 - val_loss: 0.0759
Epoch 6/15
3/3 [==============================] - 0s 65ms/step - loss: 0.1231 - val_loss: 0.0829
Epoch 7/15
3/3 [==============================] - 0s 63ms/step - loss: 0.1053 - val_loss: 0.0593
Epoch 8/15
3/3 [==============================] - ETA: 0s - loss: 0.099 - 0s 62ms/step - loss: 0.0993 - val_loss: 0.0564
Epoch 9/15
3/3 [==============================] - 0s 65ms/step - loss: 0.0973 - val_loss: 0.0532
Epoch 10/15
3/3 [==============================] - 0s 61ms/step - loss: 0.0940 - val_loss: 0.0525
Epoch 11/15
3/3 [==============================] - 0s 58ms/step - loss: 0.0912 - val_loss: 0.0522
Epoch 12/15
3/3 [==============================] - 0s 64ms/step - loss: 0.0908 - val_loss: 0.0568
Epoch 13/15
3/3 [==============================] - 0s 64ms/step - loss: 0.0889 - val_loss: 0.0521
Epoch 14/15
3/3 [==============================] - 0s 67ms/step - loss: 0.0934 - val_loss: 0.0626
Epoch 15/15
3/3 [==============================] - 0s 70ms/step - loss: 0.0921 - val_loss: 0.0564
Epoch 1/15
3/3 [==============================] - 7s 775ms/step - loss: 0.4749 - val_loss: 0.1933
Epoch 2/15
3/3 [==============================] - 0s 62ms/step - loss: 0.4087 - val_loss: 0.1080
Epoch 3/15
3/3 [==============================] - 0s 67ms/step - loss: 0.2353 - val_loss: 0.1847
Epoch 4/15
3/3 [==============================] - 0s 71ms/step - loss: 0.1748 - val_loss: 0.0843
Epoch 5/15
3/3 [==============================] - 0s 69ms/step - loss: 0.1199 - val_loss: 0.0473
Epoch 6/15
3/3 [==============================] - 0s 77ms/step - loss: 0.1194 - val_loss: 0.1083
Epoch 7/15
3/3 [==============================] - 0s 76ms/step - loss: 0.1156 - val_loss: 0.1251
Epoch 8/15
3/3 [==============================] - 0s 76ms/step - loss: 0.1076 - val_loss: 0.0666
Epoch 9/15
3/3 [==============================] - 0s 72ms/step - loss: 0.0979 - val_loss: 0.0580
Epoch 10/15
3/3 [==============================] - 0s 72ms/step - loss: 0.0946 - val_loss: 0.0728
Epoch 11/15
3/3 [==============================] - 0s 70ms/step - loss: 0.0926 - val_loss: 0.0521
Epoch 12/15
3/3 [==============================] - 0s 70ms/step - loss: 0.0865 - val_loss: 0.0472
Epoch 13/15
3/3 [==============================] - 0s 70ms/step - loss: 0.0870 - val_loss: 0.0475
Epoch 14/15
3/3 [==============================] - 0s 68ms/step - loss: 0.0845 - val_loss: 0.0498
Epoch 15/15
3/3 [==============================] - 0s 69ms/step - loss: 0.0831 - val_loss: 0.0576
../../_images/Forecaster_examples_LSTM_26_1.png
../../_images/Forecaster_examples_LSTM_26_2.png

Attempt 6

[19]:
f.manual_forecast(call_me='lstm_best',
                  lags=36,
                  batch_size=16,
                  epochs=300,
                  validation_split=.2,
                  shuffle=True,
                  activation='tanh',
                  optimizer='Adam',
                  learning_rate=0.001,
                  lstm_layer_sizes=(100,)*15,
                  dropout=(0,)*15,
                  plot_loss=True)
f.plot_test_set(order_by='LevelTestSetMAPE',models='top_2',ci=True)
Epoch 1/300
5/5 [==============================] - 26s 1s/step - loss: 0.4565 - val_loss: 0.2858
Epoch 2/300
5/5 [==============================] - 1s 295ms/step - loss: 0.1735 - val_loss: 0.2274
Epoch 3/300
5/5 [==============================] - 2s 302ms/step - loss: 0.1574 - val_loss: 0.2142
Epoch 4/300
5/5 [==============================] - 1s 300ms/step - loss: 0.1419 - val_loss: 0.3462
Epoch 5/300
5/5 [==============================] - 2s 308ms/step - loss: 0.1704 - val_loss: 0.1709
Epoch 6/300
5/5 [==============================] - 2s 306ms/step - loss: 0.1314 - val_loss: 0.2677
Epoch 7/300
5/5 [==============================] - 2s 310ms/step - loss: 0.1403 - val_loss: 0.0713
Epoch 8/300
5/5 [==============================] - 1s 300ms/step - loss: 0.1581 - val_loss: 0.1365
Epoch 9/300
5/5 [==============================] - 2s 307ms/step - loss: 0.1226 - val_loss: 0.0815
Epoch 10/300
5/5 [==============================] - 2s 304ms/step - loss: 0.1073 - val_loss: 0.1793
Epoch 11/300
5/5 [==============================] - 2s 311ms/step - loss: 0.1131 - val_loss: 0.1221
Epoch 12/300
5/5 [==============================] - 2s 325ms/step - loss: 0.1080 - val_loss: 0.0530
Epoch 13/300
5/5 [==============================] - 2s 312ms/step - loss: 0.1002 - val_loss: 0.0590
Epoch 14/300
5/5 [==============================] - 2s 317ms/step - loss: 0.0985 - val_loss: 0.0894
Epoch 15/300
5/5 [==============================] - 2s 322ms/step - loss: 0.1030 - val_loss: 0.0711
Epoch 16/300
5/5 [==============================] - 1s 284ms/step - loss: 0.0952 - val_loss: 0.0780
Epoch 17/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0967 - val_loss: 0.0663
Epoch 18/300
5/5 [==============================] - 1s 293ms/step - loss: 0.0916 - val_loss: 0.0725
Epoch 19/300
5/5 [==============================] - 2s 327ms/step - loss: 0.0938 - val_loss: 0.0640
Epoch 20/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0958 - val_loss: 0.0742
Epoch 21/300
5/5 [==============================] - 1s 287ms/step - loss: 0.0919 - val_loss: 0.0649
Epoch 22/300
5/5 [==============================] - 2s 303ms/step - loss: 0.0925 - val_loss: 0.0933
Epoch 23/300
5/5 [==============================] - 2s 315ms/step - loss: 0.0898 - val_loss: 0.0617
Epoch 24/300
5/5 [==============================] - 1s 285ms/step - loss: 0.0925 - val_loss: 0.0576
Epoch 25/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0914 - val_loss: 0.0596
Epoch 26/300
5/5 [==============================] - 1s 299ms/step - loss: 0.0920 - val_loss: 0.0646
Epoch 27/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0906 - val_loss: 0.0591
Epoch 28/300
5/5 [==============================] - 2s 312ms/step - loss: 0.0903 - val_loss: 0.0641
Epoch 29/300
5/5 [==============================] - 1s 291ms/step - loss: 0.0899 - val_loss: 0.0660
Epoch 30/300
5/5 [==============================] - 1s 273ms/step - loss: 0.0885 - val_loss: 0.0584
Epoch 31/300
5/5 [==============================] - 2s 314ms/step - loss: 0.0881 - val_loss: 0.0593
Epoch 32/300
5/5 [==============================] - 2s 312ms/step - loss: 0.0896 - val_loss: 0.0569
Epoch 33/300
5/5 [==============================] - 2s 308ms/step - loss: 0.0899 - val_loss: 0.0607
Epoch 34/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0909 - val_loss: 0.0546
Epoch 35/300
5/5 [==============================] - 2s 333ms/step - loss: 0.0934 - val_loss: 0.0601
Epoch 36/300
5/5 [==============================] - 2s 333ms/step - loss: 0.0900 - val_loss: 0.0716
Epoch 37/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0905 - val_loss: 0.0585
Epoch 38/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0893 - val_loss: 0.0731
Epoch 39/300
5/5 [==============================] - 2s 369ms/step - loss: 0.0886 - val_loss: 0.0552
Epoch 40/300
5/5 [==============================] - 2s 313ms/step - loss: 0.0883 - val_loss: 0.0563
Epoch 41/300
5/5 [==============================] - 1s 293ms/step - loss: 0.0890 - val_loss: 0.0631
Epoch 42/300
5/5 [==============================] - 2s 306ms/step - loss: 0.0876 - val_loss: 0.0577
Epoch 43/300
5/5 [==============================] - 2s 299ms/step - loss: 0.0858 - val_loss: 0.0623
Epoch 44/300
5/5 [==============================] - 1s 291ms/step - loss: 0.0883 - val_loss: 0.0586
Epoch 45/300
5/5 [==============================] - 1s 289ms/step - loss: 0.0870 - val_loss: 0.0551
Epoch 46/300
5/5 [==============================] - 1s 287ms/step - loss: 0.0899 - val_loss: 0.0592
Epoch 47/300
5/5 [==============================] - 1s 264ms/step - loss: 0.0892 - val_loss: 0.0573
Epoch 48/300
5/5 [==============================] - 1s 270ms/step - loss: 0.0865 - val_loss: 0.0590
Epoch 49/300
5/5 [==============================] - 1s 267ms/step - loss: 0.0885 - val_loss: 0.0564
Epoch 50/300
5/5 [==============================] - 1s 269ms/step - loss: 0.0874 - val_loss: 0.0546
Epoch 51/300
5/5 [==============================] - 1s 270ms/step - loss: 0.0875 - val_loss: 0.0550
Epoch 52/300
5/5 [==============================] - 1s 271ms/step - loss: 0.0868 - val_loss: 0.0613
Epoch 53/300
5/5 [==============================] - 1s 277ms/step - loss: 0.0902 - val_loss: 0.0560
Epoch 54/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0881 - val_loss: 0.0524
Epoch 55/300
5/5 [==============================] - 2s 300ms/step - loss: 0.0877 - val_loss: 0.0519
Epoch 56/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0876 - val_loss: 0.0557
Epoch 57/300
5/5 [==============================] - 1s 274ms/step - loss: 0.0874 - val_loss: 0.0610
Epoch 58/300
5/5 [==============================] - 1s 288ms/step - loss: 0.0913 - val_loss: 0.0584
Epoch 59/300
5/5 [==============================] - 1s 294ms/step - loss: 0.0889 - val_loss: 0.0546
Epoch 60/300
5/5 [==============================] - 2s 313ms/step - loss: 0.0857 - val_loss: 0.0541
Epoch 61/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0872 - val_loss: 0.0535
Epoch 62/300
5/5 [==============================] - 2s 383ms/step - loss: 0.0882 - val_loss: 0.0578
Epoch 63/300
5/5 [==============================] - 1s 275ms/step - loss: 0.0863 - val_loss: 0.0576
Epoch 64/300
5/5 [==============================] - 1s 272ms/step - loss: 0.0883 - val_loss: 0.0591
Epoch 65/300
5/5 [==============================] - 1s 295ms/step - loss: 0.0877 - val_loss: 0.0565
Epoch 66/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0869 - val_loss: 0.0550
Epoch 67/300
5/5 [==============================] - 2s 322ms/step - loss: 0.0882 - val_loss: 0.0571
Epoch 68/300
5/5 [==============================] - 2s 408ms/step - loss: 0.0858 - val_loss: 0.0552
Epoch 69/300
5/5 [==============================] - 2s 394ms/step - loss: 0.0866 - val_loss: 0.0556
Epoch 70/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0874 - val_loss: 0.0548
Epoch 71/300
5/5 [==============================] - 2s 305ms/step - loss: 0.0875 - val_loss: 0.0580
Epoch 72/300
5/5 [==============================] - 2s 322ms/step - loss: 0.0879 - val_loss: 0.0553
Epoch 73/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0870 - val_loss: 0.0584
Epoch 74/300
5/5 [==============================] - 2s 305ms/step - loss: 0.0926 - val_loss: 0.0734
Epoch 75/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0872 - val_loss: 0.0558
Epoch 76/300
5/5 [==============================] - 2s 299ms/step - loss: 0.0868 - val_loss: 0.0584
Epoch 77/300
5/5 [==============================] - 2s 311ms/step - loss: 0.0863 - val_loss: 0.0710
Epoch 78/300
5/5 [==============================] - 1s 296ms/step - loss: 0.0866 - val_loss: 0.0588
Epoch 79/300
5/5 [==============================] - 2s 316ms/step - loss: 0.0872 - val_loss: 0.0568
Epoch 80/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0883 - val_loss: 0.0607
Epoch 81/300
5/5 [==============================] - 2s 466ms/step - loss: 0.0870 - val_loss: 0.0622
Epoch 82/300
5/5 [==============================] - 2s 423ms/step - loss: 0.0865 - val_loss: 0.0584
Epoch 83/300
5/5 [==============================] - 2s 418ms/step - loss: 0.0867 - val_loss: 0.0571
Epoch 84/300
5/5 [==============================] - 2s 358ms/step - loss: 0.0870 - val_loss: 0.0603
Epoch 85/300
5/5 [==============================] - 2s 393ms/step - loss: 0.0883 - val_loss: 0.0613
Epoch 86/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0888 - val_loss: 0.0602
Epoch 87/300
5/5 [==============================] - 2s 387ms/step - loss: 0.0867 - val_loss: 0.0615
Epoch 88/300
5/5 [==============================] - 2s 352ms/step - loss: 0.0861 - val_loss: 0.0574
Epoch 89/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0868 - val_loss: 0.0564
Epoch 90/300
5/5 [==============================] - 2s 444ms/step - loss: 0.0878 - val_loss: 0.0558
Epoch 91/300
5/5 [==============================] - 2s 421ms/step - loss: 0.0866 - val_loss: 0.0544
Epoch 92/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0888 - val_loss: 0.0537
Epoch 93/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0889 - val_loss: 0.0557
Epoch 94/300
5/5 [==============================] - 2s 374ms/step - loss: 0.0884 - val_loss: 0.0586
Epoch 95/300
5/5 [==============================] - 2s 392ms/step - loss: 0.0864 - val_loss: 0.0571
Epoch 96/300
5/5 [==============================] - 3s 579ms/step - loss: 0.0883 - val_loss: 0.0570
Epoch 97/300
5/5 [==============================] - 3s 477ms/step - loss: 0.0881 - val_loss: 0.0622
Epoch 98/300
5/5 [==============================] - 3s 533ms/step - loss: 0.0865 - val_loss: 0.0563
Epoch 99/300
5/5 [==============================] - 2s 501ms/step - loss: 0.0876 - val_loss: 0.0544
Epoch 100/300
5/5 [==============================] - 3s 594ms/step - loss: 0.0864 - val_loss: 0.0555
Epoch 101/300
5/5 [==============================] - 2s 476ms/step - loss: 0.0918 - val_loss: 0.0585
Epoch 102/300
5/5 [==============================] - 2s 441ms/step - loss: 0.0909 - val_loss: 0.0691
Epoch 103/300
5/5 [==============================] - 2s 452ms/step - loss: 0.0906 - val_loss: 0.0664
Epoch 104/300
5/5 [==============================] - 2s 429ms/step - loss: 0.0885 - val_loss: 0.0652
Epoch 105/300
5/5 [==============================] - 2s 476ms/step - loss: 0.0859 - val_loss: 0.0682
Epoch 106/300
5/5 [==============================] - 2s 398ms/step - loss: 0.0881 - val_loss: 0.0664
Epoch 107/300
5/5 [==============================] - 2s 397ms/step - loss: 0.0873 - val_loss: 0.0587
Epoch 108/300
5/5 [==============================] - 2s 421ms/step - loss: 0.0878 - val_loss: 0.0524
Epoch 109/300
5/5 [==============================] - 2s 403ms/step - loss: 0.0908 - val_loss: 0.0521
Epoch 110/300
5/5 [==============================] - 2s 445ms/step - loss: 0.0873 - val_loss: 0.0549
Epoch 111/300
5/5 [==============================] - 2s 442ms/step - loss: 0.0867 - val_loss: 0.0638
Epoch 112/300
5/5 [==============================] - 2s 401ms/step - loss: 0.0884 - val_loss: 0.0573
Epoch 113/300
5/5 [==============================] - 2s 428ms/step - loss: 0.0855 - val_loss: 0.0556
Epoch 114/300
5/5 [==============================] - 2s 393ms/step - loss: 0.0846 - val_loss: 0.0563
Epoch 115/300
5/5 [==============================] - 2s 383ms/step - loss: 0.0852 - val_loss: 0.0593
Epoch 116/300
5/5 [==============================] - 2s 385ms/step - loss: 0.0856 - val_loss: 0.0605
Epoch 117/300
5/5 [==============================] - 2s 391ms/step - loss: 0.0848 - val_loss: 0.0629
Epoch 118/300
5/5 [==============================] - 2s 379ms/step - loss: 0.0846 - val_loss: 0.0638
Epoch 119/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0849 - val_loss: 0.0612
Epoch 120/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0843 - val_loss: 0.0572
Epoch 121/300
5/5 [==============================] - 2s 418ms/step - loss: 0.0835 - val_loss: 0.0567
Epoch 122/300
5/5 [==============================] - 2s 358ms/step - loss: 0.0865 - val_loss: 0.0559
Epoch 123/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0828 - val_loss: 0.0549
Epoch 124/300
5/5 [==============================] - 2s 359ms/step - loss: 0.0863 - val_loss: 0.0609
Epoch 125/300
5/5 [==============================] - 2s 356ms/step - loss: 0.0886 - val_loss: 0.0555
Epoch 126/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0863 - val_loss: 0.0637
Epoch 127/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0869 - val_loss: 0.0889
Epoch 128/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0863 - val_loss: 0.0541
Epoch 129/300
5/5 [==============================] - 2s 371ms/step - loss: 0.0859 - val_loss: 0.0522
Epoch 130/300
5/5 [==============================] - 2s 364ms/step - loss: 0.0876 - val_loss: 0.0573
Epoch 131/300
5/5 [==============================] - 2s 378ms/step - loss: 0.0864 - val_loss: 0.0683
Epoch 132/300
5/5 [==============================] - 2s 378ms/step - loss: 0.0862 - val_loss: 0.0700
Epoch 133/300
5/5 [==============================] - 2s 369ms/step - loss: 0.0884 - val_loss: 0.0646
Epoch 134/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0858 - val_loss: 0.0687
Epoch 135/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0850 - val_loss: 0.0563
Epoch 136/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0849 - val_loss: 0.0531
Epoch 137/300
5/5 [==============================] - 2s 358ms/step - loss: 0.0946 - val_loss: 0.0528
Epoch 138/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0879 - val_loss: 0.0830
Epoch 139/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0903 - val_loss: 0.0556
Epoch 140/300
5/5 [==============================] - 2s 364ms/step - loss: 0.0879 - val_loss: 0.0541
Epoch 141/300
5/5 [==============================] - 2s 379ms/step - loss: 0.0900 - val_loss: 0.0787
Epoch 142/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0886 - val_loss: 0.0597
Epoch 143/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0862 - val_loss: 0.0585
Epoch 144/300
5/5 [==============================] - 2s 377ms/step - loss: 0.0895 - val_loss: 0.0844
Epoch 145/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0917 - val_loss: 0.0628
Epoch 146/300
5/5 [==============================] - 2s 378ms/step - loss: 0.0899 - val_loss: 0.0592
Epoch 147/300
5/5 [==============================] - 2s 504ms/step - loss: 0.0899 - val_loss: 0.0569
Epoch 148/300
5/5 [==============================] - 2s 380ms/step - loss: 0.0872 - val_loss: 0.0618
Epoch 149/300
5/5 [==============================] - 2s 391ms/step - loss: 0.0885 - val_loss: 0.0658
Epoch 150/300
5/5 [==============================] - 2s 401ms/step - loss: 0.0918 - val_loss: 0.0552
Epoch 151/300
5/5 [==============================] - 2s 408ms/step - loss: 0.0913 - val_loss: 0.0785
Epoch 152/300
5/5 [==============================] - 2s 413ms/step - loss: 0.0960 - val_loss: 0.0609
Epoch 153/300
5/5 [==============================] - 2s 488ms/step - loss: 0.0899 - val_loss: 0.0641
Epoch 154/300
5/5 [==============================] - 2s 499ms/step - loss: 0.0873 - val_loss: 0.0621
Epoch 155/300
5/5 [==============================] - 2s 385ms/step - loss: 0.0867 - val_loss: 0.0742
Epoch 156/300
5/5 [==============================] - 2s 384ms/step - loss: 0.0869 - val_loss: 0.0611
Epoch 157/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0877 - val_loss: 0.0621
Epoch 158/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0881 - val_loss: 0.0633
Epoch 159/300
5/5 [==============================] - 2s 375ms/step - loss: 0.0871 - val_loss: 0.0595
Epoch 160/300
5/5 [==============================] - 2s 432ms/step - loss: 0.0863 - val_loss: 0.0566
Epoch 161/300
5/5 [==============================] - 2s 394ms/step - loss: 0.0875 - val_loss: 0.0606
Epoch 162/300
5/5 [==============================] - 2s 368ms/step - loss: 0.0872 - val_loss: 0.0583
Epoch 163/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0870 - val_loss: 0.0556
Epoch 164/300
5/5 [==============================] - 2s 427ms/step - loss: 0.0877 - val_loss: 0.0612
Epoch 165/300
5/5 [==============================] - 2s 378ms/step - loss: 0.0866 - val_loss: 0.0600
Epoch 166/300
5/5 [==============================] - 2s 492ms/step - loss: 0.0866 - val_loss: 0.0599
Epoch 167/300
5/5 [==============================] - 2s 467ms/step - loss: 0.0874 - val_loss: 0.0660
Epoch 168/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0866 - val_loss: 0.0592
Epoch 169/300
5/5 [==============================] - 2s 398ms/step - loss: 0.0871 - val_loss: 0.0578
Epoch 170/300
5/5 [==============================] - 2s 410ms/step - loss: 0.0865 - val_loss: 0.0568
Epoch 171/300
5/5 [==============================] - 2s 468ms/step - loss: 0.0868 - val_loss: 0.0611
Epoch 172/300
5/5 [==============================] - 2s 469ms/step - loss: 0.0852 - val_loss: 0.0610
Epoch 173/300
5/5 [==============================] - 2s 411ms/step - loss: 0.0870 - val_loss: 0.0617
Epoch 174/300
5/5 [==============================] - 2s 375ms/step - loss: 0.0859 - val_loss: 0.0581
Epoch 175/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0853 - val_loss: 0.0565
Epoch 176/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0851 - val_loss: 0.0572
Epoch 177/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0860 - val_loss: 0.0555
Epoch 178/300
5/5 [==============================] - 2s 332ms/step - loss: 0.0854 - val_loss: 0.0539
Epoch 179/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0863 - val_loss: 0.0548
Epoch 180/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0862 - val_loss: 0.0536
Epoch 181/300
5/5 [==============================] - 2s 394ms/step - loss: 0.0851 - val_loss: 0.0549
Epoch 182/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0848 - val_loss: 0.0547
Epoch 183/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0860 - val_loss: 0.0551
Epoch 184/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0858 - val_loss: 0.0583
Epoch 185/300
5/5 [==============================] - 2s 347ms/step - loss: 0.0877 - val_loss: 0.0597
Epoch 186/300
5/5 [==============================] - 2s 422ms/step - loss: 0.0870 - val_loss: 0.0728
Epoch 187/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0857 - val_loss: 0.0647
Epoch 188/300
5/5 [==============================] - 2s 334ms/step - loss: 0.0856 - val_loss: 0.0623
Epoch 189/300
5/5 [==============================] - 2s 366ms/step - loss: 0.0856 - val_loss: 0.0595
Epoch 190/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0876 - val_loss: 0.0546
Epoch 191/300
5/5 [==============================] - 2s 346ms/step - loss: 0.0850 - val_loss: 0.0541
Epoch 192/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0863 - val_loss: 0.0563
Epoch 193/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0861 - val_loss: 0.0562
Epoch 194/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0843 - val_loss: 0.0571
Epoch 195/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0834 - val_loss: 0.0564
Epoch 196/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0848 - val_loss: 0.0584
Epoch 197/300
5/5 [==============================] - 2s 420ms/step - loss: 0.0843 - val_loss: 0.0722
Epoch 198/300
5/5 [==============================] - 2s 429ms/step - loss: 0.0846 - val_loss: 0.0570
Epoch 199/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0930 - val_loss: 0.0813
Epoch 200/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0957 - val_loss: 0.0562
Epoch 201/300
5/5 [==============================] - 2s 439ms/step - loss: 0.0869 - val_loss: 0.0738
Epoch 202/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0880 - val_loss: 0.0631
Epoch 203/300
5/5 [==============================] - 2s 356ms/step - loss: 0.0862 - val_loss: 0.0668
Epoch 204/300
5/5 [==============================] - 2s 486ms/step - loss: 0.0876 - val_loss: 0.0704
Epoch 205/300
5/5 [==============================] - 2s 417ms/step - loss: 0.0901 - val_loss: 0.0926
Epoch 206/300
5/5 [==============================] - 2s 370ms/step - loss: 0.0925 - val_loss: 0.0622
Epoch 207/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0878 - val_loss: 0.0545
Epoch 208/300
5/5 [==============================] - 2s 414ms/step - loss: 0.0879 - val_loss: 0.0635
Epoch 209/300
5/5 [==============================] - 2s 407ms/step - loss: 0.0869 - val_loss: 0.0558
Epoch 210/300
5/5 [==============================] - 2s 400ms/step - loss: 0.0865 - val_loss: 0.0552
Epoch 211/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0863 - val_loss: 0.0572
Epoch 212/300
5/5 [==============================] - 2s 412ms/step - loss: 0.0847 - val_loss: 0.0554
Epoch 213/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0849 - val_loss: 0.0560
Epoch 214/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0855 - val_loss: 0.0555
Epoch 215/300
5/5 [==============================] - 2s 476ms/step - loss: 0.0837 - val_loss: 0.0598
Epoch 216/300
5/5 [==============================] - 2s 364ms/step - loss: 0.0826 - val_loss: 0.0623
Epoch 217/300
5/5 [==============================] - 2s 347ms/step - loss: 0.0812 - val_loss: 0.0585
Epoch 218/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0813 - val_loss: 0.0564
Epoch 219/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0825 - val_loss: 0.0554
Epoch 220/300
5/5 [==============================] - 2s 482ms/step - loss: 0.0800 - val_loss: 0.0559
Epoch 221/300
5/5 [==============================] - 2s 419ms/step - loss: 0.0821 - val_loss: 0.0570
Epoch 222/300
5/5 [==============================] - 2s 451ms/step - loss: 0.0795 - val_loss: 0.0570
Epoch 223/300
5/5 [==============================] - 2s 400ms/step - loss: 0.0798 - val_loss: 0.0627
Epoch 224/300
5/5 [==============================] - 2s 379ms/step - loss: 0.0810 - val_loss: 0.0584
Epoch 225/300
5/5 [==============================] - 2s 412ms/step - loss: 0.0801 - val_loss: 0.0594
Epoch 226/300
5/5 [==============================] - 2s 397ms/step - loss: 0.0805 - val_loss: 0.0559
Epoch 227/300
5/5 [==============================] - 2s 348ms/step - loss: 0.0822 - val_loss: 0.0587
Epoch 228/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0821 - val_loss: 0.0598
Epoch 229/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0818 - val_loss: 0.0577
Epoch 230/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0842 - val_loss: 0.0541
Epoch 231/300
5/5 [==============================] - 2s 329ms/step - loss: 0.0792 - val_loss: 0.0562
Epoch 232/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0856 - val_loss: 0.0546
Epoch 233/300
5/5 [==============================] - 2s 333ms/step - loss: 0.0808 - val_loss: 0.0621
Epoch 234/300
5/5 [==============================] - 2s 414ms/step - loss: 0.0857 - val_loss: 0.0618
Epoch 235/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0884 - val_loss: 0.0584
Epoch 236/300
5/5 [==============================] - 2s 404ms/step - loss: 0.0846 - val_loss: 0.0627
Epoch 237/300
5/5 [==============================] - 2s 488ms/step - loss: 0.0861 - val_loss: 0.0557
Epoch 238/300
5/5 [==============================] - 2s 402ms/step - loss: 0.0839 - val_loss: 0.0569
Epoch 239/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0826 - val_loss: 0.0611
Epoch 240/300
5/5 [==============================] - 2s 393ms/step - loss: 0.0839 - val_loss: 0.0558
Epoch 241/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0840 - val_loss: 0.0735
Epoch 242/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0854 - val_loss: 0.0541
Epoch 243/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0855 - val_loss: 0.0656
Epoch 244/300
5/5 [==============================] - 2s 385ms/step - loss: 0.0816 - val_loss: 0.0543
Epoch 245/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0832 - val_loss: 0.0563
Epoch 246/300
5/5 [==============================] - 2s 364ms/step - loss: 0.0847 - val_loss: 0.0526
Epoch 247/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0834 - val_loss: 0.0868
Epoch 248/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0815 - val_loss: 0.0673
Epoch 249/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0818 - val_loss: 0.0598
Epoch 250/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0856 - val_loss: 0.0769
Epoch 251/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0837 - val_loss: 0.0697
Epoch 252/300
5/5 [==============================] - 2s 372ms/step - loss: 0.0802 - val_loss: 0.0550
Epoch 253/300
5/5 [==============================] - 2s 454ms/step - loss: 0.0776 - val_loss: 0.0570
Epoch 254/300
5/5 [==============================] - 2s 467ms/step - loss: 0.0781 - val_loss: 0.0605
Epoch 255/300
5/5 [==============================] - 2s 426ms/step - loss: 0.0809 - val_loss: 0.0533
Epoch 256/300
5/5 [==============================] - 2s 375ms/step - loss: 0.0847 - val_loss: 0.0590
Epoch 257/300
5/5 [==============================] - 2s 374ms/step - loss: 0.0821 - val_loss: 0.0537
Epoch 258/300
5/5 [==============================] - 2s 359ms/step - loss: 0.0823 - val_loss: 0.0530
Epoch 259/300
5/5 [==============================] - 2s 357ms/step - loss: 0.0788 - val_loss: 0.0707
Epoch 260/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0765 - val_loss: 0.0596
Epoch 261/300
5/5 [==============================] - 2s 339ms/step - loss: 0.0810 - val_loss: 0.0573
Epoch 262/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0762 - val_loss: 0.0550
Epoch 263/300
5/5 [==============================] - 2s 372ms/step - loss: 0.0738 - val_loss: 0.0556
Epoch 264/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0727 - val_loss: 0.0559
Epoch 265/300
5/5 [==============================] - 2s 407ms/step - loss: 0.0742 - val_loss: 0.0535
Epoch 266/300
5/5 [==============================] - 2s 402ms/step - loss: 0.0703 - val_loss: 0.0726
Epoch 267/300
5/5 [==============================] - 2s 391ms/step - loss: 0.0714 - val_loss: 0.0699
Epoch 268/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0802 - val_loss: 0.0646
Epoch 269/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0826 - val_loss: 0.0529
Epoch 270/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0724 - val_loss: 0.0797
Epoch 271/300
5/5 [==============================] - 2s 386ms/step - loss: 0.0702 - val_loss: 0.0544
Epoch 272/300
5/5 [==============================] - 2s 356ms/step - loss: 0.0639 - val_loss: 0.0570
Epoch 273/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0683 - val_loss: 0.0590
Epoch 274/300
5/5 [==============================] - 2s 334ms/step - loss: 0.0660 - val_loss: 0.0583
Epoch 275/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0725 - val_loss: 0.0614
Epoch 276/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0713 - val_loss: 0.0617
Epoch 277/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0849 - val_loss: 0.0568
Epoch 278/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0811 - val_loss: 0.0762
Epoch 279/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0724 - val_loss: 0.0611
Epoch 280/300
5/5 [==============================] - 2s 335ms/step - loss: 0.0700 - val_loss: 0.0556
Epoch 281/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0674 - val_loss: 0.0551
Epoch 282/300
5/5 [==============================] - 2s 357ms/step - loss: 0.0666 - val_loss: 0.0588
Epoch 283/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0713 - val_loss: 0.0690
Epoch 284/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0691 - val_loss: 0.0515
Epoch 285/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0607 - val_loss: 0.0580
Epoch 286/300
5/5 [==============================] - 2s 358ms/step - loss: 0.0686 - val_loss: 0.0659
Epoch 287/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0821 - val_loss: 0.0516
Epoch 288/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0666 - val_loss: 0.0567
Epoch 289/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0671 - val_loss: 0.0500
Epoch 290/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0584 - val_loss: 0.0614
Epoch 291/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0620 - val_loss: 0.0604
Epoch 292/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0694 - val_loss: 0.0674
Epoch 293/300
5/5 [==============================] - 2s 352ms/step - loss: 0.0649 - val_loss: 0.0529
Epoch 294/300
5/5 [==============================] - 2s 352ms/step - loss: 0.0610 - val_loss: 0.0598
Epoch 295/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0609 - val_loss: 0.0601
Epoch 296/300
5/5 [==============================] - 2s 333ms/step - loss: 0.0538 - val_loss: 0.0647
Epoch 297/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0546 - val_loss: 0.0649
Epoch 298/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0518 - val_loss: 0.0686
Epoch 299/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0775 - val_loss: 0.0516
Epoch 300/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0681 - val_loss: 0.0715
Epoch 1/300
5/5 [==============================] - 35s 2s/step - loss: 0.4123 - val_loss: 0.4630
Epoch 2/300
5/5 [==============================] - 2s 425ms/step - loss: 0.1900 - val_loss: 0.1521
Epoch 3/300
5/5 [==============================] - 2s 496ms/step - loss: 0.1639 - val_loss: 0.2360
Epoch 4/300
5/5 [==============================] - 2s 475ms/step - loss: 0.1421 - val_loss: 0.2454
Epoch 5/300
5/5 [==============================] - 3s 525ms/step - loss: 0.1433 - val_loss: 0.2308
Epoch 6/300
5/5 [==============================] - 2s 444ms/step - loss: 0.1356 - val_loss: 0.2354
Epoch 7/300
5/5 [==============================] - 2s 355ms/step - loss: 0.1324 - val_loss: 0.2269
Epoch 8/300
5/5 [==============================] - 2s 427ms/step - loss: 0.1175 - val_loss: 0.0503
Epoch 9/300
5/5 [==============================] - 2s 431ms/step - loss: 0.1028 - val_loss: 0.0660
Epoch 10/300
5/5 [==============================] - 2s 463ms/step - loss: 0.0927 - val_loss: 0.0951
Epoch 11/300
5/5 [==============================] - 2s 445ms/step - loss: 0.0887 - val_loss: 0.0650
Epoch 12/300
5/5 [==============================] - 2s 407ms/step - loss: 0.0881 - val_loss: 0.0707
Epoch 13/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0852 - val_loss: 0.0538
Epoch 14/300
5/5 [==============================] - 2s 392ms/step - loss: 0.0838 - val_loss: 0.0695
Epoch 15/300
5/5 [==============================] - 2s 428ms/step - loss: 0.0841 - val_loss: 0.0519
Epoch 16/300
5/5 [==============================] - 2s 386ms/step - loss: 0.0846 - val_loss: 0.0516
Epoch 17/300
5/5 [==============================] - 2s 358ms/step - loss: 0.0863 - val_loss: 0.0541
Epoch 18/300
5/5 [==============================] - 2s 344ms/step - loss: 0.0869 - val_loss: 0.0497
Epoch 19/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0870 - val_loss: 0.0569
Epoch 20/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0847 - val_loss: 0.0493
Epoch 21/300
5/5 [==============================] - 2s 440ms/step - loss: 0.0833 - val_loss: 0.0548
Epoch 22/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0843 - val_loss: 0.0674
Epoch 23/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0864 - val_loss: 0.0547
Epoch 24/300
5/5 [==============================] - 2s 368ms/step - loss: 0.0862 - val_loss: 0.0503
Epoch 25/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0838 - val_loss: 0.0518
Epoch 26/300
5/5 [==============================] - 2s 346ms/step - loss: 0.0845 - val_loss: 0.0595
Epoch 27/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0869 - val_loss: 0.0500
Epoch 28/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0870 - val_loss: 0.0538
Epoch 29/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0876 - val_loss: 0.0714
Epoch 30/300
5/5 [==============================] - 2s 354ms/step - loss: 0.0905 - val_loss: 0.0490
Epoch 31/300
5/5 [==============================] - 2s 369ms/step - loss: 0.0883 - val_loss: 0.0523
Epoch 32/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0868 - val_loss: 0.0710
Epoch 33/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0888 - val_loss: 0.0501
Epoch 34/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0840 - val_loss: 0.0553
Epoch 35/300
5/5 [==============================] - 2s 511ms/step - loss: 0.0855 - val_loss: 0.0609
Epoch 36/300
5/5 [==============================] - 2s 396ms/step - loss: 0.0835 - val_loss: 0.0529
Epoch 37/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0836 - val_loss: 0.0543
Epoch 38/300
5/5 [==============================] - 2s 346ms/step - loss: 0.0845 - val_loss: 0.0506
Epoch 39/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0837 - val_loss: 0.0544
Epoch 40/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0844 - val_loss: 0.0539
Epoch 41/300
5/5 [==============================] - 2s 335ms/step - loss: 0.0842 - val_loss: 0.0518
Epoch 42/300
5/5 [==============================] - 2s 404ms/step - loss: 0.0839 - val_loss: 0.0516
Epoch 43/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0829 - val_loss: 0.0604
Epoch 44/300
5/5 [==============================] - 2s 397ms/step - loss: 0.0849 - val_loss: 0.0576
Epoch 45/300
5/5 [==============================] - 2s 360ms/step - loss: 0.0845 - val_loss: 0.0486
Epoch 46/300
5/5 [==============================] - 2s 410ms/step - loss: 0.0841 - val_loss: 0.0489
Epoch 47/300
5/5 [==============================] - 2s 463ms/step - loss: 0.0830 - val_loss: 0.0521
Epoch 48/300
5/5 [==============================] - 3s 593ms/step - loss: 0.0841 - val_loss: 0.0513
Epoch 49/300
5/5 [==============================] - 3s 551ms/step - loss: 0.0844 - val_loss: 0.0541
Epoch 50/300
5/5 [==============================] - 4s 617ms/step - loss: 0.0840 - val_loss: 0.0506
Epoch 51/300
5/5 [==============================] - 3s 548ms/step - loss: 0.0862 - val_loss: 0.0518
Epoch 52/300
5/5 [==============================] - 3s 504ms/step - loss: 0.0871 - val_loss: 0.0661
Epoch 53/300
5/5 [==============================] - 3s 506ms/step - loss: 0.0861 - val_loss: 0.0516
Epoch 54/300
5/5 [==============================] - 3s 561ms/step - loss: 0.0863 - val_loss: 0.0548
Epoch 55/300
5/5 [==============================] - 2s 503ms/step - loss: 0.0851 - val_loss: 0.0566
Epoch 56/300
5/5 [==============================] - 2s 427ms/step - loss: 0.0854 - val_loss: 0.0571
Epoch 57/300
5/5 [==============================] - 2s 497ms/step - loss: 0.0833 - val_loss: 0.0530
Epoch 58/300
5/5 [==============================] - 2s 420ms/step - loss: 0.0844 - val_loss: 0.0634
Epoch 59/300
5/5 [==============================] - 2s 476ms/step - loss: 0.0832 - val_loss: 0.0562
Epoch 60/300
5/5 [==============================] - 2s 463ms/step - loss: 0.0865 - val_loss: 0.0552
Epoch 61/300
5/5 [==============================] - 2s 425ms/step - loss: 0.0849 - val_loss: 0.0595
Epoch 62/300
5/5 [==============================] - 2s 437ms/step - loss: 0.0835 - val_loss: 0.0611
Epoch 63/300
5/5 [==============================] - 2s 476ms/step - loss: 0.0828 - val_loss: 0.0566
Epoch 64/300
5/5 [==============================] - 2s 474ms/step - loss: 0.0852 - val_loss: 0.0539
Epoch 65/300
5/5 [==============================] - 2s 410ms/step - loss: 0.0855 - val_loss: 0.0521
Epoch 66/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0838 - val_loss: 0.0626
Epoch 67/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0827 - val_loss: 0.0594
Epoch 68/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0833 - val_loss: 0.0542
Epoch 69/300
5/5 [==============================] - 2s 374ms/step - loss: 0.0815 - val_loss: 0.0498
Epoch 70/300
5/5 [==============================] - 2s 300ms/step - loss: 0.0809 - val_loss: 0.0563
Epoch 71/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0852 - val_loss: 0.0537
Epoch 72/300
5/5 [==============================] - 2s 328ms/step - loss: 0.0860 - val_loss: 0.0480
Epoch 73/300
5/5 [==============================] - 2s 325ms/step - loss: 0.0861 - val_loss: 0.0473
Epoch 74/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0867 - val_loss: 0.0544
Epoch 75/300
5/5 [==============================] - 2s 314ms/step - loss: 0.0881 - val_loss: 0.0503
Epoch 76/300
5/5 [==============================] - 1s 306ms/step - loss: 0.0887 - val_loss: 0.0521
Epoch 77/300
5/5 [==============================] - 2s 318ms/step - loss: 0.0845 - val_loss: 0.0546
Epoch 78/300
5/5 [==============================] - 2s 310ms/step - loss: 0.0856 - val_loss: 0.0556
Epoch 79/300
5/5 [==============================] - 2s 305ms/step - loss: 0.0871 - val_loss: 0.0623
Epoch 80/300
5/5 [==============================] - 1s 292ms/step - loss: 0.0851 - val_loss: 0.0545
Epoch 81/300
5/5 [==============================] - 2s 315ms/step - loss: 0.0843 - val_loss: 0.0575
Epoch 82/300
5/5 [==============================] - 2s 311ms/step - loss: 0.0825 - val_loss: 0.0645
Epoch 83/300
5/5 [==============================] - 2s 329ms/step - loss: 0.0833 - val_loss: 0.0553
Epoch 84/300
5/5 [==============================] - 1s 294ms/step - loss: 0.0878 - val_loss: 0.0622
Epoch 85/300
5/5 [==============================] - 1s 304ms/step - loss: 0.0886 - val_loss: 0.0578
Epoch 86/300
5/5 [==============================] - 2s 314ms/step - loss: 0.0868 - val_loss: 0.0507
Epoch 87/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0894 - val_loss: 0.0490
Epoch 88/300
5/5 [==============================] - 1s 296ms/step - loss: 0.0873 - val_loss: 0.0566
Epoch 89/300
5/5 [==============================] - 1s 299ms/step - loss: 0.0854 - val_loss: 0.0582
Epoch 90/300
5/5 [==============================] - 2s 311ms/step - loss: 0.0825 - val_loss: 0.0589
Epoch 91/300
5/5 [==============================] - 2s 322ms/step - loss: 0.0834 - val_loss: 0.0499
Epoch 92/300
5/5 [==============================] - 2s 312ms/step - loss: 0.0826 - val_loss: 0.0532
Epoch 93/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0815 - val_loss: 0.0530
Epoch 94/300
5/5 [==============================] - 2s 312ms/step - loss: 0.0811 - val_loss: 0.0570
Epoch 95/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0816 - val_loss: 0.0601
Epoch 96/300
5/5 [==============================] - 2s 402ms/step - loss: 0.0831 - val_loss: 0.0509
Epoch 97/300
5/5 [==============================] - 2s 418ms/step - loss: 0.0848 - val_loss: 0.0522
Epoch 98/300
5/5 [==============================] - 2s 371ms/step - loss: 0.0845 - val_loss: 0.0574
Epoch 99/300
5/5 [==============================] - 2s 356ms/step - loss: 0.0844 - val_loss: 0.0522
Epoch 100/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0821 - val_loss: 0.0517
Epoch 101/300
5/5 [==============================] - 2s 389ms/step - loss: 0.0806 - val_loss: 0.0565
Epoch 102/300
5/5 [==============================] - 2s 329ms/step - loss: 0.0820 - val_loss: 0.0606
Epoch 103/300
5/5 [==============================] - 1s 304ms/step - loss: 0.0793 - val_loss: 0.0493
Epoch 104/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0773 - val_loss: 0.0463
Epoch 105/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0752 - val_loss: 0.0495
Epoch 106/300
5/5 [==============================] - 2s 329ms/step - loss: 0.0777 - val_loss: 0.0467
Epoch 107/300
5/5 [==============================] - 2s 508ms/step - loss: 0.0847 - val_loss: 0.0537
Epoch 108/300
5/5 [==============================] - 2s 459ms/step - loss: 0.0805 - val_loss: 0.0588
Epoch 109/300
5/5 [==============================] - 2s 398ms/step - loss: 0.0849 - val_loss: 0.0499
Epoch 110/300
5/5 [==============================] - 2s 400ms/step - loss: 0.0832 - val_loss: 0.0479
Epoch 111/300
5/5 [==============================] - 2s 384ms/step - loss: 0.0805 - val_loss: 0.0609
Epoch 112/300
5/5 [==============================] - 2s 404ms/step - loss: 0.0777 - val_loss: 0.0503
Epoch 113/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0789 - val_loss: 0.0498
Epoch 114/300
5/5 [==============================] - 2s 313ms/step - loss: 0.0789 - val_loss: 0.0510
Epoch 115/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0748 - val_loss: 0.0512
Epoch 116/300
5/5 [==============================] - 1s 295ms/step - loss: 0.0737 - val_loss: 0.0483
Epoch 117/300
5/5 [==============================] - 1s 294ms/step - loss: 0.0774 - val_loss: 0.0477
Epoch 118/300
5/5 [==============================] - 1s 300ms/step - loss: 0.0759 - val_loss: 0.0520
Epoch 119/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0798 - val_loss: 0.0497
Epoch 120/300
5/5 [==============================] - 2s 380ms/step - loss: 0.0789 - val_loss: 0.0843
Epoch 121/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0775 - val_loss: 0.0528
Epoch 122/300
5/5 [==============================] - 1s 304ms/step - loss: 0.0770 - val_loss: 0.0488
Epoch 123/300
5/5 [==============================] - 2s 307ms/step - loss: 0.0797 - val_loss: 0.0508
Epoch 124/300
5/5 [==============================] - 2s 303ms/step - loss: 0.0754 - val_loss: 0.0511
Epoch 125/300
5/5 [==============================] - 2s 397ms/step - loss: 0.0754 - val_loss: 0.0475
Epoch 126/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0780 - val_loss: 0.0458
Epoch 127/300
5/5 [==============================] - 2s 337ms/step - loss: 0.0773 - val_loss: 0.0461
Epoch 128/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0758 - val_loss: 0.0466
Epoch 129/300
5/5 [==============================] - 2s 312ms/step - loss: 0.0787 - val_loss: 0.0524
Epoch 130/300
5/5 [==============================] - 1s 303ms/step - loss: 0.0776 - val_loss: 0.0469
Epoch 131/300
5/5 [==============================] - 1s 298ms/step - loss: 0.0764 - val_loss: 0.0690
Epoch 132/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0803 - val_loss: 0.0484
Epoch 133/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0721 - val_loss: 0.0473
Epoch 134/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0710 - val_loss: 0.0463
Epoch 135/300
5/5 [==============================] - 2s 321ms/step - loss: 0.0712 - val_loss: 0.0478
Epoch 136/300
5/5 [==============================] - 2s 307ms/step - loss: 0.0692 - val_loss: 0.0471
Epoch 137/300
5/5 [==============================] - 2s 313ms/step - loss: 0.0686 - val_loss: 0.0467
Epoch 138/300
5/5 [==============================] - 2s 348ms/step - loss: 0.0673 - val_loss: 0.0466
Epoch 139/300
5/5 [==============================] - 2s 306ms/step - loss: 0.0666 - val_loss: 0.0543
Epoch 140/300
5/5 [==============================] - 1s 303ms/step - loss: 0.0699 - val_loss: 0.0525
Epoch 141/300
5/5 [==============================] - 1s 298ms/step - loss: 0.0726 - val_loss: 0.0582
Epoch 142/300
5/5 [==============================] - 1s 300ms/step - loss: 0.0678 - val_loss: 0.0479
Epoch 143/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0729 - val_loss: 0.0450
Epoch 144/300
5/5 [==============================] - 1s 297ms/step - loss: 0.0832 - val_loss: 0.0638
Epoch 145/300
5/5 [==============================] - 1s 305ms/step - loss: 0.0820 - val_loss: 0.0580
Epoch 146/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0739 - val_loss: 0.0507
Epoch 147/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0755 - val_loss: 0.0750
Epoch 148/300
5/5 [==============================] - 1s 305ms/step - loss: 0.0721 - val_loss: 0.0723
Epoch 149/300
5/5 [==============================] - 2s 314ms/step - loss: 0.0733 - val_loss: 0.0460
Epoch 150/300
5/5 [==============================] - 1s 306ms/step - loss: 0.0698 - val_loss: 0.0564
Epoch 151/300
5/5 [==============================] - 1s 300ms/step - loss: 0.0691 - val_loss: 0.0498
Epoch 152/300
5/5 [==============================] - 2s 302ms/step - loss: 0.0690 - val_loss: 0.0478
Epoch 153/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0694 - val_loss: 0.0676
Epoch 154/300
5/5 [==============================] - 2s 306ms/step - loss: 0.0742 - val_loss: 0.0572
Epoch 155/300
5/5 [==============================] - 2s 316ms/step - loss: 0.0718 - val_loss: 0.0546
Epoch 156/300
5/5 [==============================] - 1s 298ms/step - loss: 0.0691 - val_loss: 0.0451
Epoch 157/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0680 - val_loss: 0.0464
Epoch 158/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0643 - val_loss: 0.0569
Epoch 159/300
5/5 [==============================] - 1s 299ms/step - loss: 0.0648 - val_loss: 0.0653
Epoch 160/300
5/5 [==============================] - 2s 314ms/step - loss: 0.0623 - val_loss: 0.0569
Epoch 161/300
5/5 [==============================] - 1s 295ms/step - loss: 0.0643 - val_loss: 0.0437
Epoch 162/300
5/5 [==============================] - 1s 298ms/step - loss: 0.0633 - val_loss: 0.0481
Epoch 163/300
5/5 [==============================] - 1s 297ms/step - loss: 0.0625 - val_loss: 0.0711
Epoch 164/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0641 - val_loss: 0.0535
Epoch 165/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0651 - val_loss: 0.0503
Epoch 166/300
5/5 [==============================] - 2s 306ms/step - loss: 0.0621 - val_loss: 0.0599
Epoch 167/300
5/5 [==============================] - 2s 321ms/step - loss: 0.0571 - val_loss: 0.0633
Epoch 168/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0566 - val_loss: 0.0401
Epoch 169/300
5/5 [==============================] - 2s 335ms/step - loss: 0.0580 - val_loss: 0.0432
Epoch 170/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0584 - val_loss: 0.0521
Epoch 171/300
5/5 [==============================] - 2s 368ms/step - loss: 0.0575 - val_loss: 0.0683
Epoch 172/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0575 - val_loss: 0.0480
Epoch 173/300
5/5 [==============================] - 2s 307ms/step - loss: 0.0523 - val_loss: 0.0425
Epoch 174/300
5/5 [==============================] - 2s 352ms/step - loss: 0.0508 - val_loss: 0.0486
Epoch 175/300
5/5 [==============================] - 2s 338ms/step - loss: 0.0482 - val_loss: 0.0460
Epoch 176/300
5/5 [==============================] - 2s 407ms/step - loss: 0.0557 - val_loss: 0.0439
Epoch 177/300
5/5 [==============================] - 2s 410ms/step - loss: 0.0562 - val_loss: 0.0609
Epoch 178/300
5/5 [==============================] - 2s 339ms/step - loss: 0.0637 - val_loss: 0.0506
Epoch 179/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0565 - val_loss: 0.0511
Epoch 180/300
5/5 [==============================] - 1s 295ms/step - loss: 0.0520 - val_loss: 0.0511
Epoch 181/300
5/5 [==============================] - 1s 296ms/step - loss: 0.0464 - val_loss: 0.0490
Epoch 182/300
5/5 [==============================] - 2s 319ms/step - loss: 0.0462 - val_loss: 0.0885
Epoch 183/300
5/5 [==============================] - 1s 303ms/step - loss: 0.0455 - val_loss: 0.0619
Epoch 184/300
5/5 [==============================] - 2s 306ms/step - loss: 0.0424 - val_loss: 0.0441
Epoch 185/300
5/5 [==============================] - 1s 298ms/step - loss: 0.0431 - val_loss: 0.0395
Epoch 186/300
5/5 [==============================] - 2s 392ms/step - loss: 0.0418 - val_loss: 0.0444
Epoch 187/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0416 - val_loss: 0.0558
Epoch 188/300
5/5 [==============================] - 2s 309ms/step - loss: 0.0409 - val_loss: 0.0415
Epoch 189/300
5/5 [==============================] - 2s 315ms/step - loss: 0.0414 - val_loss: 0.0537
Epoch 190/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0443 - val_loss: 0.0578
Epoch 191/300
5/5 [==============================] - 2s 332ms/step - loss: 0.0413 - val_loss: 0.0520
Epoch 192/300
5/5 [==============================] - 2s 362ms/step - loss: 0.0400 - val_loss: 0.0456
Epoch 193/300
5/5 [==============================] - 2s 385ms/step - loss: 0.0459 - val_loss: 0.0440
Epoch 194/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0437 - val_loss: 0.0562
Epoch 195/300
5/5 [==============================] - 1s 304ms/step - loss: 0.0426 - val_loss: 0.0436
Epoch 196/300
5/5 [==============================] - 1s 296ms/step - loss: 0.0399 - val_loss: 0.0467
Epoch 197/300
5/5 [==============================] - 1s 296ms/step - loss: 0.0405 - val_loss: 0.0586
Epoch 198/300
5/5 [==============================] - 2s 377ms/step - loss: 0.0404 - val_loss: 0.0567
Epoch 199/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0375 - val_loss: 0.0559
Epoch 200/300
5/5 [==============================] - 2s 356ms/step - loss: 0.0341 - val_loss: 0.0390
Epoch 201/300
5/5 [==============================] - 2s 367ms/step - loss: 0.0333 - val_loss: 0.0463
Epoch 202/300
5/5 [==============================] - 2s 310ms/step - loss: 0.0368 - val_loss: 0.0382
Epoch 203/300
5/5 [==============================] - 2s 307ms/step - loss: 0.0358 - val_loss: 0.0519
Epoch 204/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0349 - val_loss: 0.0325
Epoch 205/300
5/5 [==============================] - 2s 319ms/step - loss: 0.0429 - val_loss: 0.0314
Epoch 206/300
5/5 [==============================] - 1s 302ms/step - loss: 0.0444 - val_loss: 0.1003
Epoch 207/300
5/5 [==============================] - 2s 324ms/step - loss: 0.0460 - val_loss: 0.0481
Epoch 208/300
5/5 [==============================] - 1s 301ms/step - loss: 0.0408 - val_loss: 0.0436
Epoch 209/300
5/5 [==============================] - 1s 300ms/step - loss: 0.0467 - val_loss: 0.0651
Epoch 210/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0423 - val_loss: 0.0350
Epoch 211/300
5/5 [==============================] - 2s 349ms/step - loss: 0.0429 - val_loss: 0.0402
Epoch 212/300
5/5 [==============================] - 2s 322ms/step - loss: 0.0373 - val_loss: 0.0430
Epoch 213/300
5/5 [==============================] - 2s 331ms/step - loss: 0.0404 - val_loss: 0.0449
Epoch 214/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0381 - val_loss: 0.0396
Epoch 215/300
5/5 [==============================] - 2s 308ms/step - loss: 0.0363 - val_loss: 0.0482
Epoch 216/300
5/5 [==============================] - 2s 320ms/step - loss: 0.0303 - val_loss: 0.0427
Epoch 217/300
5/5 [==============================] - 2s 350ms/step - loss: 0.0325 - val_loss: 0.0431
Epoch 218/300
5/5 [==============================] - 2s 313ms/step - loss: 0.0305 - val_loss: 0.0357
Epoch 219/300
5/5 [==============================] - 2s 383ms/step - loss: 0.0329 - val_loss: 0.0364
Epoch 220/300
5/5 [==============================] - 2s 320ms/step - loss: 0.0327 - val_loss: 0.0395
Epoch 221/300
5/5 [==============================] - 2s 302ms/step - loss: 0.0272 - val_loss: 0.0433
Epoch 222/300
5/5 [==============================] - 2s 372ms/step - loss: 0.0313 - val_loss: 0.0433
Epoch 223/300
5/5 [==============================] - 2s 324ms/step - loss: 0.0298 - val_loss: 0.0334
Epoch 224/300
5/5 [==============================] - 2s 374ms/step - loss: 0.0301 - val_loss: 0.0540
Epoch 225/300
5/5 [==============================] - 2s 393ms/step - loss: 0.0500 - val_loss: 0.0547
Epoch 226/300
5/5 [==============================] - 2s 418ms/step - loss: 0.0471 - val_loss: 0.0436
Epoch 227/300
5/5 [==============================] - 2s 361ms/step - loss: 0.0480 - val_loss: 0.0606
Epoch 228/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0506 - val_loss: 0.0418
Epoch 229/300
5/5 [==============================] - 2s 386ms/step - loss: 0.0423 - val_loss: 0.0470
Epoch 230/300
5/5 [==============================] - 2s 402ms/step - loss: 0.0393 - val_loss: 0.0414
Epoch 231/300
5/5 [==============================] - 2s 448ms/step - loss: 0.0383 - val_loss: 0.0343
Epoch 232/300
5/5 [==============================] - 2s 424ms/step - loss: 0.0378 - val_loss: 0.0385
Epoch 233/300
5/5 [==============================] - 2s 437ms/step - loss: 0.0386 - val_loss: 0.0345
Epoch 234/300
5/5 [==============================] - 2s 471ms/step - loss: 0.0368 - val_loss: 0.0596
Epoch 235/300
5/5 [==============================] - 2s 456ms/step - loss: 0.0365 - val_loss: 0.0487
Epoch 236/300
5/5 [==============================] - 2s 409ms/step - loss: 0.0392 - val_loss: 0.0328
Epoch 237/300
5/5 [==============================] - 2s 370ms/step - loss: 0.0354 - val_loss: 0.0508
Epoch 238/300
5/5 [==============================] - 2s 415ms/step - loss: 0.0365 - val_loss: 0.0420
Epoch 239/300
5/5 [==============================] - 2s 391ms/step - loss: 0.0417 - val_loss: 0.0347
Epoch 240/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0390 - val_loss: 0.0402
Epoch 241/300
5/5 [==============================] - 2s 335ms/step - loss: 0.0333 - val_loss: 0.0401
Epoch 242/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0333 - val_loss: 0.0497
Epoch 243/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0309 - val_loss: 0.0602
Epoch 244/300
5/5 [==============================] - 2s 394ms/step - loss: 0.0300 - val_loss: 0.0353
Epoch 245/300
5/5 [==============================] - 2s 388ms/step - loss: 0.0280 - val_loss: 0.0314
Epoch 246/300
5/5 [==============================] - 2s 389ms/step - loss: 0.0273 - val_loss: 0.0362
Epoch 247/300
5/5 [==============================] - 2s 398ms/step - loss: 0.0290 - val_loss: 0.0412
Epoch 248/300
5/5 [==============================] - 2s 372ms/step - loss: 0.0282 - val_loss: 0.0349
Epoch 249/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0293 - val_loss: 0.0290
Epoch 250/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0246 - val_loss: 0.0326
Epoch 251/300
5/5 [==============================] - 2s 353ms/step - loss: 0.0240 - val_loss: 0.0434
Epoch 252/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0241 - val_loss: 0.0291
Epoch 253/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0248 - val_loss: 0.0499
Epoch 254/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0250 - val_loss: 0.0325
Epoch 255/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0245 - val_loss: 0.0305
Epoch 256/300
5/5 [==============================] - 2s 376ms/step - loss: 0.0256 - val_loss: 0.0312
Epoch 257/300
5/5 [==============================] - 2s 347ms/step - loss: 0.0306 - val_loss: 0.0430
Epoch 258/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0299 - val_loss: 0.0417
Epoch 259/300
5/5 [==============================] - 2s 403ms/step - loss: 0.0280 - val_loss: 0.0361
Epoch 260/300
5/5 [==============================] - 2s 355ms/step - loss: 0.0315 - val_loss: 0.0446
Epoch 261/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0297 - val_loss: 0.0315
Epoch 262/300
5/5 [==============================] - 2s 459ms/step - loss: 0.0274 - val_loss: 0.0386
Epoch 263/300
5/5 [==============================] - 2s 413ms/step - loss: 0.0248 - val_loss: 0.0351
Epoch 264/300
5/5 [==============================] - 2s 431ms/step - loss: 0.0254 - val_loss: 0.0398
Epoch 265/300
5/5 [==============================] - 2s 451ms/step - loss: 0.0239 - val_loss: 0.0368
Epoch 266/300
5/5 [==============================] - 2s 414ms/step - loss: 0.0241 - val_loss: 0.0313
Epoch 267/300
5/5 [==============================] - 2s 444ms/step - loss: 0.0230 - val_loss: 0.0280
Epoch 268/300
5/5 [==============================] - 2s 465ms/step - loss: 0.0241 - val_loss: 0.0295
Epoch 269/300
5/5 [==============================] - 2s 426ms/step - loss: 0.0249 - val_loss: 0.0350
Epoch 270/300
5/5 [==============================] - 2s 434ms/step - loss: 0.0260 - val_loss: 0.0422
Epoch 271/300
5/5 [==============================] - 2s 425ms/step - loss: 0.0247 - val_loss: 0.0324
Epoch 272/300
5/5 [==============================] - 2s 455ms/step - loss: 0.0240 - val_loss: 0.0350
Epoch 273/300
5/5 [==============================] - 2s 452ms/step - loss: 0.0221 - val_loss: 0.0475
Epoch 274/300
5/5 [==============================] - 2s 454ms/step - loss: 0.0210 - val_loss: 0.0292
Epoch 275/300
5/5 [==============================] - 2s 441ms/step - loss: 0.0259 - val_loss: 0.0326
Epoch 276/300
5/5 [==============================] - 2s 444ms/step - loss: 0.0223 - val_loss: 0.0286
Epoch 277/300
5/5 [==============================] - 2s 394ms/step - loss: 0.0218 - val_loss: 0.0317
Epoch 278/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0226 - val_loss: 0.0368
Epoch 279/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0229 - val_loss: 0.0268
Epoch 280/300
5/5 [==============================] - 2s 346ms/step - loss: 0.0231 - val_loss: 0.0368
Epoch 281/300
5/5 [==============================] - 2s 340ms/step - loss: 0.0245 - val_loss: 0.0307
Epoch 282/300
5/5 [==============================] - 2s 359ms/step - loss: 0.0257 - val_loss: 0.0297
Epoch 283/300
5/5 [==============================] - 2s 343ms/step - loss: 0.0282 - val_loss: 0.0345
Epoch 284/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0263 - val_loss: 0.0375
Epoch 285/300
5/5 [==============================] - 2s 341ms/step - loss: 0.0259 - val_loss: 0.0283
Epoch 286/300
5/5 [==============================] - 2s 363ms/step - loss: 0.0244 - val_loss: 0.0375
Epoch 287/300
5/5 [==============================] - 2s 365ms/step - loss: 0.0217 - val_loss: 0.0302
Epoch 288/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0196 - val_loss: 0.0303
Epoch 289/300
5/5 [==============================] - 2s 406ms/step - loss: 0.0219 - val_loss: 0.0367
Epoch 290/300
5/5 [==============================] - 2s 382ms/step - loss: 0.0267 - val_loss: 0.0308
Epoch 291/300
5/5 [==============================] - 2s 373ms/step - loss: 0.0261 - val_loss: 0.0334
Epoch 292/300
5/5 [==============================] - 2s 383ms/step - loss: 0.0312 - val_loss: 0.0319
Epoch 293/300
5/5 [==============================] - 2s 379ms/step - loss: 0.0295 - val_loss: 0.0366
Epoch 294/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0331 - val_loss: 0.0342
Epoch 295/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0312 - val_loss: 0.0432
Epoch 296/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0297 - val_loss: 0.0449
Epoch 297/300
5/5 [==============================] - 2s 342ms/step - loss: 0.0289 - val_loss: 0.0356
Epoch 298/300
5/5 [==============================] - 2s 345ms/step - loss: 0.0254 - val_loss: 0.0326
Epoch 299/300
5/5 [==============================] - 2s 351ms/step - loss: 0.0228 - val_loss: 0.0327
Epoch 300/300
5/5 [==============================] - 2s 336ms/step - loss: 0.0192 - val_loss: 0.0332
../../_images/Forecaster_examples_LSTM_28_1.png
../../_images/Forecaster_examples_LSTM_28_2.png

MLR Modeling

[20]:
f.set_estimator('mlr')
f.add_ar_terms(24)
f.add_seasonal_regressors('month','quarter',dummy=True)
f.add_seasonal_regressors('year')
f.add_time_trend()
f.diff()
[28]:
f.manual_forecast()
f.plot_test_set(order_by='LevelTestSetMAPE',models='top_3')
../../_images/Forecaster_examples_LSTM_31_0.png
[22]:
f.plot_test_set(models='mlr',ci=True)
../../_images/Forecaster_examples_LSTM_32_0.png
[29]:
f.plot(order_by='LevelTestSetMAPE',models='top_3')
../../_images/Forecaster_examples_LSTM_33_0.png

Benchmarking

[24]:
f.export('model_summaries',determine_best_by='LevelTestSetMAPE')[
    ['ModelNickname','LevelTestSetMAPE','LevelTestSetRMSE','LevelTestSetR2','best_model']
]
[24]:
ModelNickname LevelTestSetMAPE LevelTestSetRMSE LevelTestSetR2 best_model
0 mlr 0.023420 13.932124 0.964960 True
1 lstm_24lags_4layers_15epochs 0.114650 81.210940 -0.190586 False
2 lstm_best 0.154356 93.090478 -0.564380 False
3 lstm_24lags_earlystop_3layers 0.375480 198.842435 -6.137560 False
4 lstm_24lags 0.712334 362.077454 -22.666530 False
5 lstm_24lags_5epochs 0.719090 347.686110 -20.822590 False
6 lstm_default 0.779915 379.298924 -24.971368 False

Export Results

[25]:
f.export_forecasts_with_cis('mlr')
[25]:
DATE UpperForecast Forecast LowerForecast ModelNickname CILevel
0 1961-01-01 47.711290 30.246319 12.781347 mlr 0.95
1 1961-02-01 -27.057760 -44.522731 -61.987702 mlr 0.95
2 1961-03-01 61.142910 43.677939 26.212968 mlr 0.95
3 1961-04-01 46.528396 29.063425 11.598454 mlr 0.95
4 1961-05-01 29.862361 12.397390 -5.067581 mlr 0.95
5 1961-06-01 90.724956 73.259985 55.795014 mlr 0.95
6 1961-07-01 123.005797 105.540825 88.075854 mlr 0.95
7 1961-08-01 -16.122173 -33.587144 -51.052115 mlr 0.95
8 1961-09-01 -73.067354 -90.532325 -107.997296 mlr 0.95
9 1961-10-01 -41.907826 -59.372797 -76.837768 mlr 0.95
10 1961-11-01 -54.975544 -72.440515 -89.905487 mlr 0.95
11 1961-12-01 66.035665 48.570694 31.105723 mlr 0.95
[26]:
f.export_test_set_preds_with_cis('mlr')
[26]:
DATE UpperPreds Preds Actuals LowerPreds ModelNickname CILevel
0 1960-01-01 28.066682 10.601711 12.0 -6.863261 mlr 0.95
1 1960-02-01 0.741874 -16.723097 -26.0 -34.188068 mlr 0.95
2 1960-03-01 65.886942 48.421971 28.0 30.956999 mlr 0.95
3 1960-04-01 2.898761 -14.566210 42.0 -32.031182 mlr 0.95
4 1960-05-01 50.348916 32.883945 11.0 15.418974 mlr 0.95
5 1960-06-01 78.363001 60.898029 63.0 43.433058 mlr 0.95
6 1960-07-01 103.669621 86.204650 87.0 68.739679 mlr 0.95
7 1960-08-01 23.470167 6.005196 -16.0 -11.459775 mlr 0.95
8 1960-09-01 -86.780501 -104.245472 -98.0 -121.710443 mlr 0.95
9 1960-10-01 -51.156100 -68.621071 -47.0 -86.086042 mlr 0.95
10 1960-11-01 -43.972634 -61.437606 -71.0 -78.902577 mlr 0.95
11 1960-12-01 66.544516 49.079545 42.0 31.614574 mlr 0.95

Export Feature Info

[27]:
f.save_feature_importance()
f.export_feature_importance('mlr')
[27]:
weight std
feature
AR1 0.360410 0.057652
AR12 0.218659 0.026116
AR4 0.214434 0.031875
AR10 0.184816 0.017657
AR21 0.131055 0.009248
AR13 0.116967 0.011384
AR18 0.111233 0.012176
AR2 0.088423 0.019647
AR11 0.085070 0.011649
AR20 0.080938 0.011092
AR3 0.071639 0.014407
AR22 0.065199 0.003837
AR7 0.063136 0.003587
AR8 0.057833 0.006383
month_9 0.044074 0.003543
AR9 0.043885 0.002790
AR14 0.034640 0.007735
AR17 0.033160 0.002823
AR24 0.031625 0.003686
AR23 0.030453 0.002730
AR19 0.020941 0.001888
month_7 0.013514 0.002930
AR5 0.008518 0.000783
month_8 0.007976 0.002222
month_2 0.007844 0.002097
AR16 0.006507 0.002998
month_10 0.006464 0.001107
month_11 0.003286 0.000805
AR6 0.003027 0.000783
month_3 0.002440 0.001107
year 0.002268 0.000767
month 0.001389 0.000671
month_12 0.001126 0.000959
quarter_2 0.000940 0.000174
month_1 0.000728 0.001333
AR15 0.000653 0.000701
t 0.000439 0.000580
month_6 0.000257 0.000201
month_5 0.000195 0.000508
month_4 0.000147 0.000201
quarter_1 0.000085 0.000297
quarter_4 0.000043 0.000198
quarter 0.000022 0.000066
quarter_3 0.000012 0.000040
[ ]: