This example illustrates how simple it is to train a classifier using side information.
It illustrates the exemplary use of the multi-view pattern; for more info on how to use other patterns, check out the other examples.
In [18]:
from __future__ import print_function
import concarne
import concarne.patterns
import concarne.training
import lasagne
import theano.tensor as T
%pylab inline
try:
import sklearn.linear_model as sklm
except:
print (
"""You don't have scikit-learn installed; install it to compare
learning with side information to simple supervised learning""")
sklm = None
import numpy as np
Populating the interactive namespace from numpy and matplotlib
WARNING: pylab import has clobbered these variables: ['beta', 'split']
`%matplotlib` prevents importing * from pylab and numpy
In [19]:
num_samples = 300
input_dim = 50
side_dim = 50
In [20]:
# generate some random data with 100 samples
# and 5 dimensions
X = np.random.randn(num_samples, input_dim)
# select the third dimension as the relevant
# for our classification task
S = X[:, 2:3]
# The labels are simply the sign of S
# (note the downcast to int32 - this is required
# by theano)
y = np.asarray(S > 0, dtype='int32').reshape( (-1,) )
# This means we have 2 classes - we will use
# that later for building the pattern
num_classes = 2
In [21]:
plt.plot(S)
plt.plot(y)
Out[21]:
[<matplotlib.lines.Line2D at 0x7f0c7101c510>]
Now let's define some side information: we simulate an additional sensorwhich contains S, but embedded into a different space.
In [22]:
Z = np.random.randn(num_samples, side_dim)
# set second dimension of Z to correspond to S
Z[:, 1] = S[:,0]
Let's make it harder to find S in X and Z by applying a random rotations to both data sets
In [23]:
# random rotation 1
R = np.linalg.qr(np.random.randn(input_dim, input_dim))[0]
X = X.dot(R)
# random rotation 2
Q = np.linalg.qr(np.random.randn(side_dim, side_dim))[0]
Z = Z.dot(Q)
Finally, split our data into training, test, and validation data
In [24]:
split = num_samples/3
X_train = X[:split]
X_val = X[split:2*split]
X_test = X[2*split:]
y_train = y[:split]
y_val = y[split:2*split]
y_test = y[2*split:]
Z_train = Z[:split]
Z_val = Z[split:2*split]
Z_test = Z[2*split:]
Let's check how hard the problem is for supervised learning alone.
In [25]:
if sklm is not None:
# let's try different regularizations
for c in [1e-5, 1e-1, 1, 10, 100, 1e5]:
lr = sklm.LogisticRegression(C=c)
lr.fit(X_train, y_train)
print ("Logistic Regression (C=%f)\n accuracy = %.3f %%" % (c, 100*lr.score(X_test, y_test)))
Logistic Regression (C=0.000010)
accuracy = 78.000 %
Logistic Regression (C=0.100000)
accuracy = 81.000 %
Logistic Regression (C=1.000000)
accuracy = 78.000 %
Logistic Regression (C=10.000000)
accuracy = 79.000 %
Logistic Regression (C=100.000000)
accuracy = 80.000 %
Logistic Regression (C=100000.000000)
accuracy = 76.000 %
In [26]:
# Let's first define the theano variables which will represent our data
input_var = T.matrix('inputs') # for X
target_var = T.ivector('targets') # for Y
side_var = T.matrix('sideinfo') # for Z
In [27]:
# Size of the intermediate representation phi(X);
# since S is 1-dim, phi(X) can also map to a
# 1-dim vector
representation_dim = 1
Now define the functions - we choose linear functions.
concarne internally relies on lasagne which encodes functions as (sets of) layers. Additionally, concarne supports nolearn style initialization of lasagne layers as follows:
In [28]:
phi = [ (lasagne.layers.DenseLayer,
{ 'num_units': concarne.patterns.Pattern.PHI_OUTPUT_SHAPE,
'nonlinearity':None, 'b':None })]
psi = [(lasagne.layers.DenseLayer,
{ 'num_units': concarne.patterns.Pattern.PSI_OUTPUT_SHAPE,
'nonlinearity':lasagne.nonlinearities.softmax, 'b':None })]
beta = [(lasagne.layers.DenseLayer,
{ 'num_units': concarne.patterns.Pattern.BETA_OUTPUT_SHAPE,
'nonlinearity':None, 'b':None })]
For the variable of your layer that denotes the output of the network you should use the markers PHI_OUTPUT_SHAPE, PSI_OUTPUT_SHAPE and BETA_OUTPUT_SHAPE, so that the pattern can automatically infer the correct shape.
In [29]:
pattern = concarne.patterns.MultiViewPattern(
phi=phi, psi=psi, beta=beta,
# the following parameters are required to
# build the functions and the losses
input_var=input_var,
target_var=target_var,
side_var=side_var,
input_shape=input_dim,
target_shape=num_classes,
side_shape=side_dim,
representation_shape=representation_dim,
# we have to define two loss functions:
# 1) the target loss deals with
# optimizing psi and phi wrt. X & Y
target_loss=lasagne.objectives.categorical_crossentropy,
# 2) the side loss deals with
# optimizing beta and phi wrt. X & Z,
# for multi-view it is beta(Z)~phi(X)
side_loss=lasagne.objectives.squared_error)
To train a pattern, you can use the PatternTrainer which trains the pattern via stochastic gradient descent. It also supports different procedures to train the pattern.
In [30]:
trainer = concarne.training.PatternTrainer(
pattern,
procedure='simultaneous',
num_epochs=500,
batch_size=10,
update=lasagne.updates.nesterov_momentum,
update_learning_rate=0.01,
update_momentum=0.9,
)
Let's train!
In [31]:
trainer.fit_XYZ(X_train, y_train, [Z_train],
X_val=X_val, y_val=y_val,
side_val=[X_val, Z_val],
verbose=True)
pass
Training procedure: simultaneous
Optimize phi & psi & beta using a weighted sum of target and side objective
Update: nesterov_momentum(learning_rate=0.01, momentum=0.9)
-> standard mode with single training function
Starting training...
Epoch 1 of 500 took 0.004s
training loss: 1.495710
(target: 0.355893, side: 1.139818 (absolute))
validation loss: 0.703471
validation accuracy: 49.00 %
side validation loss: 0.491789
side validation accuracy: nan %
Epoch 2 of 500 took 0.005s
training loss: 0.544633
(target: 0.346933, side: 0.197700 (absolute))
validation loss: 0.697445
validation accuracy: 50.00 %
side validation loss: 0.259028
side validation accuracy: nan %
Epoch 3 of 500 took 0.004s
training loss: 0.389688
(target: 0.344722, side: 0.044966 (absolute))
validation loss: 0.694272
validation accuracy: 52.00 %
side validation loss: 0.171608
side validation accuracy: nan %
Epoch 4 of 500 took 0.004s
training loss: 0.354294
(target: 0.341375, side: 0.012919 (absolute))
validation loss: 0.689545
validation accuracy: 50.00 %
side validation loss: 0.143140
side validation accuracy: nan %
Epoch 5 of 500 took 0.004s
training loss: 0.343095
(target: 0.335787, side: 0.007308 (absolute))
validation loss: 0.679085
validation accuracy: 62.00 %
side validation loss: 0.125700
side validation accuracy: nan %
Epoch 6 of 500 took 0.004s
training loss: 0.329919
(target: 0.323550, side: 0.006369 (absolute))
validation loss: 0.653539
validation accuracy: 70.00 %
side validation loss: 0.116395
side validation accuracy: nan %
Epoch 7 of 500 took 0.004s
training loss: 0.307952
(target: 0.301399, side: 0.006553 (absolute))
validation loss: 0.605843
validation accuracy: 73.00 %
side validation loss: 0.116766
side validation accuracy: nan %
Epoch 8 of 500 took 0.005s
training loss: 0.273438
(target: 0.264832, side: 0.008606 (absolute))
validation loss: 0.534339
validation accuracy: 80.00 %
side validation loss: 0.129018
side validation accuracy: nan %
Epoch 9 of 500 took 0.004s
training loss: 0.229569
(target: 0.217878, side: 0.011691 (absolute))
validation loss: 0.451535
validation accuracy: 88.00 %
side validation loss: 0.148894
side validation accuracy: nan %
Epoch 10 of 500 took 0.004s
training loss: 0.185251
(target: 0.172525, side: 0.012726 (absolute))
validation loss: 0.386586
validation accuracy: 88.00 %
side validation loss: 0.170254
side validation accuracy: nan %
Epoch 11 of 500 took 0.004s
training loss: 0.150119
(target: 0.136290, side: 0.013829 (absolute))
validation loss: 0.336170
validation accuracy: 87.00 %
side validation loss: 0.195884
side validation accuracy: nan %
Epoch 12 of 500 took 0.005s
training loss: 0.126842
(target: 0.111549, side: 0.015293 (absolute))
validation loss: 0.302726
validation accuracy: 88.00 %
side validation loss: 0.211059
side validation accuracy: nan %
Epoch 13 of 500 took 0.004s
training loss: 0.107820
(target: 0.093784, side: 0.014036 (absolute))
validation loss: 0.281181
validation accuracy: 89.00 %
side validation loss: 0.233755
side validation accuracy: nan %
Epoch 14 of 500 took 0.004s
training loss: 0.095174
(target: 0.081015, side: 0.014159 (absolute))
validation loss: 0.261562
validation accuracy: 89.00 %
side validation loss: 0.248399
side validation accuracy: nan %
Epoch 15 of 500 took 0.005s
training loss: 0.085662
(target: 0.070896, side: 0.014766 (absolute))
validation loss: 0.250615
validation accuracy: 89.00 %
side validation loss: 0.255128
side validation accuracy: nan %
Epoch 16 of 500 took 0.004s
training loss: 0.078066
(target: 0.065006, side: 0.013059 (absolute))
validation loss: 0.241982
validation accuracy: 89.00 %
side validation loss: 0.270143
side validation accuracy: nan %
Epoch 17 of 500 took 0.004s
training loss: 0.071939
(target: 0.057959, side: 0.013979 (absolute))
validation loss: 0.233651
validation accuracy: 90.00 %
side validation loss: 0.284076
side validation accuracy: nan %
Epoch 18 of 500 took 0.005s
training loss: 0.067285
(target: 0.053793, side: 0.013493 (absolute))
validation loss: 0.224974
validation accuracy: 89.00 %
side validation loss: 0.290438
side validation accuracy: nan %
Epoch 19 of 500 took 0.004s
training loss: 0.063564
(target: 0.049657, side: 0.013907 (absolute))
validation loss: 0.215998
validation accuracy: 90.00 %
side validation loss: 0.294067
side validation accuracy: nan %
Epoch 20 of 500 took 0.004s
training loss: 0.059572
(target: 0.046471, side: 0.013101 (absolute))
validation loss: 0.215216
validation accuracy: 91.00 %
side validation loss: 0.303529
side validation accuracy: nan %
Epoch 21 of 500 took 0.004s
training loss: 0.056929
(target: 0.043655, side: 0.013274 (absolute))
validation loss: 0.209001
validation accuracy: 90.00 %
side validation loss: 0.310360
side validation accuracy: nan %
Epoch 22 of 500 took 0.004s
training loss: 0.053668
(target: 0.040438, side: 0.013230 (absolute))
validation loss: 0.205045
validation accuracy: 90.00 %
side validation loss: 0.315774
side validation accuracy: nan %
Epoch 23 of 500 took 0.005s
training loss: 0.051662
(target: 0.038800, side: 0.012862 (absolute))
validation loss: 0.198977
validation accuracy: 91.00 %
side validation loss: 0.317712
side validation accuracy: nan %
Epoch 24 of 500 took 0.004s
training loss: 0.049286
(target: 0.036805, side: 0.012480 (absolute))
validation loss: 0.196500
validation accuracy: 90.00 %
side validation loss: 0.326186
side validation accuracy: nan %
Epoch 25 of 500 took 0.004s
training loss: 0.047118
(target: 0.035371, side: 0.011747 (absolute))
validation loss: 0.193916
validation accuracy: 91.00 %
side validation loss: 0.325923
side validation accuracy: nan %
Epoch 26 of 500 took 0.006s
training loss: 0.045590
(target: 0.033217, side: 0.012373 (absolute))
validation loss: 0.185526
validation accuracy: 92.00 %
side validation loss: 0.322538
side validation accuracy: nan %
Epoch 27 of 500 took 0.004s
training loss: 0.043795
(target: 0.031535, side: 0.012260 (absolute))
validation loss: 0.181353
validation accuracy: 91.00 %
side validation loss: 0.327163
side validation accuracy: nan %
Epoch 28 of 500 took 0.004s
training loss: 0.042319
(target: 0.030975, side: 0.011344 (absolute))
validation loss: 0.182624
validation accuracy: 91.00 %
side validation loss: 0.329322
side validation accuracy: nan %
Epoch 29 of 500 took 0.004s
training loss: 0.041094
(target: 0.029805, side: 0.011289 (absolute))
validation loss: 0.174700
validation accuracy: 92.00 %
side validation loss: 0.334443
side validation accuracy: nan %
Epoch 30 of 500 took 0.004s
training loss: 0.039618
(target: 0.028485, side: 0.011133 (absolute))
validation loss: 0.173896
validation accuracy: 92.00 %
side validation loss: 0.333489
side validation accuracy: nan %
Epoch 31 of 500 took 0.004s
training loss: 0.038662
(target: 0.027173, side: 0.011489 (absolute))
validation loss: 0.168231
validation accuracy: 92.00 %
side validation loss: 0.331792
side validation accuracy: nan %
Epoch 32 of 500 took 0.004s
training loss: 0.037865
(target: 0.027136, side: 0.010728 (absolute))
validation loss: 0.170550
validation accuracy: 92.00 %
side validation loss: 0.329533
side validation accuracy: nan %
Epoch 33 of 500 took 0.005s
training loss: 0.036525
(target: 0.025485, side: 0.011041 (absolute))
validation loss: 0.162026
validation accuracy: 93.00 %
side validation loss: 0.334490
side validation accuracy: nan %
Epoch 34 of 500 took 0.004s
training loss: 0.035359
(target: 0.025181, side: 0.010177 (absolute))
validation loss: 0.159865
validation accuracy: 93.00 %
side validation loss: 0.334889
side validation accuracy: nan %
Epoch 35 of 500 took 0.004s
training loss: 0.034428
(target: 0.024135, side: 0.010294 (absolute))
validation loss: 0.157652
validation accuracy: 93.00 %
side validation loss: 0.334438
side validation accuracy: nan %
Epoch 36 of 500 took 0.004s
training loss: 0.033437
(target: 0.023370, side: 0.010067 (absolute))
validation loss: 0.154351
validation accuracy: 94.00 %
side validation loss: 0.331729
side validation accuracy: nan %
Epoch 37 of 500 took 0.005s
training loss: 0.033183
(target: 0.022941, side: 0.010242 (absolute))
validation loss: 0.153141
validation accuracy: 94.00 %
side validation loss: 0.331355
side validation accuracy: nan %
Epoch 38 of 500 took 0.005s
training loss: 0.032465
(target: 0.022599, side: 0.009866 (absolute))
validation loss: 0.150296
validation accuracy: 94.00 %
side validation loss: 0.332353
side validation accuracy: nan %
Epoch 39 of 500 took 0.004s
training loss: 0.031797
(target: 0.021608, side: 0.010189 (absolute))
validation loss: 0.147681
validation accuracy: 94.00 %
side validation loss: 0.333668
side validation accuracy: nan %
Epoch 40 of 500 took 0.005s
training loss: 0.030855
(target: 0.021452, side: 0.009403 (absolute))
validation loss: 0.143274
validation accuracy: 94.00 %
side validation loss: 0.330764
side validation accuracy: nan %
Epoch 41 of 500 took 0.004s
training loss: 0.030114
(target: 0.021095, side: 0.009019 (absolute))
validation loss: 0.141862
validation accuracy: 96.00 %
side validation loss: 0.331862
side validation accuracy: nan %
Epoch 42 of 500 took 0.004s
training loss: 0.029509
(target: 0.019811, side: 0.009698 (absolute))
validation loss: 0.142732
validation accuracy: 94.00 %
side validation loss: 0.331570
side validation accuracy: nan %
Epoch 43 of 500 took 0.004s
training loss: 0.028967
(target: 0.019900, side: 0.009068 (absolute))
validation loss: 0.136740
validation accuracy: 94.00 %
side validation loss: 0.328178
side validation accuracy: nan %
Epoch 44 of 500 took 0.004s
training loss: 0.028317
(target: 0.019456, side: 0.008861 (absolute))
validation loss: 0.137746
validation accuracy: 95.00 %
side validation loss: 0.330612
side validation accuracy: nan %
Epoch 45 of 500 took 0.004s
training loss: 0.027761
(target: 0.018771, side: 0.008990 (absolute))
validation loss: 0.134292
validation accuracy: 95.00 %
side validation loss: 0.328936
side validation accuracy: nan %
Epoch 46 of 500 took 0.004s
training loss: 0.027429
(target: 0.018822, side: 0.008607 (absolute))
validation loss: 0.133241
validation accuracy: 95.00 %
side validation loss: 0.328137
side validation accuracy: nan %
Epoch 47 of 500 took 0.004s
training loss: 0.026675
(target: 0.017895, side: 0.008780 (absolute))
validation loss: 0.127976
validation accuracy: 96.00 %
side validation loss: 0.328430
side validation accuracy: nan %
Epoch 48 of 500 took 0.004s
training loss: 0.026356
(target: 0.018095, side: 0.008261 (absolute))
validation loss: 0.131623
validation accuracy: 96.00 %
side validation loss: 0.325831
side validation accuracy: nan %
Epoch 49 of 500 took 0.004s
training loss: 0.025525
(target: 0.017111, side: 0.008414 (absolute))
validation loss: 0.124800
validation accuracy: 96.00 %
side validation loss: 0.323627
side validation accuracy: nan %
Epoch 50 of 500 took 0.004s
training loss: 0.025407
(target: 0.017223, side: 0.008184 (absolute))
validation loss: 0.126369
validation accuracy: 96.00 %
side validation loss: 0.324632
side validation accuracy: nan %
Epoch 51 of 500 took 0.005s
training loss: 0.025321
(target: 0.016653, side: 0.008668 (absolute))
validation loss: 0.124269
validation accuracy: 96.00 %
side validation loss: 0.322760
side validation accuracy: nan %
Epoch 52 of 500 took 0.004s
training loss: 0.024486
(target: 0.016732, side: 0.007754 (absolute))
validation loss: 0.122231
validation accuracy: 96.00 %
side validation loss: 0.321113
side validation accuracy: nan %
Epoch 53 of 500 took 0.004s
training loss: 0.024422
(target: 0.016311, side: 0.008111 (absolute))
validation loss: 0.123595
validation accuracy: 96.00 %
side validation loss: 0.321341
side validation accuracy: nan %
Epoch 54 of 500 took 0.004s
training loss: 0.023975
(target: 0.016033, side: 0.007941 (absolute))
validation loss: 0.119500
validation accuracy: 96.00 %
side validation loss: 0.318932
side validation accuracy: nan %
Epoch 55 of 500 took 0.004s
training loss: 0.023411
(target: 0.015682, side: 0.007730 (absolute))
validation loss: 0.118870
validation accuracy: 96.00 %
side validation loss: 0.319469
side validation accuracy: nan %
Epoch 56 of 500 took 0.004s
training loss: 0.022988
(target: 0.015583, side: 0.007405 (absolute))
validation loss: 0.117347
validation accuracy: 96.00 %
side validation loss: 0.318360
side validation accuracy: nan %
Epoch 57 of 500 took 0.004s
training loss: 0.022704
(target: 0.015030, side: 0.007675 (absolute))
validation loss: 0.116047
validation accuracy: 96.00 %
side validation loss: 0.314206
side validation accuracy: nan %
Epoch 58 of 500 took 0.004s
training loss: 0.022452
(target: 0.014565, side: 0.007888 (absolute))
validation loss: 0.113139
validation accuracy: 96.00 %
side validation loss: 0.317324
side validation accuracy: nan %
Epoch 59 of 500 took 0.004s
training loss: 0.022006
(target: 0.014956, side: 0.007050 (absolute))
validation loss: 0.113767
validation accuracy: 96.00 %
side validation loss: 0.315694
side validation accuracy: nan %
Epoch 60 of 500 took 0.004s
training loss: 0.021894
(target: 0.014389, side: 0.007505 (absolute))
validation loss: 0.112226
validation accuracy: 96.00 %
side validation loss: 0.315493
side validation accuracy: nan %
Epoch 61 of 500 took 0.004s
training loss: 0.021182
(target: 0.014017, side: 0.007165 (absolute))
validation loss: 0.109786
validation accuracy: 96.00 %
side validation loss: 0.313653
side validation accuracy: nan %
Epoch 62 of 500 took 0.004s
training loss: 0.021211
(target: 0.014067, side: 0.007144 (absolute))
validation loss: 0.110283
validation accuracy: 96.00 %
side validation loss: 0.313228
side validation accuracy: nan %
Epoch 63 of 500 took 0.005s
training loss: 0.020902
(target: 0.013846, side: 0.007055 (absolute))
validation loss: 0.108824
validation accuracy: 95.00 %
side validation loss: 0.309794
side validation accuracy: nan %
Epoch 64 of 500 took 0.004s
training loss: 0.020623
(target: 0.013666, side: 0.006957 (absolute))
validation loss: 0.106229
validation accuracy: 95.00 %
side validation loss: 0.308682
side validation accuracy: nan %
Epoch 65 of 500 took 0.004s
training loss: 0.020775
(target: 0.013707, side: 0.007068 (absolute))
validation loss: 0.105619
validation accuracy: 96.00 %
side validation loss: 0.309612
side validation accuracy: nan %
Epoch 66 of 500 took 0.004s
training loss: 0.020297
(target: 0.012991, side: 0.007306 (absolute))
validation loss: 0.104738
validation accuracy: 96.00 %
side validation loss: 0.310867
side validation accuracy: nan %
Epoch 67 of 500 took 0.004s
training loss: 0.020109
(target: 0.013212, side: 0.006897 (absolute))
validation loss: 0.104918
validation accuracy: 95.00 %
side validation loss: 0.307160
side validation accuracy: nan %
Epoch 68 of 500 took 0.004s
training loss: 0.019533
(target: 0.012881, side: 0.006652 (absolute))
validation loss: 0.102962
validation accuracy: 95.00 %
side validation loss: 0.305019
side validation accuracy: nan %
Epoch 69 of 500 took 0.004s
training loss: 0.019213
(target: 0.012451, side: 0.006762 (absolute))
validation loss: 0.100030
validation accuracy: 95.00 %
side validation loss: 0.303997
side validation accuracy: nan %
Epoch 70 of 500 took 0.004s
training loss: 0.019049
(target: 0.012734, side: 0.006316 (absolute))
validation loss: 0.101087
validation accuracy: 95.00 %
side validation loss: 0.301968
side validation accuracy: nan %
Epoch 71 of 500 took 0.004s
training loss: 0.019175
(target: 0.012180, side: 0.006995 (absolute))
validation loss: 0.099722
validation accuracy: 95.00 %
side validation loss: 0.305347
side validation accuracy: nan %
Epoch 72 of 500 took 0.004s
training loss: 0.019184
(target: 0.012409, side: 0.006775 (absolute))
validation loss: 0.097972
validation accuracy: 96.00 %
side validation loss: 0.303356
side validation accuracy: nan %
Epoch 73 of 500 took 0.004s
training loss: 0.018342
(target: 0.011903, side: 0.006439 (absolute))
validation loss: 0.097976
validation accuracy: 96.00 %
side validation loss: 0.301587
side validation accuracy: nan %
Epoch 74 of 500 took 0.004s
training loss: 0.018151
(target: 0.012030, side: 0.006121 (absolute))
validation loss: 0.098428
validation accuracy: 96.00 %
side validation loss: 0.300333
side validation accuracy: nan %
Epoch 75 of 500 took 0.005s
training loss: 0.018015
(target: 0.011711, side: 0.006305 (absolute))
validation loss: 0.097169
validation accuracy: 96.00 %
side validation loss: 0.300944
side validation accuracy: nan %
Epoch 76 of 500 took 0.005s
training loss: 0.017678
(target: 0.011493, side: 0.006185 (absolute))
validation loss: 0.095381
validation accuracy: 96.00 %
side validation loss: 0.297223
side validation accuracy: nan %
Epoch 77 of 500 took 0.004s
training loss: 0.017509
(target: 0.011299, side: 0.006210 (absolute))
validation loss: 0.095558
validation accuracy: 96.00 %
side validation loss: 0.299461
side validation accuracy: nan %
Epoch 78 of 500 took 0.004s
training loss: 0.017552
(target: 0.011260, side: 0.006292 (absolute))
validation loss: 0.095382
validation accuracy: 96.00 %
side validation loss: 0.296979
side validation accuracy: nan %
Epoch 79 of 500 took 0.004s
training loss: 0.017277
(target: 0.011223, side: 0.006054 (absolute))
validation loss: 0.093336
validation accuracy: 96.00 %
side validation loss: 0.295375
side validation accuracy: nan %
Epoch 80 of 500 took 0.004s
training loss: 0.017021
(target: 0.011281, side: 0.005741 (absolute))
validation loss: 0.092981
validation accuracy: 96.00 %
side validation loss: 0.294143
side validation accuracy: nan %
Epoch 81 of 500 took 0.004s
training loss: 0.016881
(target: 0.010650, side: 0.006232 (absolute))
validation loss: 0.092431
validation accuracy: 96.00 %
side validation loss: 0.294741
side validation accuracy: nan %
Epoch 82 of 500 took 0.004s
training loss: 0.016542
(target: 0.010817, side: 0.005725 (absolute))
validation loss: 0.092818
validation accuracy: 96.00 %
side validation loss: 0.296416
side validation accuracy: nan %
Epoch 83 of 500 took 0.004s
training loss: 0.016383
(target: 0.010739, side: 0.005644 (absolute))
validation loss: 0.091988
validation accuracy: 96.00 %
side validation loss: 0.293836
side validation accuracy: nan %
Epoch 84 of 500 took 0.004s
training loss: 0.016716
(target: 0.010714, side: 0.006002 (absolute))
validation loss: 0.090369
validation accuracy: 96.00 %
side validation loss: 0.292274
side validation accuracy: nan %
Epoch 85 of 500 took 0.004s
training loss: 0.016570
(target: 0.010606, side: 0.005964 (absolute))
validation loss: 0.090890
validation accuracy: 96.00 %
side validation loss: 0.293267
side validation accuracy: nan %
Epoch 86 of 500 took 0.004s
training loss: 0.016019
(target: 0.010014, side: 0.006005 (absolute))
validation loss: 0.089484
validation accuracy: 96.00 %
side validation loss: 0.289198
side validation accuracy: nan %
Epoch 87 of 500 took 0.004s
training loss: 0.015799
(target: 0.010275, side: 0.005524 (absolute))
validation loss: 0.088033
validation accuracy: 96.00 %
side validation loss: 0.289553
side validation accuracy: nan %
Epoch 88 of 500 took 0.005s
training loss: 0.015504
(target: 0.009990, side: 0.005514 (absolute))
validation loss: 0.088204
validation accuracy: 96.00 %
side validation loss: 0.288490
side validation accuracy: nan %
Epoch 89 of 500 took 0.004s
training loss: 0.015648
(target: 0.010125, side: 0.005523 (absolute))
validation loss: 0.087050
validation accuracy: 96.00 %
side validation loss: 0.288676
side validation accuracy: nan %
Epoch 90 of 500 took 0.004s
training loss: 0.015188
(target: 0.009666, side: 0.005521 (absolute))
validation loss: 0.087913
validation accuracy: 96.00 %
side validation loss: 0.286120
side validation accuracy: nan %
Epoch 91 of 500 took 0.004s
training loss: 0.015244
(target: 0.009922, side: 0.005321 (absolute))
validation loss: 0.087222
validation accuracy: 96.00 %
side validation loss: 0.285192
side validation accuracy: nan %
Epoch 92 of 500 took 0.004s
training loss: 0.014951
(target: 0.009653, side: 0.005299 (absolute))
validation loss: 0.086427
validation accuracy: 96.00 %
side validation loss: 0.285585
side validation accuracy: nan %
Epoch 93 of 500 took 0.004s
training loss: 0.014830
(target: 0.009535, side: 0.005295 (absolute))
validation loss: 0.085988
validation accuracy: 96.00 %
side validation loss: 0.283849
side validation accuracy: nan %
Epoch 94 of 500 took 0.004s
training loss: 0.014689
(target: 0.009486, side: 0.005203 (absolute))
validation loss: 0.087142
validation accuracy: 96.00 %
side validation loss: 0.282886
side validation accuracy: nan %
Epoch 95 of 500 took 0.004s
training loss: 0.014523
(target: 0.009281, side: 0.005241 (absolute))
validation loss: 0.084195
validation accuracy: 96.00 %
side validation loss: 0.284497
side validation accuracy: nan %
Epoch 96 of 500 took 0.004s
training loss: 0.014814
(target: 0.009400, side: 0.005414 (absolute))
validation loss: 0.084841
validation accuracy: 96.00 %
side validation loss: 0.284031
side validation accuracy: nan %
Epoch 97 of 500 took 0.004s
training loss: 0.014364
(target: 0.009082, side: 0.005282 (absolute))
validation loss: 0.083317
validation accuracy: 96.00 %
side validation loss: 0.281725
side validation accuracy: nan %
Epoch 98 of 500 took 0.004s
training loss: 0.014381
(target: 0.009266, side: 0.005115 (absolute))
validation loss: 0.083161
validation accuracy: 96.00 %
side validation loss: 0.281044
side validation accuracy: nan %
Epoch 99 of 500 took 0.004s
training loss: 0.014109
(target: 0.008938, side: 0.005171 (absolute))
validation loss: 0.084377
validation accuracy: 96.00 %
side validation loss: 0.284493
side validation accuracy: nan %
Epoch 100 of 500 took 0.005s
training loss: 0.013793
(target: 0.008821, side: 0.004972 (absolute))
validation loss: 0.083534
validation accuracy: 96.00 %
side validation loss: 0.281331
side validation accuracy: nan %
Epoch 101 of 500 took 0.004s
training loss: 0.013951
(target: 0.008849, side: 0.005102 (absolute))
validation loss: 0.082167
validation accuracy: 96.00 %
side validation loss: 0.279460
side validation accuracy: nan %
Epoch 102 of 500 took 0.004s
training loss: 0.013870
(target: 0.008749, side: 0.005121 (absolute))
validation loss: 0.083514
validation accuracy: 96.00 %
side validation loss: 0.281133
side validation accuracy: nan %
Epoch 103 of 500 took 0.004s
training loss: 0.013632
(target: 0.008623, side: 0.005008 (absolute))
validation loss: 0.081451
validation accuracy: 96.00 %
side validation loss: 0.276354
side validation accuracy: nan %
Epoch 104 of 500 took 0.004s
training loss: 0.013827
(target: 0.008715, side: 0.005113 (absolute))
validation loss: 0.081758
validation accuracy: 96.00 %
side validation loss: 0.277682
side validation accuracy: nan %
Epoch 105 of 500 took 0.004s
training loss: 0.013451
(target: 0.008460, side: 0.004992 (absolute))
validation loss: 0.081427
validation accuracy: 96.00 %
side validation loss: 0.277245
side validation accuracy: nan %
Epoch 106 of 500 took 0.004s
training loss: 0.013092
(target: 0.008675, side: 0.004417 (absolute))
validation loss: 0.081798
validation accuracy: 96.00 %
side validation loss: 0.275114
side validation accuracy: nan %
Epoch 107 of 500 took 0.004s
training loss: 0.013532
(target: 0.008195, side: 0.005337 (absolute))
validation loss: 0.079224
validation accuracy: 96.00 %
side validation loss: 0.275072
side validation accuracy: nan %
Epoch 108 of 500 took 0.004s
training loss: 0.013196
(target: 0.008283, side: 0.004914 (absolute))
validation loss: 0.080364
validation accuracy: 96.00 %
side validation loss: 0.276942
side validation accuracy: nan %
Epoch 109 of 500 took 0.004s
training loss: 0.013089
(target: 0.008254, side: 0.004835 (absolute))
validation loss: 0.081006
validation accuracy: 96.00 %
side validation loss: 0.274696
side validation accuracy: nan %
Epoch 110 of 500 took 0.004s
training loss: 0.012702
(target: 0.008165, side: 0.004537 (absolute))
validation loss: 0.080721
validation accuracy: 96.00 %
side validation loss: 0.274184
side validation accuracy: nan %
Epoch 111 of 500 took 0.004s
training loss: 0.012649
(target: 0.008096, side: 0.004553 (absolute))
validation loss: 0.079943
validation accuracy: 96.00 %
side validation loss: 0.274055
side validation accuracy: nan %
Epoch 112 of 500 took 0.005s
training loss: 0.012888
(target: 0.008061, side: 0.004827 (absolute))
validation loss: 0.078010
validation accuracy: 96.00 %
side validation loss: 0.272151
side validation accuracy: nan %
Epoch 113 of 500 took 0.004s
training loss: 0.012686
(target: 0.007987, side: 0.004699 (absolute))
validation loss: 0.080232
validation accuracy: 96.00 %
side validation loss: 0.271284
side validation accuracy: nan %
Epoch 114 of 500 took 0.004s
training loss: 0.012823
(target: 0.008103, side: 0.004719 (absolute))
validation loss: 0.080249
validation accuracy: 96.00 %
side validation loss: 0.271568
side validation accuracy: nan %
Epoch 115 of 500 took 0.004s
training loss: 0.012446
(target: 0.007495, side: 0.004951 (absolute))
validation loss: 0.077400
validation accuracy: 96.00 %
side validation loss: 0.272055
side validation accuracy: nan %
Epoch 116 of 500 took 0.004s
training loss: 0.012470
(target: 0.007947, side: 0.004523 (absolute))
validation loss: 0.077024
validation accuracy: 96.00 %
side validation loss: 0.268501
side validation accuracy: nan %
Epoch 117 of 500 took 0.004s
training loss: 0.012362
(target: 0.007922, side: 0.004440 (absolute))
validation loss: 0.076522
validation accuracy: 96.00 %
side validation loss: 0.269189
side validation accuracy: nan %
Epoch 118 of 500 took 0.004s
training loss: 0.012395
(target: 0.007601, side: 0.004794 (absolute))
validation loss: 0.078893
validation accuracy: 96.00 %
side validation loss: 0.268044
side validation accuracy: nan %
Epoch 119 of 500 took 0.004s
training loss: 0.011929
(target: 0.007593, side: 0.004336 (absolute))
validation loss: 0.077931
validation accuracy: 96.00 %
side validation loss: 0.270490
side validation accuracy: nan %
Epoch 120 of 500 took 0.004s
training loss: 0.012180
(target: 0.007583, side: 0.004597 (absolute))
validation loss: 0.076753
validation accuracy: 96.00 %
side validation loss: 0.269310
side validation accuracy: nan %
Epoch 121 of 500 took 0.004s
training loss: 0.012046
(target: 0.007541, side: 0.004505 (absolute))
validation loss: 0.077043
validation accuracy: 96.00 %
side validation loss: 0.267839
side validation accuracy: nan %
Epoch 122 of 500 took 0.004s
training loss: 0.012046
(target: 0.007470, side: 0.004576 (absolute))
validation loss: 0.078308
validation accuracy: 95.00 %
side validation loss: 0.267806
side validation accuracy: nan %
Epoch 123 of 500 took 0.004s
training loss: 0.011648
(target: 0.007328, side: 0.004320 (absolute))
validation loss: 0.076239
validation accuracy: 96.00 %
side validation loss: 0.264105
side validation accuracy: nan %
Epoch 124 of 500 took 0.005s
training loss: 0.011741
(target: 0.007373, side: 0.004368 (absolute))
validation loss: 0.076606
validation accuracy: 96.00 %
side validation loss: 0.268902
side validation accuracy: nan %
Epoch 125 of 500 took 0.004s
training loss: 0.011766
(target: 0.007164, side: 0.004602 (absolute))
validation loss: 0.076364
validation accuracy: 96.00 %
side validation loss: 0.266800
side validation accuracy: nan %
Epoch 126 of 500 took 0.004s
training loss: 0.011413
(target: 0.007261, side: 0.004152 (absolute))
validation loss: 0.076125
validation accuracy: 95.00 %
side validation loss: 0.263902
side validation accuracy: nan %
Epoch 127 of 500 took 0.004s
training loss: 0.011325
(target: 0.006948, side: 0.004378 (absolute))
validation loss: 0.075544
validation accuracy: 96.00 %
side validation loss: 0.262695
side validation accuracy: nan %
Epoch 128 of 500 took 0.004s
training loss: 0.011394
(target: 0.007143, side: 0.004252 (absolute))
validation loss: 0.074537
validation accuracy: 96.00 %
side validation loss: 0.265554
side validation accuracy: nan %
Epoch 129 of 500 took 0.004s
training loss: 0.011282
(target: 0.007011, side: 0.004272 (absolute))
validation loss: 0.075749
validation accuracy: 96.00 %
side validation loss: 0.261951
side validation accuracy: nan %
Epoch 130 of 500 took 0.004s
training loss: 0.011183
(target: 0.006964, side: 0.004219 (absolute))
validation loss: 0.074716
validation accuracy: 96.00 %
side validation loss: 0.265126
side validation accuracy: nan %
Epoch 131 of 500 took 0.004s
training loss: 0.011158
(target: 0.006843, side: 0.004315 (absolute))
validation loss: 0.075369
validation accuracy: 97.00 %
side validation loss: 0.261870
side validation accuracy: nan %
Epoch 132 of 500 took 0.004s
training loss: 0.011041
(target: 0.006820, side: 0.004221 (absolute))
validation loss: 0.075947
validation accuracy: 96.00 %
side validation loss: 0.262590
side validation accuracy: nan %
Epoch 133 of 500 took 0.004s
training loss: 0.010886
(target: 0.006820, side: 0.004066 (absolute))
validation loss: 0.073614
validation accuracy: 96.00 %
side validation loss: 0.261571
side validation accuracy: nan %
Epoch 134 of 500 took 0.004s
training loss: 0.011033
(target: 0.006909, side: 0.004124 (absolute))
validation loss: 0.073147
validation accuracy: 97.00 %
side validation loss: 0.258853
side validation accuracy: nan %
Epoch 135 of 500 took 0.004s
training loss: 0.010885
(target: 0.006664, side: 0.004221 (absolute))
validation loss: 0.075256
validation accuracy: 96.00 %
side validation loss: 0.260251
side validation accuracy: nan %
Epoch 136 of 500 took 0.004s
training loss: 0.010898
(target: 0.006582, side: 0.004316 (absolute))
validation loss: 0.072874
validation accuracy: 97.00 %
side validation loss: 0.259571
side validation accuracy: nan %
Epoch 137 of 500 took 0.004s
training loss: 0.010577
(target: 0.006551, side: 0.004026 (absolute))
validation loss: 0.073943
validation accuracy: 96.00 %
side validation loss: 0.258918
side validation accuracy: nan %
Epoch 138 of 500 took 0.004s
training loss: 0.010670
(target: 0.006559, side: 0.004110 (absolute))
validation loss: 0.073504
validation accuracy: 96.00 %
side validation loss: 0.257952
side validation accuracy: nan %
Epoch 139 of 500 took 0.004s
training loss: 0.010765
(target: 0.006545, side: 0.004221 (absolute))
validation loss: 0.074283
validation accuracy: 97.00 %
side validation loss: 0.257860
side validation accuracy: nan %
Epoch 140 of 500 took 0.004s
training loss: 0.010505
(target: 0.006604, side: 0.003901 (absolute))
validation loss: 0.074412
validation accuracy: 96.00 %
side validation loss: 0.256816
side validation accuracy: nan %
Epoch 141 of 500 took 0.004s
training loss: 0.010452
(target: 0.006458, side: 0.003994 (absolute))
validation loss: 0.072778
validation accuracy: 96.00 %
side validation loss: 0.256645
side validation accuracy: nan %
Epoch 142 of 500 took 0.004s
training loss: 0.010403
(target: 0.006447, side: 0.003957 (absolute))
validation loss: 0.074182
validation accuracy: 96.00 %
side validation loss: 0.256472
side validation accuracy: nan %
Epoch 143 of 500 took 0.004s
training loss: 0.010451
(target: 0.006254, side: 0.004196 (absolute))
validation loss: 0.073684
validation accuracy: 96.00 %
side validation loss: 0.253969
side validation accuracy: nan %
Epoch 144 of 500 took 0.004s
training loss: 0.010278
(target: 0.006465, side: 0.003814 (absolute))
validation loss: 0.073210
validation accuracy: 96.00 %
side validation loss: 0.254538
side validation accuracy: nan %
Epoch 145 of 500 took 0.004s
training loss: 0.010009
(target: 0.006196, side: 0.003813 (absolute))
validation loss: 0.074377
validation accuracy: 96.00 %
side validation loss: 0.254055
side validation accuracy: nan %
Epoch 146 of 500 took 0.004s
training loss: 0.010233
(target: 0.006297, side: 0.003936 (absolute))
validation loss: 0.074488
validation accuracy: 96.00 %
side validation loss: 0.254896
side validation accuracy: nan %
Epoch 147 of 500 took 0.004s
training loss: 0.010072
(target: 0.006400, side: 0.003671 (absolute))
validation loss: 0.072135
validation accuracy: 97.00 %
side validation loss: 0.254716
side validation accuracy: nan %
Epoch 148 of 500 took 0.004s
training loss: 0.009989
(target: 0.005980, side: 0.004008 (absolute))
validation loss: 0.074699
validation accuracy: 96.00 %
side validation loss: 0.252569
side validation accuracy: nan %
Epoch 149 of 500 took 0.005s
training loss: 0.009847
(target: 0.006192, side: 0.003655 (absolute))
validation loss: 0.072174
validation accuracy: 96.00 %
side validation loss: 0.253663
side validation accuracy: nan %
Epoch 150 of 500 took 0.004s
training loss: 0.009918
(target: 0.005893, side: 0.004025 (absolute))
validation loss: 0.073436
validation accuracy: 96.00 %
side validation loss: 0.255130
side validation accuracy: nan %
Epoch 151 of 500 took 0.004s
training loss: 0.009958
(target: 0.006287, side: 0.003671 (absolute))
validation loss: 0.071129
validation accuracy: 96.00 %
side validation loss: 0.253602
side validation accuracy: nan %
Epoch 152 of 500 took 0.004s
training loss: 0.009803
(target: 0.006098, side: 0.003704 (absolute))
validation loss: 0.072233
validation accuracy: 97.00 %
side validation loss: 0.250011
side validation accuracy: nan %
Epoch 153 of 500 took 0.004s
training loss: 0.009827
(target: 0.005845, side: 0.003982 (absolute))
validation loss: 0.071391
validation accuracy: 96.00 %
side validation loss: 0.250823
side validation accuracy: nan %
Epoch 154 of 500 took 0.005s
training loss: 0.009596
(target: 0.005901, side: 0.003694 (absolute))
validation loss: 0.071699
validation accuracy: 96.00 %
side validation loss: 0.251258
side validation accuracy: nan %
Epoch 155 of 500 took 0.005s
training loss: 0.009525
(target: 0.005941, side: 0.003584 (absolute))
validation loss: 0.070708
validation accuracy: 97.00 %
side validation loss: 0.249907
side validation accuracy: nan %
Epoch 156 of 500 took 0.004s
training loss: 0.009606
(target: 0.005871, side: 0.003734 (absolute))
validation loss: 0.071815
validation accuracy: 96.00 %
side validation loss: 0.251711
side validation accuracy: nan %
Epoch 157 of 500 took 0.004s
training loss: 0.009431
(target: 0.005890, side: 0.003541 (absolute))
validation loss: 0.071990
validation accuracy: 96.00 %
side validation loss: 0.248411
side validation accuracy: nan %
Epoch 158 of 500 took 0.006s
training loss: 0.009337
(target: 0.005734, side: 0.003604 (absolute))
validation loss: 0.071762
validation accuracy: 97.00 %
side validation loss: 0.251329
side validation accuracy: nan %
Epoch 159 of 500 took 0.004s
training loss: 0.009255
(target: 0.005713, side: 0.003542 (absolute))
validation loss: 0.070844
validation accuracy: 96.00 %
side validation loss: 0.250464
side validation accuracy: nan %
Epoch 160 of 500 took 0.004s
training loss: 0.009319
(target: 0.005598, side: 0.003720 (absolute))
validation loss: 0.071588
validation accuracy: 96.00 %
side validation loss: 0.247922
side validation accuracy: nan %
Epoch 161 of 500 took 0.004s
training loss: 0.009439
(target: 0.005700, side: 0.003739 (absolute))
validation loss: 0.072826
validation accuracy: 96.00 %
side validation loss: 0.248150
side validation accuracy: nan %
Epoch 162 of 500 took 0.005s
training loss: 0.009333
(target: 0.005669, side: 0.003664 (absolute))
validation loss: 0.071386
validation accuracy: 96.00 %
side validation loss: 0.246700
side validation accuracy: nan %
Epoch 163 of 500 took 0.004s
training loss: 0.009103
(target: 0.005731, side: 0.003372 (absolute))
validation loss: 0.069724
validation accuracy: 96.00 %
side validation loss: 0.247602
side validation accuracy: nan %
Epoch 164 of 500 took 0.004s
training loss: 0.009046
(target: 0.005584, side: 0.003461 (absolute))
validation loss: 0.071422
validation accuracy: 96.00 %
side validation loss: 0.247245
side validation accuracy: nan %
Epoch 165 of 500 took 0.004s
training loss: 0.009071
(target: 0.005454, side: 0.003617 (absolute))
validation loss: 0.071231
validation accuracy: 96.00 %
side validation loss: 0.244878
side validation accuracy: nan %
Epoch 166 of 500 took 0.004s
training loss: 0.009246
(target: 0.005642, side: 0.003605 (absolute))
validation loss: 0.070043
validation accuracy: 96.00 %
side validation loss: 0.244702
side validation accuracy: nan %
Epoch 167 of 500 took 0.004s
training loss: 0.009064
(target: 0.005402, side: 0.003662 (absolute))
validation loss: 0.072135
validation accuracy: 96.00 %
side validation loss: 0.245404
side validation accuracy: nan %
Epoch 168 of 500 took 0.004s
training loss: 0.008949
(target: 0.005519, side: 0.003430 (absolute))
validation loss: 0.070494
validation accuracy: 96.00 %
side validation loss: 0.246125
side validation accuracy: nan %
Epoch 169 of 500 took 0.004s
training loss: 0.009030
(target: 0.005564, side: 0.003466 (absolute))
validation loss: 0.071401
validation accuracy: 96.00 %
side validation loss: 0.244390
side validation accuracy: nan %
Epoch 170 of 500 took 0.004s
training loss: 0.008987
(target: 0.005378, side: 0.003609 (absolute))
validation loss: 0.069448
validation accuracy: 96.00 %
side validation loss: 0.244974
side validation accuracy: nan %
Epoch 171 of 500 took 0.004s
training loss: 0.008733
(target: 0.005319, side: 0.003414 (absolute))
validation loss: 0.070461
validation accuracy: 96.00 %
side validation loss: 0.244470
side validation accuracy: nan %
Epoch 172 of 500 took 0.004s
training loss: 0.008789
(target: 0.005357, side: 0.003432 (absolute))
validation loss: 0.069856
validation accuracy: 96.00 %
side validation loss: 0.244962
side validation accuracy: nan %
Epoch 173 of 500 took 0.005s
training loss: 0.008761
(target: 0.005256, side: 0.003505 (absolute))
validation loss: 0.070377
validation accuracy: 96.00 %
side validation loss: 0.245581
side validation accuracy: nan %
Epoch 174 of 500 took 0.004s
training loss: 0.008879
(target: 0.005369, side: 0.003510 (absolute))
validation loss: 0.069840
validation accuracy: 96.00 %
side validation loss: 0.243171
side validation accuracy: nan %
Epoch 175 of 500 took 0.004s
training loss: 0.008761
(target: 0.005151, side: 0.003610 (absolute))
validation loss: 0.070374
validation accuracy: 96.00 %
side validation loss: 0.241337
side validation accuracy: nan %
Epoch 176 of 500 took 0.004s
training loss: 0.008770
(target: 0.005481, side: 0.003289 (absolute))
validation loss: 0.070377
validation accuracy: 96.00 %
side validation loss: 0.243597
side validation accuracy: nan %
Epoch 177 of 500 took 0.004s
training loss: 0.008496
(target: 0.005167, side: 0.003329 (absolute))
validation loss: 0.069165
validation accuracy: 96.00 %
side validation loss: 0.243807
side validation accuracy: nan %
Epoch 178 of 500 took 0.004s
training loss: 0.008612
(target: 0.005097, side: 0.003515 (absolute))
validation loss: 0.070276
validation accuracy: 96.00 %
side validation loss: 0.240651
side validation accuracy: nan %
Epoch 179 of 500 took 0.004s
training loss: 0.008545
(target: 0.005141, side: 0.003404 (absolute))
validation loss: 0.069957
validation accuracy: 96.00 %
side validation loss: 0.240492
side validation accuracy: nan %
Epoch 180 of 500 took 0.004s
training loss: 0.008527
(target: 0.005087, side: 0.003440 (absolute))
validation loss: 0.068947
validation accuracy: 96.00 %
side validation loss: 0.239533
side validation accuracy: nan %
Epoch 181 of 500 took 0.004s
training loss: 0.008556
(target: 0.005245, side: 0.003311 (absolute))
validation loss: 0.069168
validation accuracy: 96.00 %
side validation loss: 0.239650
side validation accuracy: nan %
Epoch 182 of 500 took 0.004s
training loss: 0.008319
(target: 0.004964, side: 0.003355 (absolute))
validation loss: 0.069222
validation accuracy: 96.00 %
side validation loss: 0.239019
side validation accuracy: nan %
Epoch 183 of 500 took 0.004s
training loss: 0.008698
(target: 0.005132, side: 0.003566 (absolute))
validation loss: 0.070409
validation accuracy: 96.00 %
side validation loss: 0.238624
side validation accuracy: nan %
Epoch 184 of 500 took 0.004s
training loss: 0.008330
(target: 0.005024, side: 0.003306 (absolute))
validation loss: 0.071317
validation accuracy: 96.00 %
side validation loss: 0.239270
side validation accuracy: nan %
Epoch 185 of 500 took 0.004s
training loss: 0.008343
(target: 0.005045, side: 0.003299 (absolute))
validation loss: 0.069230
validation accuracy: 96.00 %
side validation loss: 0.239362
side validation accuracy: nan %
Epoch 186 of 500 took 0.005s
training loss: 0.008171
(target: 0.004913, side: 0.003258 (absolute))
validation loss: 0.069213
validation accuracy: 96.00 %
side validation loss: 0.235936
side validation accuracy: nan %
Epoch 187 of 500 took 0.004s
training loss: 0.008265
(target: 0.004976, side: 0.003289 (absolute))
validation loss: 0.068587
validation accuracy: 96.00 %
side validation loss: 0.238184
side validation accuracy: nan %
Epoch 188 of 500 took 0.004s
training loss: 0.008232
(target: 0.004901, side: 0.003331 (absolute))
validation loss: 0.070250
validation accuracy: 96.00 %
side validation loss: 0.235738
side validation accuracy: nan %
Epoch 189 of 500 took 0.004s
training loss: 0.008235
(target: 0.004929, side: 0.003306 (absolute))
validation loss: 0.069525
validation accuracy: 96.00 %
side validation loss: 0.235656
side validation accuracy: nan %
Epoch 190 of 500 took 0.004s
training loss: 0.008069
(target: 0.004894, side: 0.003176 (absolute))
validation loss: 0.069485
validation accuracy: 96.00 %
side validation loss: 0.234905
side validation accuracy: nan %
Epoch 191 of 500 took 0.004s
training loss: 0.008244
(target: 0.004892, side: 0.003352 (absolute))
validation loss: 0.069853
validation accuracy: 96.00 %
side validation loss: 0.236101
side validation accuracy: nan %
Epoch 192 of 500 took 0.004s
training loss: 0.007987
(target: 0.004813, side: 0.003174 (absolute))
validation loss: 0.069687
validation accuracy: 96.00 %
side validation loss: 0.235266
side validation accuracy: nan %
Epoch 193 of 500 took 0.004s
training loss: 0.007904
(target: 0.004733, side: 0.003171 (absolute))
validation loss: 0.068435
validation accuracy: 96.00 %
side validation loss: 0.235120
side validation accuracy: nan %
Epoch 194 of 500 took 0.004s
training loss: 0.008009
(target: 0.004765, side: 0.003244 (absolute))
validation loss: 0.069736
validation accuracy: 96.00 %
side validation loss: 0.235537
side validation accuracy: nan %
Epoch 195 of 500 took 0.004s
training loss: 0.007899
(target: 0.004754, side: 0.003145 (absolute))
validation loss: 0.069988
validation accuracy: 96.00 %
side validation loss: 0.234655
side validation accuracy: nan %
Epoch 196 of 500 took 0.004s
training loss: 0.008036
(target: 0.004783, side: 0.003254 (absolute))
validation loss: 0.070086
validation accuracy: 96.00 %
side validation loss: 0.234733
side validation accuracy: nan %
Epoch 197 of 500 took 0.004s
training loss: 0.007997
(target: 0.004627, side: 0.003369 (absolute))
validation loss: 0.069379
validation accuracy: 96.00 %
side validation loss: 0.231893
side validation accuracy: nan %
Epoch 198 of 500 took 0.005s
training loss: 0.007802
(target: 0.004719, side: 0.003083 (absolute))
validation loss: 0.067682
validation accuracy: 96.00 %
side validation loss: 0.233858
side validation accuracy: nan %
Epoch 199 of 500 took 0.004s
training loss: 0.007984
(target: 0.004794, side: 0.003191 (absolute))
validation loss: 0.068931
validation accuracy: 96.00 %
side validation loss: 0.232064
side validation accuracy: nan %
Epoch 200 of 500 took 0.004s
training loss: 0.007750
(target: 0.004704, side: 0.003046 (absolute))
validation loss: 0.069294
validation accuracy: 96.00 %
side validation loss: 0.234957
side validation accuracy: nan %
Epoch 201 of 500 took 0.004s
training loss: 0.007627
(target: 0.004477, side: 0.003151 (absolute))
validation loss: 0.068362
validation accuracy: 96.00 %
side validation loss: 0.231443
side validation accuracy: nan %
Epoch 202 of 500 took 0.004s
training loss: 0.007758
(target: 0.004606, side: 0.003152 (absolute))
validation loss: 0.069090
validation accuracy: 96.00 %
side validation loss: 0.232357
side validation accuracy: nan %
Epoch 203 of 500 took 0.004s
training loss: 0.007518
(target: 0.004533, side: 0.002985 (absolute))
validation loss: 0.068925
validation accuracy: 96.00 %
side validation loss: 0.230370
side validation accuracy: nan %
Epoch 204 of 500 took 0.004s
training loss: 0.007518
(target: 0.004597, side: 0.002921 (absolute))
validation loss: 0.067320
validation accuracy: 96.00 %
side validation loss: 0.231946
side validation accuracy: nan %
Epoch 205 of 500 took 0.004s
training loss: 0.007764
(target: 0.004496, side: 0.003268 (absolute))
validation loss: 0.069294
validation accuracy: 96.00 %
side validation loss: 0.230030
side validation accuracy: nan %
Epoch 206 of 500 took 0.004s
training loss: 0.007473
(target: 0.004472, side: 0.003001 (absolute))
validation loss: 0.067419
validation accuracy: 96.00 %
side validation loss: 0.231121
side validation accuracy: nan %
Epoch 207 of 500 took 0.004s
training loss: 0.007474
(target: 0.004580, side: 0.002894 (absolute))
validation loss: 0.067331
validation accuracy: 96.00 %
side validation loss: 0.228437
side validation accuracy: nan %
Epoch 208 of 500 took 0.004s
training loss: 0.007506
(target: 0.004450, side: 0.003056 (absolute))
validation loss: 0.068017
validation accuracy: 96.00 %
side validation loss: 0.230585
side validation accuracy: nan %
Epoch 209 of 500 took 0.004s
training loss: 0.007416
(target: 0.004415, side: 0.003000 (absolute))
validation loss: 0.066787
validation accuracy: 96.00 %
side validation loss: 0.228854
side validation accuracy: nan %
Epoch 210 of 500 took 0.005s
training loss: 0.007361
(target: 0.004422, side: 0.002939 (absolute))
validation loss: 0.067285
validation accuracy: 96.00 %
side validation loss: 0.229092
side validation accuracy: nan %
Epoch 211 of 500 took 0.004s
training loss: 0.007454
(target: 0.004380, side: 0.003074 (absolute))
validation loss: 0.068055
validation accuracy: 96.00 %
side validation loss: 0.227879
side validation accuracy: nan %
Epoch 212 of 500 took 0.004s
training loss: 0.007361
(target: 0.004438, side: 0.002923 (absolute))
validation loss: 0.066717
validation accuracy: 96.00 %
side validation loss: 0.226481
side validation accuracy: nan %
Epoch 213 of 500 took 0.004s
training loss: 0.007379
(target: 0.004409, side: 0.002970 (absolute))
validation loss: 0.067844
validation accuracy: 96.00 %
side validation loss: 0.226681
side validation accuracy: nan %
Epoch 214 of 500 took 0.004s
training loss: 0.007250
(target: 0.004309, side: 0.002941 (absolute))
validation loss: 0.068071
validation accuracy: 96.00 %
side validation loss: 0.227773
side validation accuracy: nan %
Epoch 215 of 500 took 0.004s
training loss: 0.007159
(target: 0.004343, side: 0.002816 (absolute))
validation loss: 0.067462
validation accuracy: 96.00 %
side validation loss: 0.227829
side validation accuracy: nan %
Epoch 216 of 500 took 0.004s
training loss: 0.007317
(target: 0.004288, side: 0.003029 (absolute))
validation loss: 0.067678
validation accuracy: 96.00 %
side validation loss: 0.225970
side validation accuracy: nan %
Epoch 217 of 500 took 0.004s
training loss: 0.007253
(target: 0.004305, side: 0.002948 (absolute))
validation loss: 0.067261
validation accuracy: 96.00 %
side validation loss: 0.226899
side validation accuracy: nan %
Epoch 218 of 500 took 0.004s
training loss: 0.007201
(target: 0.004250, side: 0.002951 (absolute))
validation loss: 0.067158
validation accuracy: 96.00 %
side validation loss: 0.225978
side validation accuracy: nan %
Epoch 219 of 500 took 0.004s
training loss: 0.007193
(target: 0.004286, side: 0.002907 (absolute))
validation loss: 0.067697
validation accuracy: 96.00 %
side validation loss: 0.226210
side validation accuracy: nan %
Epoch 220 of 500 took 0.004s
training loss: 0.007068
(target: 0.004164, side: 0.002905 (absolute))
validation loss: 0.068680
validation accuracy: 96.00 %
side validation loss: 0.225415
side validation accuracy: nan %
Epoch 221 of 500 took 0.004s
training loss: 0.007050
(target: 0.004215, side: 0.002835 (absolute))
validation loss: 0.068179
validation accuracy: 96.00 %
side validation loss: 0.224623
side validation accuracy: nan %
Epoch 222 of 500 took 0.005s
training loss: 0.007198
(target: 0.004184, side: 0.003014 (absolute))
validation loss: 0.067254
validation accuracy: 96.00 %
side validation loss: 0.224346
side validation accuracy: nan %
Epoch 223 of 500 took 0.004s
training loss: 0.007161
(target: 0.004241, side: 0.002919 (absolute))
validation loss: 0.066039
validation accuracy: 96.00 %
side validation loss: 0.225044
side validation accuracy: nan %
Epoch 224 of 500 took 0.004s
training loss: 0.007260
(target: 0.004142, side: 0.003118 (absolute))
validation loss: 0.067194
validation accuracy: 96.00 %
side validation loss: 0.224824
side validation accuracy: nan %
Epoch 225 of 500 took 0.004s
training loss: 0.007020
(target: 0.004255, side: 0.002765 (absolute))
validation loss: 0.069200
validation accuracy: 96.00 %
side validation loss: 0.222674
side validation accuracy: nan %
Epoch 226 of 500 took 0.004s
training loss: 0.007040
(target: 0.004152, side: 0.002888 (absolute))
validation loss: 0.067099
validation accuracy: 96.00 %
side validation loss: 0.225070
side validation accuracy: nan %
Epoch 227 of 500 took 0.004s
training loss: 0.006955
(target: 0.004137, side: 0.002818 (absolute))
validation loss: 0.067032
validation accuracy: 96.00 %
side validation loss: 0.222839
side validation accuracy: nan %
Epoch 228 of 500 took 0.004s
training loss: 0.006816
(target: 0.004096, side: 0.002719 (absolute))
validation loss: 0.065816
validation accuracy: 96.00 %
side validation loss: 0.221982
side validation accuracy: nan %
Epoch 229 of 500 took 0.004s
training loss: 0.006975
(target: 0.004050, side: 0.002925 (absolute))
validation loss: 0.067802
validation accuracy: 96.00 %
side validation loss: 0.221483
side validation accuracy: nan %
Epoch 230 of 500 took 0.004s
training loss: 0.006969
(target: 0.004086, side: 0.002883 (absolute))
validation loss: 0.067316
validation accuracy: 96.00 %
side validation loss: 0.221705
side validation accuracy: nan %
Epoch 231 of 500 took 0.004s
training loss: 0.006813
(target: 0.004039, side: 0.002774 (absolute))
validation loss: 0.068497
validation accuracy: 96.00 %
side validation loss: 0.220396
side validation accuracy: nan %
Epoch 232 of 500 took 0.004s
training loss: 0.006894
(target: 0.004030, side: 0.002864 (absolute))
validation loss: 0.066863
validation accuracy: 96.00 %
side validation loss: 0.221740
side validation accuracy: nan %
Epoch 233 of 500 took 0.004s
training loss: 0.006926
(target: 0.003981, side: 0.002945 (absolute))
validation loss: 0.066686
validation accuracy: 96.00 %
side validation loss: 0.219907
side validation accuracy: nan %
Epoch 234 of 500 took 0.004s
training loss: 0.006695
(target: 0.004009, side: 0.002686 (absolute))
validation loss: 0.067941
validation accuracy: 96.00 %
side validation loss: 0.220131
side validation accuracy: nan %
Epoch 235 of 500 took 0.005s
training loss: 0.006773
(target: 0.003944, side: 0.002829 (absolute))
validation loss: 0.068480
validation accuracy: 97.00 %
side validation loss: 0.219360
side validation accuracy: nan %
Epoch 236 of 500 took 0.004s
training loss: 0.006750
(target: 0.004027, side: 0.002723 (absolute))
validation loss: 0.068676
validation accuracy: 96.00 %
side validation loss: 0.219162
side validation accuracy: nan %
Epoch 237 of 500 took 0.004s
training loss: 0.006812
(target: 0.003905, side: 0.002907 (absolute))
validation loss: 0.066932
validation accuracy: 96.00 %
side validation loss: 0.218875
side validation accuracy: nan %
Epoch 238 of 500 took 0.004s
training loss: 0.006717
(target: 0.003930, side: 0.002787 (absolute))
validation loss: 0.067011
validation accuracy: 96.00 %
side validation loss: 0.217310
side validation accuracy: nan %
Epoch 239 of 500 took 0.004s
training loss: 0.006635
(target: 0.003957, side: 0.002678 (absolute))
validation loss: 0.067912
validation accuracy: 96.00 %
side validation loss: 0.215180
side validation accuracy: nan %
Epoch 240 of 500 took 0.004s
training loss: 0.006523
(target: 0.003938, side: 0.002585 (absolute))
validation loss: 0.068054
validation accuracy: 97.00 %
side validation loss: 0.218223
side validation accuracy: nan %
Epoch 241 of 500 took 0.004s
training loss: 0.006630
(target: 0.003926, side: 0.002705 (absolute))
validation loss: 0.066132
validation accuracy: 96.00 %
side validation loss: 0.216546
side validation accuracy: nan %
Epoch 242 of 500 took 0.004s
training loss: 0.006571
(target: 0.003865, side: 0.002706 (absolute))
validation loss: 0.066870
validation accuracy: 97.00 %
side validation loss: 0.217160
side validation accuracy: nan %
Epoch 243 of 500 took 0.004s
training loss: 0.006707
(target: 0.003903, side: 0.002804 (absolute))
validation loss: 0.066832
validation accuracy: 96.00 %
side validation loss: 0.216875
side validation accuracy: nan %
Epoch 244 of 500 took 0.004s
training loss: 0.006642
(target: 0.003897, side: 0.002745 (absolute))
validation loss: 0.066414
validation accuracy: 97.00 %
side validation loss: 0.215719
side validation accuracy: nan %
Epoch 245 of 500 took 0.004s
training loss: 0.006484
(target: 0.003790, side: 0.002693 (absolute))
validation loss: 0.067225
validation accuracy: 96.00 %
side validation loss: 0.217314
side validation accuracy: nan %
Epoch 246 of 500 took 0.004s
training loss: 0.006627
(target: 0.003840, side: 0.002786 (absolute))
validation loss: 0.065764
validation accuracy: 97.00 %
side validation loss: 0.214508
side validation accuracy: nan %
Epoch 247 of 500 took 0.005s
training loss: 0.006729
(target: 0.003723, side: 0.003006 (absolute))
validation loss: 0.066590
validation accuracy: 96.00 %
side validation loss: 0.215457
side validation accuracy: nan %
Epoch 248 of 500 took 0.004s
training loss: 0.006645
(target: 0.003955, side: 0.002690 (absolute))
validation loss: 0.066429
validation accuracy: 96.00 %
side validation loss: 0.215901
side validation accuracy: nan %
Epoch 249 of 500 took 0.004s
training loss: 0.006412
(target: 0.003831, side: 0.002581 (absolute))
validation loss: 0.067551
validation accuracy: 97.00 %
side validation loss: 0.215015
side validation accuracy: nan %
Epoch 250 of 500 took 0.004s
training loss: 0.006394
(target: 0.003695, side: 0.002699 (absolute))
validation loss: 0.067491
validation accuracy: 97.00 %
side validation loss: 0.213802
side validation accuracy: nan %
Epoch 251 of 500 took 0.004s
training loss: 0.006295
(target: 0.003830, side: 0.002465 (absolute))
validation loss: 0.066717
validation accuracy: 96.00 %
side validation loss: 0.213823
side validation accuracy: nan %
Epoch 252 of 500 took 0.004s
training loss: 0.006366
(target: 0.003672, side: 0.002694 (absolute))
validation loss: 0.066666
validation accuracy: 97.00 %
side validation loss: 0.212981
side validation accuracy: nan %
Epoch 253 of 500 took 0.004s
training loss: 0.006402
(target: 0.003772, side: 0.002630 (absolute))
validation loss: 0.067236
validation accuracy: 96.00 %
side validation loss: 0.214999
side validation accuracy: nan %
Epoch 254 of 500 took 0.004s
training loss: 0.006409
(target: 0.003673, side: 0.002736 (absolute))
validation loss: 0.066765
validation accuracy: 97.00 %
side validation loss: 0.213395
side validation accuracy: nan %
Epoch 255 of 500 took 0.004s
training loss: 0.006331
(target: 0.003876, side: 0.002456 (absolute))
validation loss: 0.066347
validation accuracy: 96.00 %
side validation loss: 0.212840
side validation accuracy: nan %
Epoch 256 of 500 took 0.004s
training loss: 0.006255
(target: 0.003562, side: 0.002693 (absolute))
validation loss: 0.066141
validation accuracy: 96.00 %
side validation loss: 0.212227
side validation accuracy: nan %
Epoch 257 of 500 took 0.004s
training loss: 0.006268
(target: 0.003697, side: 0.002571 (absolute))
validation loss: 0.065848
validation accuracy: 97.00 %
side validation loss: 0.211326
side validation accuracy: nan %
Epoch 258 of 500 took 0.004s
training loss: 0.006342
(target: 0.003582, side: 0.002761 (absolute))
validation loss: 0.066707
validation accuracy: 97.00 %
side validation loss: 0.210561
side validation accuracy: nan %
Epoch 259 of 500 took 0.005s
training loss: 0.006235
(target: 0.003723, side: 0.002512 (absolute))
validation loss: 0.065671
validation accuracy: 97.00 %
side validation loss: 0.211080
side validation accuracy: nan %
Epoch 260 of 500 took 0.004s
training loss: 0.006308
(target: 0.003568, side: 0.002740 (absolute))
validation loss: 0.066798
validation accuracy: 96.00 %
side validation loss: 0.210955
side validation accuracy: nan %
Epoch 261 of 500 took 0.004s
training loss: 0.006246
(target: 0.003797, side: 0.002449 (absolute))
validation loss: 0.066036
validation accuracy: 97.00 %
side validation loss: 0.211799
side validation accuracy: nan %
Epoch 262 of 500 took 0.004s
training loss: 0.006218
(target: 0.003497, side: 0.002721 (absolute))
validation loss: 0.067062
validation accuracy: 97.00 %
side validation loss: 0.210428
side validation accuracy: nan %
Epoch 263 of 500 took 0.004s
training loss: 0.006162
(target: 0.003649, side: 0.002513 (absolute))
validation loss: 0.065481
validation accuracy: 97.00 %
side validation loss: 0.211516
side validation accuracy: nan %
Epoch 264 of 500 took 0.004s
training loss: 0.006175
(target: 0.003447, side: 0.002727 (absolute))
validation loss: 0.066369
validation accuracy: 96.00 %
side validation loss: 0.209683
side validation accuracy: nan %
Epoch 265 of 500 took 0.004s
training loss: 0.006196
(target: 0.003595, side: 0.002600 (absolute))
validation loss: 0.066085
validation accuracy: 97.00 %
side validation loss: 0.209052
side validation accuracy: nan %
Epoch 266 of 500 took 0.004s
training loss: 0.006064
(target: 0.003608, side: 0.002456 (absolute))
validation loss: 0.066142
validation accuracy: 97.00 %
side validation loss: 0.210394
side validation accuracy: nan %
Epoch 267 of 500 took 0.004s
training loss: 0.006049
(target: 0.003602, side: 0.002447 (absolute))
validation loss: 0.066586
validation accuracy: 97.00 %
side validation loss: 0.208905
side validation accuracy: nan %
Epoch 268 of 500 took 0.004s
training loss: 0.005993
(target: 0.003536, side: 0.002456 (absolute))
validation loss: 0.066237
validation accuracy: 97.00 %
side validation loss: 0.210030
side validation accuracy: nan %
Epoch 269 of 500 took 0.004s
training loss: 0.006013
(target: 0.003457, side: 0.002556 (absolute))
validation loss: 0.066390
validation accuracy: 97.00 %
side validation loss: 0.209042
side validation accuracy: nan %
Epoch 270 of 500 took 0.004s
training loss: 0.005961
(target: 0.003594, side: 0.002367 (absolute))
validation loss: 0.066158
validation accuracy: 97.00 %
side validation loss: 0.209111
side validation accuracy: nan %
Epoch 271 of 500 took 0.005s
training loss: 0.006127
(target: 0.003454, side: 0.002673 (absolute))
validation loss: 0.066051
validation accuracy: 97.00 %
side validation loss: 0.206935
side validation accuracy: nan %
Epoch 272 of 500 took 0.004s
training loss: 0.006001
(target: 0.003531, side: 0.002470 (absolute))
validation loss: 0.065170
validation accuracy: 97.00 %
side validation loss: 0.207768
side validation accuracy: nan %
Epoch 273 of 500 took 0.004s
training loss: 0.005931
(target: 0.003427, side: 0.002504 (absolute))
validation loss: 0.064576
validation accuracy: 97.00 %
side validation loss: 0.207300
side validation accuracy: nan %
Epoch 274 of 500 took 0.004s
training loss: 0.005985
(target: 0.003398, side: 0.002586 (absolute))
validation loss: 0.066402
validation accuracy: 97.00 %
side validation loss: 0.207272
side validation accuracy: nan %
Epoch 275 of 500 took 0.004s
training loss: 0.005997
(target: 0.003419, side: 0.002578 (absolute))
validation loss: 0.065458
validation accuracy: 98.00 %
side validation loss: 0.207377
side validation accuracy: nan %
Epoch 276 of 500 took 0.004s
training loss: 0.005894
(target: 0.003505, side: 0.002389 (absolute))
validation loss: 0.065822
validation accuracy: 97.00 %
side validation loss: 0.205813
side validation accuracy: nan %
Epoch 277 of 500 took 0.004s
training loss: 0.005876
(target: 0.003398, side: 0.002478 (absolute))
validation loss: 0.065559
validation accuracy: 97.00 %
side validation loss: 0.206163
side validation accuracy: nan %
Epoch 278 of 500 took 0.004s
training loss: 0.005929
(target: 0.003424, side: 0.002505 (absolute))
validation loss: 0.065172
validation accuracy: 97.00 %
side validation loss: 0.205117
side validation accuracy: nan %
Epoch 279 of 500 took 0.004s
training loss: 0.006013
(target: 0.003463, side: 0.002551 (absolute))
validation loss: 0.066835
validation accuracy: 97.00 %
side validation loss: 0.206228
side validation accuracy: nan %
Epoch 280 of 500 took 0.004s
training loss: 0.005897
(target: 0.003382, side: 0.002514 (absolute))
validation loss: 0.064461
validation accuracy: 97.00 %
side validation loss: 0.205127
side validation accuracy: nan %
Epoch 281 of 500 took 0.004s
training loss: 0.005797
(target: 0.003317, side: 0.002480 (absolute))
validation loss: 0.065987
validation accuracy: 97.00 %
side validation loss: 0.205067
side validation accuracy: nan %
Epoch 282 of 500 took 0.004s
training loss: 0.005755
(target: 0.003389, side: 0.002366 (absolute))
validation loss: 0.065734
validation accuracy: 97.00 %
side validation loss: 0.205466
side validation accuracy: nan %
Epoch 283 of 500 took 0.005s
training loss: 0.005850
(target: 0.003387, side: 0.002463 (absolute))
validation loss: 0.064346
validation accuracy: 97.00 %
side validation loss: 0.203526
side validation accuracy: nan %
Epoch 284 of 500 took 0.004s
training loss: 0.005651
(target: 0.003355, side: 0.002296 (absolute))
validation loss: 0.065356
validation accuracy: 97.00 %
side validation loss: 0.203942
side validation accuracy: nan %
Epoch 285 of 500 took 0.004s
training loss: 0.005771
(target: 0.003311, side: 0.002460 (absolute))
validation loss: 0.064323
validation accuracy: 97.00 %
side validation loss: 0.203106
side validation accuracy: nan %
Epoch 286 of 500 took 0.004s
training loss: 0.005791
(target: 0.003401, side: 0.002390 (absolute))
validation loss: 0.064820
validation accuracy: 97.00 %
side validation loss: 0.202011
side validation accuracy: nan %
Epoch 287 of 500 took 0.004s
training loss: 0.005771
(target: 0.003346, side: 0.002425 (absolute))
validation loss: 0.064741
validation accuracy: 98.00 %
side validation loss: 0.203806
side validation accuracy: nan %
Epoch 288 of 500 took 0.004s
training loss: 0.005752
(target: 0.003298, side: 0.002454 (absolute))
validation loss: 0.063682
validation accuracy: 98.00 %
side validation loss: 0.202554
side validation accuracy: nan %
Epoch 289 of 500 took 0.004s
training loss: 0.005767
(target: 0.003248, side: 0.002518 (absolute))
validation loss: 0.064727
validation accuracy: 97.00 %
side validation loss: 0.202846
side validation accuracy: nan %
Epoch 290 of 500 took 0.004s
training loss: 0.005669
(target: 0.003258, side: 0.002411 (absolute))
validation loss: 0.064398
validation accuracy: 97.00 %
side validation loss: 0.202358
side validation accuracy: nan %
Epoch 291 of 500 took 0.004s
training loss: 0.005777
(target: 0.003411, side: 0.002367 (absolute))
validation loss: 0.065405
validation accuracy: 97.00 %
side validation loss: 0.201639
side validation accuracy: nan %
Epoch 292 of 500 took 0.004s
training loss: 0.005667
(target: 0.003205, side: 0.002462 (absolute))
validation loss: 0.065175
validation accuracy: 97.00 %
side validation loss: 0.201391
side validation accuracy: nan %
Epoch 293 of 500 took 0.004s
training loss: 0.005759
(target: 0.003241, side: 0.002518 (absolute))
validation loss: 0.064939
validation accuracy: 97.00 %
side validation loss: 0.201698
side validation accuracy: nan %
Epoch 294 of 500 took 0.004s
training loss: 0.005585
(target: 0.003255, side: 0.002330 (absolute))
validation loss: 0.065095
validation accuracy: 97.00 %
side validation loss: 0.201398
side validation accuracy: nan %
Epoch 295 of 500 took 0.004s
training loss: 0.005679
(target: 0.003254, side: 0.002425 (absolute))
validation loss: 0.064178
validation accuracy: 97.00 %
side validation loss: 0.201473
side validation accuracy: nan %
Epoch 296 of 500 took 0.005s
training loss: 0.005646
(target: 0.003322, side: 0.002324 (absolute))
validation loss: 0.065210
validation accuracy: 98.00 %
side validation loss: 0.201693
side validation accuracy: nan %
Epoch 297 of 500 took 0.004s
training loss: 0.005666
(target: 0.003164, side: 0.002502 (absolute))
validation loss: 0.066385
validation accuracy: 97.00 %
side validation loss: 0.200058
side validation accuracy: nan %
Epoch 298 of 500 took 0.004s
training loss: 0.005551
(target: 0.003247, side: 0.002304 (absolute))
validation loss: 0.065117
validation accuracy: 97.00 %
side validation loss: 0.200076
side validation accuracy: nan %
Epoch 299 of 500 took 0.004s
training loss: 0.005666
(target: 0.003256, side: 0.002410 (absolute))
validation loss: 0.063203
validation accuracy: 98.00 %
side validation loss: 0.199411
side validation accuracy: nan %
Epoch 300 of 500 took 0.004s
training loss: 0.005585
(target: 0.003152, side: 0.002433 (absolute))
validation loss: 0.065426
validation accuracy: 98.00 %
side validation loss: 0.199427
side validation accuracy: nan %
Epoch 301 of 500 took 0.004s
training loss: 0.005481
(target: 0.003226, side: 0.002255 (absolute))
validation loss: 0.066222
validation accuracy: 97.00 %
side validation loss: 0.201277
side validation accuracy: nan %
Epoch 302 of 500 took 0.004s
training loss: 0.005459
(target: 0.003119, side: 0.002341 (absolute))
validation loss: 0.063442
validation accuracy: 98.00 %
side validation loss: 0.199899
side validation accuracy: nan %
Epoch 303 of 500 took 0.004s
training loss: 0.005572
(target: 0.003190, side: 0.002382 (absolute))
validation loss: 0.064178
validation accuracy: 98.00 %
side validation loss: 0.199943
side validation accuracy: nan %
Epoch 304 of 500 took 0.004s
training loss: 0.005661
(target: 0.003101, side: 0.002560 (absolute))
validation loss: 0.065722
validation accuracy: 98.00 %
side validation loss: 0.196673
side validation accuracy: nan %
Epoch 305 of 500 took 0.004s
training loss: 0.005476
(target: 0.003218, side: 0.002258 (absolute))
validation loss: 0.064633
validation accuracy: 97.00 %
side validation loss: 0.197601
side validation accuracy: nan %
Epoch 306 of 500 took 0.004s
training loss: 0.005440
(target: 0.003096, side: 0.002344 (absolute))
validation loss: 0.063555
validation accuracy: 97.00 %
side validation loss: 0.199523
side validation accuracy: nan %
Epoch 307 of 500 took 0.004s
training loss: 0.005429
(target: 0.003147, side: 0.002282 (absolute))
validation loss: 0.065404
validation accuracy: 97.00 %
side validation loss: 0.198376
side validation accuracy: nan %
Epoch 308 of 500 took 0.005s
training loss: 0.005619
(target: 0.003089, side: 0.002530 (absolute))
validation loss: 0.063919
validation accuracy: 98.00 %
side validation loss: 0.196743
side validation accuracy: nan %
Epoch 309 of 500 took 0.004s
training loss: 0.005562
(target: 0.003276, side: 0.002286 (absolute))
validation loss: 0.064751
validation accuracy: 97.00 %
side validation loss: 0.198817
side validation accuracy: nan %
Epoch 310 of 500 took 0.004s
training loss: 0.005382
(target: 0.003045, side: 0.002337 (absolute))
validation loss: 0.064145
validation accuracy: 98.00 %
side validation loss: 0.196599
side validation accuracy: nan %
Epoch 311 of 500 took 0.004s
training loss: 0.005285
(target: 0.003063, side: 0.002222 (absolute))
validation loss: 0.064272
validation accuracy: 97.00 %
side validation loss: 0.196817
side validation accuracy: nan %
Epoch 312 of 500 took 0.004s
training loss: 0.005366
(target: 0.003003, side: 0.002363 (absolute))
validation loss: 0.065413
validation accuracy: 97.00 %
side validation loss: 0.198179
side validation accuracy: nan %
Epoch 313 of 500 took 0.004s
training loss: 0.005349
(target: 0.003185, side: 0.002164 (absolute))
validation loss: 0.065334
validation accuracy: 97.00 %
side validation loss: 0.196359
side validation accuracy: nan %
Epoch 314 of 500 took 0.004s
training loss: 0.005297
(target: 0.003048, side: 0.002249 (absolute))
validation loss: 0.065355
validation accuracy: 97.00 %
side validation loss: 0.196406
side validation accuracy: nan %
Epoch 315 of 500 took 0.004s
training loss: 0.005491
(target: 0.003049, side: 0.002442 (absolute))
validation loss: 0.064548
validation accuracy: 98.00 %
side validation loss: 0.196607
side validation accuracy: nan %
Epoch 316 of 500 took 0.004s
training loss: 0.005317
(target: 0.003023, side: 0.002294 (absolute))
validation loss: 0.065634
validation accuracy: 97.00 %
side validation loss: 0.195278
side validation accuracy: nan %
Epoch 317 of 500 took 0.004s
training loss: 0.005417
(target: 0.003081, side: 0.002336 (absolute))
validation loss: 0.064007
validation accuracy: 97.00 %
side validation loss: 0.195845
side validation accuracy: nan %
Epoch 318 of 500 took 0.004s
training loss: 0.005294
(target: 0.003012, side: 0.002282 (absolute))
validation loss: 0.064230
validation accuracy: 98.00 %
side validation loss: 0.194812
side validation accuracy: nan %
Epoch 319 of 500 took 0.004s
training loss: 0.005287
(target: 0.002964, side: 0.002323 (absolute))
validation loss: 0.063680
validation accuracy: 97.00 %
side validation loss: 0.194536
side validation accuracy: nan %
Epoch 320 of 500 took 0.005s
training loss: 0.005242
(target: 0.003015, side: 0.002227 (absolute))
validation loss: 0.064738
validation accuracy: 98.00 %
side validation loss: 0.195156
side validation accuracy: nan %
Epoch 321 of 500 took 0.004s
training loss: 0.005340
(target: 0.003094, side: 0.002246 (absolute))
validation loss: 0.063971
validation accuracy: 97.00 %
side validation loss: 0.193963
side validation accuracy: nan %
Epoch 322 of 500 took 0.004s
training loss: 0.005292
(target: 0.003006, side: 0.002286 (absolute))
validation loss: 0.063132
validation accuracy: 98.00 %
side validation loss: 0.194442
side validation accuracy: nan %
Epoch 323 of 500 took 0.004s
training loss: 0.005220
(target: 0.002945, side: 0.002275 (absolute))
validation loss: 0.064205
validation accuracy: 98.00 %
side validation loss: 0.194018
side validation accuracy: nan %
Epoch 324 of 500 took 0.004s
training loss: 0.005164
(target: 0.002933, side: 0.002231 (absolute))
validation loss: 0.064342
validation accuracy: 98.00 %
side validation loss: 0.193089
side validation accuracy: nan %
Epoch 325 of 500 took 0.004s
training loss: 0.005191
(target: 0.002956, side: 0.002235 (absolute))
validation loss: 0.064063
validation accuracy: 98.00 %
side validation loss: 0.192757
side validation accuracy: nan %
Epoch 326 of 500 took 0.004s
training loss: 0.005178
(target: 0.003032, side: 0.002146 (absolute))
validation loss: 0.064425
validation accuracy: 98.00 %
side validation loss: 0.193109
side validation accuracy: nan %
Epoch 327 of 500 took 0.004s
training loss: 0.005185
(target: 0.002900, side: 0.002284 (absolute))
validation loss: 0.064431
validation accuracy: 97.00 %
side validation loss: 0.193357
side validation accuracy: nan %
Epoch 328 of 500 took 0.004s
training loss: 0.005122
(target: 0.002919, side: 0.002202 (absolute))
validation loss: 0.063195
validation accuracy: 98.00 %
side validation loss: 0.193855
side validation accuracy: nan %
Epoch 329 of 500 took 0.004s
training loss: 0.005202
(target: 0.002949, side: 0.002253 (absolute))
validation loss: 0.064037
validation accuracy: 97.00 %
side validation loss: 0.191808
side validation accuracy: nan %
Epoch 330 of 500 took 0.004s
training loss: 0.005314
(target: 0.002962, side: 0.002352 (absolute))
validation loss: 0.062352
validation accuracy: 98.00 %
side validation loss: 0.191805
side validation accuracy: nan %
Epoch 331 of 500 took 0.004s
training loss: 0.005235
(target: 0.002928, side: 0.002307 (absolute))
validation loss: 0.062488
validation accuracy: 98.00 %
side validation loss: 0.191259
side validation accuracy: nan %
Epoch 332 of 500 took 0.005s
training loss: 0.005245
(target: 0.002949, side: 0.002296 (absolute))
validation loss: 0.063178
validation accuracy: 98.00 %
side validation loss: 0.190387
side validation accuracy: nan %
Epoch 333 of 500 took 0.005s
training loss: 0.005096
(target: 0.002933, side: 0.002164 (absolute))
validation loss: 0.063618
validation accuracy: 98.00 %
side validation loss: 0.190596
side validation accuracy: nan %
Epoch 334 of 500 took 0.004s
training loss: 0.005101
(target: 0.002818, side: 0.002283 (absolute))
validation loss: 0.063337
validation accuracy: 98.00 %
side validation loss: 0.190542
side validation accuracy: nan %
Epoch 335 of 500 took 0.004s
training loss: 0.005053
(target: 0.002906, side: 0.002147 (absolute))
validation loss: 0.064536
validation accuracy: 98.00 %
side validation loss: 0.190327
side validation accuracy: nan %
Epoch 336 of 500 took 0.004s
training loss: 0.005113
(target: 0.002929, side: 0.002184 (absolute))
validation loss: 0.063368
validation accuracy: 98.00 %
side validation loss: 0.189794
side validation accuracy: nan %
Epoch 337 of 500 took 0.004s
training loss: 0.005090
(target: 0.002862, side: 0.002228 (absolute))
validation loss: 0.063493
validation accuracy: 98.00 %
side validation loss: 0.190647
side validation accuracy: nan %
Epoch 338 of 500 took 0.004s
training loss: 0.005082
(target: 0.002890, side: 0.002192 (absolute))
validation loss: 0.062668
validation accuracy: 98.00 %
side validation loss: 0.188265
side validation accuracy: nan %
Epoch 339 of 500 took 0.004s
training loss: 0.005016
(target: 0.002856, side: 0.002159 (absolute))
validation loss: 0.063378
validation accuracy: 98.00 %
side validation loss: 0.189191
side validation accuracy: nan %
Epoch 340 of 500 took 0.004s
training loss: 0.005002
(target: 0.002861, side: 0.002141 (absolute))
validation loss: 0.063092
validation accuracy: 98.00 %
side validation loss: 0.189687
side validation accuracy: nan %
Epoch 341 of 500 took 0.004s
training loss: 0.005007
(target: 0.002820, side: 0.002187 (absolute))
validation loss: 0.063026
validation accuracy: 98.00 %
side validation loss: 0.188710
side validation accuracy: nan %
Epoch 342 of 500 took 0.004s
training loss: 0.005007
(target: 0.002938, side: 0.002070 (absolute))
validation loss: 0.063449
validation accuracy: 98.00 %
side validation loss: 0.188547
side validation accuracy: nan %
Epoch 343 of 500 took 0.004s
training loss: 0.004938
(target: 0.002762, side: 0.002177 (absolute))
validation loss: 0.062801
validation accuracy: 98.00 %
side validation loss: 0.188711
side validation accuracy: nan %
Epoch 344 of 500 took 0.005s
training loss: 0.004955
(target: 0.002860, side: 0.002095 (absolute))
validation loss: 0.063754
validation accuracy: 98.00 %
side validation loss: 0.187513
side validation accuracy: nan %
Epoch 345 of 500 took 0.005s
training loss: 0.005052
(target: 0.002800, side: 0.002252 (absolute))
validation loss: 0.063286
validation accuracy: 98.00 %
side validation loss: 0.187139
side validation accuracy: nan %
Epoch 346 of 500 took 0.004s
training loss: 0.005115
(target: 0.002850, side: 0.002265 (absolute))
validation loss: 0.065094
validation accuracy: 97.00 %
side validation loss: 0.186801
side validation accuracy: nan %
Epoch 347 of 500 took 0.004s
training loss: 0.004953
(target: 0.002749, side: 0.002203 (absolute))
validation loss: 0.063193
validation accuracy: 98.00 %
side validation loss: 0.188499
side validation accuracy: nan %
Epoch 348 of 500 took 0.004s
training loss: 0.004918
(target: 0.002841, side: 0.002076 (absolute))
validation loss: 0.062194
validation accuracy: 98.00 %
side validation loss: 0.187228
side validation accuracy: nan %
Epoch 349 of 500 took 0.004s
training loss: 0.004877
(target: 0.002809, side: 0.002068 (absolute))
validation loss: 0.063718
validation accuracy: 98.00 %
side validation loss: 0.187804
side validation accuracy: nan %
Epoch 350 of 500 took 0.004s
training loss: 0.004870
(target: 0.002733, side: 0.002138 (absolute))
validation loss: 0.064491
validation accuracy: 97.00 %
side validation loss: 0.186510
side validation accuracy: nan %
Epoch 351 of 500 took 0.004s
training loss: 0.004913
(target: 0.002795, side: 0.002118 (absolute))
validation loss: 0.062353
validation accuracy: 98.00 %
side validation loss: 0.188027
side validation accuracy: nan %
Epoch 352 of 500 took 0.004s
training loss: 0.004804
(target: 0.002793, side: 0.002011 (absolute))
validation loss: 0.063606
validation accuracy: 98.00 %
side validation loss: 0.185435
side validation accuracy: nan %
Epoch 353 of 500 took 0.004s
training loss: 0.004948
(target: 0.002768, side: 0.002179 (absolute))
validation loss: 0.063665
validation accuracy: 98.00 %
side validation loss: 0.185109
side validation accuracy: nan %
Epoch 354 of 500 took 0.004s
training loss: 0.004798
(target: 0.002706, side: 0.002093 (absolute))
validation loss: 0.063025
validation accuracy: 98.00 %
side validation loss: 0.186872
side validation accuracy: nan %
Epoch 355 of 500 took 0.004s
training loss: 0.004885
(target: 0.002735, side: 0.002150 (absolute))
validation loss: 0.062992
validation accuracy: 98.00 %
side validation loss: 0.185678
side validation accuracy: nan %
Epoch 356 of 500 took 0.004s
training loss: 0.004896
(target: 0.002732, side: 0.002164 (absolute))
validation loss: 0.062763
validation accuracy: 98.00 %
side validation loss: 0.184275
side validation accuracy: nan %
Epoch 357 of 500 took 0.005s
training loss: 0.004807
(target: 0.002781, side: 0.002026 (absolute))
validation loss: 0.064147
validation accuracy: 98.00 %
side validation loss: 0.186372
side validation accuracy: nan %
Epoch 358 of 500 took 0.004s
training loss: 0.004791
(target: 0.002666, side: 0.002125 (absolute))
validation loss: 0.062950
validation accuracy: 98.00 %
side validation loss: 0.184200
side validation accuracy: nan %
Epoch 359 of 500 took 0.005s
training loss: 0.004714
(target: 0.002635, side: 0.002078 (absolute))
validation loss: 0.062833
validation accuracy: 98.00 %
side validation loss: 0.185198
side validation accuracy: nan %
Epoch 360 of 500 took 0.004s
training loss: 0.004840
(target: 0.002779, side: 0.002060 (absolute))
validation loss: 0.062383
validation accuracy: 98.00 %
side validation loss: 0.185864
side validation accuracy: nan %
Epoch 361 of 500 took 0.004s
training loss: 0.004796
(target: 0.002702, side: 0.002094 (absolute))
validation loss: 0.063422
validation accuracy: 98.00 %
side validation loss: 0.184131
side validation accuracy: nan %
Epoch 362 of 500 took 0.004s
training loss: 0.004971
(target: 0.002742, side: 0.002229 (absolute))
validation loss: 0.062444
validation accuracy: 98.00 %
side validation loss: 0.183128
side validation accuracy: nan %
Epoch 363 of 500 took 0.004s
training loss: 0.004944
(target: 0.002768, side: 0.002176 (absolute))
validation loss: 0.063718
validation accuracy: 98.00 %
side validation loss: 0.184156
side validation accuracy: nan %
Epoch 364 of 500 took 0.004s
training loss: 0.004816
(target: 0.002697, side: 0.002119 (absolute))
validation loss: 0.062798
validation accuracy: 98.00 %
side validation loss: 0.183564
side validation accuracy: nan %
Epoch 365 of 500 took 0.004s
training loss: 0.004721
(target: 0.002704, side: 0.002017 (absolute))
validation loss: 0.063115
validation accuracy: 98.00 %
side validation loss: 0.183269
side validation accuracy: nan %
Epoch 366 of 500 took 0.004s
training loss: 0.004739
(target: 0.002674, side: 0.002065 (absolute))
validation loss: 0.063018
validation accuracy: 98.00 %
side validation loss: 0.181980
side validation accuracy: nan %
Epoch 367 of 500 took 0.004s
training loss: 0.004769
(target: 0.002617, side: 0.002152 (absolute))
validation loss: 0.062866
validation accuracy: 98.00 %
side validation loss: 0.182685
side validation accuracy: nan %
Epoch 368 of 500 took 0.004s
training loss: 0.004723
(target: 0.002729, side: 0.001994 (absolute))
validation loss: 0.063028
validation accuracy: 98.00 %
side validation loss: 0.182309
side validation accuracy: nan %
Epoch 369 of 500 took 0.005s
training loss: 0.004695
(target: 0.002659, side: 0.002036 (absolute))
validation loss: 0.062893
validation accuracy: 98.00 %
side validation loss: 0.182669
side validation accuracy: nan %
Epoch 370 of 500 took 0.004s
training loss: 0.004733
(target: 0.002628, side: 0.002105 (absolute))
validation loss: 0.063107
validation accuracy: 98.00 %
side validation loss: 0.181808
side validation accuracy: nan %
Epoch 371 of 500 took 0.004s
training loss: 0.004652
(target: 0.002676, side: 0.001975 (absolute))
validation loss: 0.062718
validation accuracy: 98.00 %
side validation loss: 0.183349
side validation accuracy: nan %
Epoch 372 of 500 took 0.004s
training loss: 0.004665
(target: 0.002687, side: 0.001979 (absolute))
validation loss: 0.063138
validation accuracy: 98.00 %
side validation loss: 0.182352
side validation accuracy: nan %
Epoch 373 of 500 took 0.004s
training loss: 0.004681
(target: 0.002543, side: 0.002138 (absolute))
validation loss: 0.062868
validation accuracy: 98.00 %
side validation loss: 0.181372
side validation accuracy: nan %
Epoch 374 of 500 took 0.004s
training loss: 0.004662
(target: 0.002667, side: 0.001995 (absolute))
validation loss: 0.063157
validation accuracy: 98.00 %
side validation loss: 0.182634
side validation accuracy: nan %
Epoch 375 of 500 took 0.004s
training loss: 0.004607
(target: 0.002554, side: 0.002052 (absolute))
validation loss: 0.062329
validation accuracy: 98.00 %
side validation loss: 0.182790
side validation accuracy: nan %
Epoch 376 of 500 took 0.004s
training loss: 0.004639
(target: 0.002624, side: 0.002016 (absolute))
validation loss: 0.062204
validation accuracy: 98.00 %
side validation loss: 0.181096
side validation accuracy: nan %
Epoch 377 of 500 took 0.004s
training loss: 0.004605
(target: 0.002572, side: 0.002033 (absolute))
validation loss: 0.062309
validation accuracy: 98.00 %
side validation loss: 0.181730
side validation accuracy: nan %
Epoch 378 of 500 took 0.004s
training loss: 0.004671
(target: 0.002607, side: 0.002064 (absolute))
validation loss: 0.062818
validation accuracy: 98.00 %
side validation loss: 0.179899
side validation accuracy: nan %
Epoch 379 of 500 took 0.004s
training loss: 0.004634
(target: 0.002598, side: 0.002036 (absolute))
validation loss: 0.063079
validation accuracy: 98.00 %
side validation loss: 0.181205
side validation accuracy: nan %
Epoch 380 of 500 took 0.004s
training loss: 0.004676
(target: 0.002587, side: 0.002089 (absolute))
validation loss: 0.062318
validation accuracy: 98.00 %
side validation loss: 0.180848
side validation accuracy: nan %
Epoch 381 of 500 took 0.005s
training loss: 0.004600
(target: 0.002561, side: 0.002039 (absolute))
validation loss: 0.062612
validation accuracy: 98.00 %
side validation loss: 0.179407
side validation accuracy: nan %
Epoch 382 of 500 took 0.004s
training loss: 0.004565
(target: 0.002573, side: 0.001992 (absolute))
validation loss: 0.061737
validation accuracy: 98.00 %
side validation loss: 0.179728
side validation accuracy: nan %
Epoch 383 of 500 took 0.004s
training loss: 0.004590
(target: 0.002551, side: 0.002039 (absolute))
validation loss: 0.062785
validation accuracy: 98.00 %
side validation loss: 0.180754
side validation accuracy: nan %
Epoch 384 of 500 took 0.004s
training loss: 0.004578
(target: 0.002614, side: 0.001964 (absolute))
validation loss: 0.062162
validation accuracy: 98.00 %
side validation loss: 0.179787
side validation accuracy: nan %
Epoch 385 of 500 took 0.004s
training loss: 0.004565
(target: 0.002575, side: 0.001990 (absolute))
validation loss: 0.061538
validation accuracy: 98.00 %
side validation loss: 0.180025
side validation accuracy: nan %
Epoch 386 of 500 took 0.004s
training loss: 0.004475
(target: 0.002449, side: 0.002026 (absolute))
validation loss: 0.061775
validation accuracy: 98.00 %
side validation loss: 0.179525
side validation accuracy: nan %
Epoch 387 of 500 took 0.004s
training loss: 0.004556
(target: 0.002556, side: 0.001999 (absolute))
validation loss: 0.062876
validation accuracy: 98.00 %
side validation loss: 0.179387
side validation accuracy: nan %
Epoch 388 of 500 took 0.004s
training loss: 0.004545
(target: 0.002599, side: 0.001947 (absolute))
validation loss: 0.062022
validation accuracy: 98.00 %
side validation loss: 0.179144
side validation accuracy: nan %
Epoch 389 of 500 took 0.004s
training loss: 0.004446
(target: 0.002405, side: 0.002041 (absolute))
validation loss: 0.062176
validation accuracy: 98.00 %
side validation loss: 0.178932
side validation accuracy: nan %
Epoch 390 of 500 took 0.004s
training loss: 0.004506
(target: 0.002624, side: 0.001882 (absolute))
validation loss: 0.061755
validation accuracy: 98.00 %
side validation loss: 0.177827
side validation accuracy: nan %
Epoch 391 of 500 took 0.004s
training loss: 0.004527
(target: 0.002463, side: 0.002065 (absolute))
validation loss: 0.061200
validation accuracy: 98.00 %
side validation loss: 0.177675
side validation accuracy: nan %
Epoch 392 of 500 took 0.004s
training loss: 0.004495
(target: 0.002513, side: 0.001982 (absolute))
validation loss: 0.060669
validation accuracy: 98.00 %
side validation loss: 0.178880
side validation accuracy: nan %
Epoch 393 of 500 took 0.005s
training loss: 0.004462
(target: 0.002513, side: 0.001949 (absolute))
validation loss: 0.063002
validation accuracy: 98.00 %
side validation loss: 0.178406
side validation accuracy: nan %
Epoch 394 of 500 took 0.005s
training loss: 0.004564
(target: 0.002498, side: 0.002067 (absolute))
validation loss: 0.060998
validation accuracy: 98.00 %
side validation loss: 0.176296
side validation accuracy: nan %
Epoch 395 of 500 took 0.004s
training loss: 0.004463
(target: 0.002513, side: 0.001950 (absolute))
validation loss: 0.062732
validation accuracy: 98.00 %
side validation loss: 0.175924
side validation accuracy: nan %
Epoch 396 of 500 took 0.004s
training loss: 0.004565
(target: 0.002441, side: 0.002124 (absolute))
validation loss: 0.061264
validation accuracy: 98.00 %
side validation loss: 0.176723
side validation accuracy: nan %
Epoch 397 of 500 took 0.004s
training loss: 0.004414
(target: 0.002469, side: 0.001945 (absolute))
validation loss: 0.062559
validation accuracy: 98.00 %
side validation loss: 0.176822
side validation accuracy: nan %
Epoch 398 of 500 took 0.004s
training loss: 0.004350
(target: 0.002453, side: 0.001896 (absolute))
validation loss: 0.061915
validation accuracy: 98.00 %
side validation loss: 0.176112
side validation accuracy: nan %
Epoch 399 of 500 took 0.004s
training loss: 0.004418
(target: 0.002453, side: 0.001964 (absolute))
validation loss: 0.061656
validation accuracy: 98.00 %
side validation loss: 0.177579
side validation accuracy: nan %
Epoch 400 of 500 took 0.004s
training loss: 0.004432
(target: 0.002476, side: 0.001956 (absolute))
validation loss: 0.060936
validation accuracy: 98.00 %
side validation loss: 0.174240
side validation accuracy: nan %
Epoch 401 of 500 took 0.004s
training loss: 0.004376
(target: 0.002486, side: 0.001890 (absolute))
validation loss: 0.060936
validation accuracy: 98.00 %
side validation loss: 0.175088
side validation accuracy: nan %
Epoch 402 of 500 took 0.004s
training loss: 0.004398
(target: 0.002433, side: 0.001965 (absolute))
validation loss: 0.061738
validation accuracy: 98.00 %
side validation loss: 0.176436
side validation accuracy: nan %
Epoch 403 of 500 took 0.004s
training loss: 0.004518
(target: 0.002514, side: 0.002004 (absolute))
validation loss: 0.061849
validation accuracy: 98.00 %
side validation loss: 0.174797
side validation accuracy: nan %
Epoch 404 of 500 took 0.004s
training loss: 0.004341
(target: 0.002428, side: 0.001914 (absolute))
validation loss: 0.061416
validation accuracy: 98.00 %
side validation loss: 0.175385
side validation accuracy: nan %
Epoch 405 of 500 took 0.004s
training loss: 0.004340
(target: 0.002471, side: 0.001869 (absolute))
validation loss: 0.060859
validation accuracy: 98.00 %
side validation loss: 0.175232
side validation accuracy: nan %
Epoch 406 of 500 took 0.005s
training loss: 0.004381
(target: 0.002431, side: 0.001950 (absolute))
validation loss: 0.061377
validation accuracy: 98.00 %
side validation loss: 0.174764
side validation accuracy: nan %
Epoch 407 of 500 took 0.004s
training loss: 0.004419
(target: 0.002399, side: 0.002020 (absolute))
validation loss: 0.060697
validation accuracy: 98.00 %
side validation loss: 0.173881
side validation accuracy: nan %
Epoch 408 of 500 took 0.004s
training loss: 0.004337
(target: 0.002428, side: 0.001910 (absolute))
validation loss: 0.061702
validation accuracy: 98.00 %
side validation loss: 0.174635
side validation accuracy: nan %
Epoch 409 of 500 took 0.004s
training loss: 0.004364
(target: 0.002489, side: 0.001875 (absolute))
validation loss: 0.062438
validation accuracy: 98.00 %
side validation loss: 0.176253
side validation accuracy: nan %
Epoch 410 of 500 took 0.004s
training loss: 0.004383
(target: 0.002355, side: 0.002028 (absolute))
validation loss: 0.061210
validation accuracy: 98.00 %
side validation loss: 0.174366
side validation accuracy: nan %
Epoch 411 of 500 took 0.004s
training loss: 0.004244
(target: 0.002419, side: 0.001825 (absolute))
validation loss: 0.061288
validation accuracy: 98.00 %
side validation loss: 0.173753
side validation accuracy: nan %
Epoch 412 of 500 took 0.004s
training loss: 0.004298
(target: 0.002410, side: 0.001887 (absolute))
validation loss: 0.061107
validation accuracy: 98.00 %
side validation loss: 0.173859
side validation accuracy: nan %
Epoch 413 of 500 took 0.004s
training loss: 0.004337
(target: 0.002440, side: 0.001897 (absolute))
validation loss: 0.060366
validation accuracy: 98.00 %
side validation loss: 0.174240
side validation accuracy: nan %
Epoch 414 of 500 took 0.004s
training loss: 0.004219
(target: 0.002312, side: 0.001907 (absolute))
validation loss: 0.060344
validation accuracy: 98.00 %
side validation loss: 0.173675
side validation accuracy: nan %
Epoch 415 of 500 took 0.004s
training loss: 0.004299
(target: 0.002445, side: 0.001854 (absolute))
validation loss: 0.061754
validation accuracy: 98.00 %
side validation loss: 0.172855
side validation accuracy: nan %
Epoch 416 of 500 took 0.004s
training loss: 0.004384
(target: 0.002354, side: 0.002030 (absolute))
validation loss: 0.061160
validation accuracy: 98.00 %
side validation loss: 0.171671
side validation accuracy: nan %
Epoch 417 of 500 took 0.004s
training loss: 0.004319
(target: 0.002442, side: 0.001877 (absolute))
validation loss: 0.061711
validation accuracy: 98.00 %
side validation loss: 0.174183
side validation accuracy: nan %
Epoch 418 of 500 took 0.005s
training loss: 0.004525
(target: 0.002367, side: 0.002158 (absolute))
validation loss: 0.060718
validation accuracy: 98.00 %
side validation loss: 0.174133
side validation accuracy: nan %
Epoch 419 of 500 took 0.004s
training loss: 0.004385
(target: 0.002346, side: 0.002039 (absolute))
validation loss: 0.061210
validation accuracy: 98.00 %
side validation loss: 0.171397
side validation accuracy: nan %
Epoch 420 of 500 took 0.004s
training loss: 0.004321
(target: 0.002387, side: 0.001934 (absolute))
validation loss: 0.061792
validation accuracy: 98.00 %
side validation loss: 0.170418
side validation accuracy: nan %
Epoch 421 of 500 took 0.004s
training loss: 0.004311
(target: 0.002353, side: 0.001959 (absolute))
validation loss: 0.061015
validation accuracy: 98.00 %
side validation loss: 0.171339
side validation accuracy: nan %
Epoch 422 of 500 took 0.004s
training loss: 0.004214
(target: 0.002401, side: 0.001812 (absolute))
validation loss: 0.061855
validation accuracy: 98.00 %
side validation loss: 0.171814
side validation accuracy: nan %
Epoch 423 of 500 took 0.004s
training loss: 0.004208
(target: 0.002329, side: 0.001879 (absolute))
validation loss: 0.060544
validation accuracy: 98.00 %
side validation loss: 0.172246
side validation accuracy: nan %
Epoch 424 of 500 took 0.004s
training loss: 0.004180
(target: 0.002317, side: 0.001863 (absolute))
validation loss: 0.062089
validation accuracy: 98.00 %
side validation loss: 0.170819
side validation accuracy: nan %
Epoch 425 of 500 took 0.004s
training loss: 0.004259
(target: 0.002371, side: 0.001887 (absolute))
validation loss: 0.061313
validation accuracy: 98.00 %
side validation loss: 0.172253
side validation accuracy: nan %
Epoch 426 of 500 took 0.004s
training loss: 0.004258
(target: 0.002347, side: 0.001912 (absolute))
validation loss: 0.060282
validation accuracy: 98.00 %
side validation loss: 0.170462
side validation accuracy: nan %
Epoch 427 of 500 took 0.004s
training loss: 0.004191
(target: 0.002360, side: 0.001831 (absolute))
validation loss: 0.061394
validation accuracy: 98.00 %
side validation loss: 0.171625
side validation accuracy: nan %
Epoch 428 of 500 took 0.004s
training loss: 0.004203
(target: 0.002304, side: 0.001900 (absolute))
validation loss: 0.061194
validation accuracy: 98.00 %
side validation loss: 0.171293
side validation accuracy: nan %
Epoch 429 of 500 took 0.004s
training loss: 0.004143
(target: 0.002347, side: 0.001797 (absolute))
validation loss: 0.060522
validation accuracy: 98.00 %
side validation loss: 0.170297
side validation accuracy: nan %
Epoch 430 of 500 took 0.005s
training loss: 0.004098
(target: 0.002254, side: 0.001845 (absolute))
validation loss: 0.061389
validation accuracy: 98.00 %
side validation loss: 0.171096
side validation accuracy: nan %
Epoch 431 of 500 took 0.004s
training loss: 0.004262
(target: 0.002303, side: 0.001959 (absolute))
validation loss: 0.059968
validation accuracy: 98.00 %
side validation loss: 0.169753
side validation accuracy: nan %
Epoch 432 of 500 took 0.004s
training loss: 0.004127
(target: 0.002337, side: 0.001791 (absolute))
validation loss: 0.060652
validation accuracy: 98.00 %
side validation loss: 0.169754
side validation accuracy: nan %
Epoch 433 of 500 took 0.004s
training loss: 0.004027
(target: 0.002303, side: 0.001725 (absolute))
validation loss: 0.059661
validation accuracy: 98.00 %
side validation loss: 0.170902
side validation accuracy: nan %
Epoch 434 of 500 took 0.004s
training loss: 0.004149
(target: 0.002296, side: 0.001853 (absolute))
validation loss: 0.061232
validation accuracy: 98.00 %
side validation loss: 0.170218
side validation accuracy: nan %
Epoch 435 of 500 took 0.004s
training loss: 0.004215
(target: 0.002355, side: 0.001859 (absolute))
validation loss: 0.059981
validation accuracy: 98.00 %
side validation loss: 0.168863
side validation accuracy: nan %
Epoch 436 of 500 took 0.004s
training loss: 0.004261
(target: 0.002276, side: 0.001986 (absolute))
validation loss: 0.059967
validation accuracy: 98.00 %
side validation loss: 0.169996
side validation accuracy: nan %
Epoch 437 of 500 took 0.004s
training loss: 0.004129
(target: 0.002276, side: 0.001853 (absolute))
validation loss: 0.060520
validation accuracy: 98.00 %
side validation loss: 0.169564
side validation accuracy: nan %
Epoch 438 of 500 took 0.004s
training loss: 0.004140
(target: 0.002339, side: 0.001802 (absolute))
validation loss: 0.061498
validation accuracy: 98.00 %
side validation loss: 0.169800
side validation accuracy: nan %
Epoch 439 of 500 took 0.004s
training loss: 0.004079
(target: 0.002258, side: 0.001821 (absolute))
validation loss: 0.060409
validation accuracy: 98.00 %
side validation loss: 0.169638
side validation accuracy: nan %
Epoch 440 of 500 took 0.004s
training loss: 0.004111
(target: 0.002262, side: 0.001849 (absolute))
validation loss: 0.060120
validation accuracy: 98.00 %
side validation loss: 0.169170
side validation accuracy: nan %
Epoch 441 of 500 took 0.004s
training loss: 0.004102
(target: 0.002271, side: 0.001831 (absolute))
validation loss: 0.060521
validation accuracy: 98.00 %
side validation loss: 0.167756
side validation accuracy: nan %
Epoch 442 of 500 took 0.005s
training loss: 0.004080
(target: 0.002210, side: 0.001870 (absolute))
validation loss: 0.060509
validation accuracy: 98.00 %
side validation loss: 0.167760
side validation accuracy: nan %
Epoch 443 of 500 took 0.005s
training loss: 0.004088
(target: 0.002298, side: 0.001790 (absolute))
validation loss: 0.061653
validation accuracy: 98.00 %
side validation loss: 0.168491
side validation accuracy: nan %
Epoch 444 of 500 took 0.004s
training loss: 0.004081
(target: 0.002255, side: 0.001826 (absolute))
validation loss: 0.060446
validation accuracy: 98.00 %
side validation loss: 0.168669
side validation accuracy: nan %
Epoch 445 of 500 took 0.004s
training loss: 0.004040
(target: 0.002234, side: 0.001806 (absolute))
validation loss: 0.060230
validation accuracy: 98.00 %
side validation loss: 0.167596
side validation accuracy: nan %
Epoch 446 of 500 took 0.004s
training loss: 0.004040
(target: 0.002209, side: 0.001831 (absolute))
validation loss: 0.061327
validation accuracy: 98.00 %
side validation loss: 0.168032
side validation accuracy: nan %
Epoch 447 of 500 took 0.004s
training loss: 0.004038
(target: 0.002294, side: 0.001744 (absolute))
validation loss: 0.060619
validation accuracy: 98.00 %
side validation loss: 0.167871
side validation accuracy: nan %
Epoch 448 of 500 took 0.004s
training loss: 0.004040
(target: 0.002165, side: 0.001876 (absolute))
validation loss: 0.060814
validation accuracy: 98.00 %
side validation loss: 0.168016
side validation accuracy: nan %
Epoch 449 of 500 took 0.004s
training loss: 0.004059
(target: 0.002256, side: 0.001803 (absolute))
validation loss: 0.060901
validation accuracy: 98.00 %
side validation loss: 0.167709
side validation accuracy: nan %
Epoch 450 of 500 took 0.004s
training loss: 0.004147
(target: 0.002196, side: 0.001952 (absolute))
validation loss: 0.060105
validation accuracy: 98.00 %
side validation loss: 0.166511
side validation accuracy: nan %
Epoch 451 of 500 took 0.004s
training loss: 0.004099
(target: 0.002235, side: 0.001864 (absolute))
validation loss: 0.059999
validation accuracy: 98.00 %
side validation loss: 0.167060
side validation accuracy: nan %
Epoch 452 of 500 took 0.004s
training loss: 0.004052
(target: 0.002242, side: 0.001810 (absolute))
validation loss: 0.060346
validation accuracy: 98.00 %
side validation loss: 0.166814
side validation accuracy: nan %
Epoch 453 of 500 took 0.004s
training loss: 0.004001
(target: 0.002265, side: 0.001736 (absolute))
validation loss: 0.060628
validation accuracy: 98.00 %
side validation loss: 0.166915
side validation accuracy: nan %
Epoch 454 of 500 took 0.004s
training loss: 0.004032
(target: 0.002178, side: 0.001854 (absolute))
validation loss: 0.060533
validation accuracy: 98.00 %
side validation loss: 0.166669
side validation accuracy: nan %
Epoch 455 of 500 took 0.005s
training loss: 0.004023
(target: 0.002233, side: 0.001790 (absolute))
validation loss: 0.059262
validation accuracy: 98.00 %
side validation loss: 0.166895
side validation accuracy: nan %
Epoch 456 of 500 took 0.004s
training loss: 0.004090
(target: 0.002233, side: 0.001857 (absolute))
validation loss: 0.060169
validation accuracy: 98.00 %
side validation loss: 0.166143
side validation accuracy: nan %
Epoch 457 of 500 took 0.004s
training loss: 0.004102
(target: 0.002216, side: 0.001886 (absolute))
validation loss: 0.060401
validation accuracy: 98.00 %
side validation loss: 0.166845
side validation accuracy: nan %
Epoch 458 of 500 took 0.004s
training loss: 0.004064
(target: 0.002207, side: 0.001857 (absolute))
validation loss: 0.060651
validation accuracy: 98.00 %
side validation loss: 0.166166
side validation accuracy: nan %
Epoch 459 of 500 took 0.004s
training loss: 0.003961
(target: 0.002195, side: 0.001766 (absolute))
validation loss: 0.060941
validation accuracy: 98.00 %
side validation loss: 0.166128
side validation accuracy: nan %
Epoch 460 of 500 took 0.004s
training loss: 0.004164
(target: 0.002205, side: 0.001959 (absolute))
validation loss: 0.061264
validation accuracy: 98.00 %
side validation loss: 0.165625
side validation accuracy: nan %
Epoch 461 of 500 took 0.004s
training loss: 0.004053
(target: 0.002223, side: 0.001831 (absolute))
validation loss: 0.060561
validation accuracy: 98.00 %
side validation loss: 0.165219
side validation accuracy: nan %
Epoch 462 of 500 took 0.004s
training loss: 0.003975
(target: 0.002158, side: 0.001817 (absolute))
validation loss: 0.060430
validation accuracy: 98.00 %
side validation loss: 0.164982
side validation accuracy: nan %
Epoch 463 of 500 took 0.004s
training loss: 0.003918
(target: 0.002201, side: 0.001717 (absolute))
validation loss: 0.059591
validation accuracy: 98.00 %
side validation loss: 0.164420
side validation accuracy: nan %
Epoch 464 of 500 took 0.004s
training loss: 0.003959
(target: 0.002131, side: 0.001829 (absolute))
validation loss: 0.058754
validation accuracy: 98.00 %
side validation loss: 0.165686
side validation accuracy: nan %
Epoch 465 of 500 took 0.004s
training loss: 0.003923
(target: 0.002215, side: 0.001708 (absolute))
validation loss: 0.060121
validation accuracy: 98.00 %
side validation loss: 0.165166
side validation accuracy: nan %
Epoch 466 of 500 took 0.004s
training loss: 0.003892
(target: 0.002152, side: 0.001740 (absolute))
validation loss: 0.058452
validation accuracy: 98.00 %
side validation loss: 0.165395
side validation accuracy: nan %
Epoch 467 of 500 took 0.005s
training loss: 0.003971
(target: 0.002163, side: 0.001808 (absolute))
validation loss: 0.059714
validation accuracy: 98.00 %
side validation loss: 0.163596
side validation accuracy: nan %
Epoch 468 of 500 took 0.004s
training loss: 0.003935
(target: 0.002130, side: 0.001805 (absolute))
validation loss: 0.059751
validation accuracy: 98.00 %
side validation loss: 0.164066
side validation accuracy: nan %
Epoch 469 of 500 took 0.004s
training loss: 0.003930
(target: 0.002164, side: 0.001767 (absolute))
validation loss: 0.059232
validation accuracy: 98.00 %
side validation loss: 0.163732
side validation accuracy: nan %
Epoch 470 of 500 took 0.004s
training loss: 0.003915
(target: 0.002156, side: 0.001759 (absolute))
validation loss: 0.059717
validation accuracy: 98.00 %
side validation loss: 0.165217
side validation accuracy: nan %
Epoch 471 of 500 took 0.004s
training loss: 0.003892
(target: 0.002100, side: 0.001792 (absolute))
validation loss: 0.059991
validation accuracy: 98.00 %
side validation loss: 0.163301
side validation accuracy: nan %
Epoch 472 of 500 took 0.004s
training loss: 0.003920
(target: 0.002168, side: 0.001752 (absolute))
validation loss: 0.059792
validation accuracy: 98.00 %
side validation loss: 0.163499
side validation accuracy: nan %
Epoch 473 of 500 took 0.004s
training loss: 0.003982
(target: 0.002092, side: 0.001890 (absolute))
validation loss: 0.058418
validation accuracy: 98.00 %
side validation loss: 0.164114
side validation accuracy: nan %
Epoch 474 of 500 took 0.004s
training loss: 0.003920
(target: 0.002174, side: 0.001746 (absolute))
validation loss: 0.061460
validation accuracy: 98.00 %
side validation loss: 0.163353
side validation accuracy: nan %
Epoch 475 of 500 took 0.004s
training loss: 0.003881
(target: 0.002143, side: 0.001738 (absolute))
validation loss: 0.059588
validation accuracy: 98.00 %
side validation loss: 0.163339
side validation accuracy: nan %
Epoch 476 of 500 took 0.004s
training loss: 0.003990
(target: 0.002052, side: 0.001938 (absolute))
validation loss: 0.058967
validation accuracy: 98.00 %
side validation loss: 0.162110
side validation accuracy: nan %
Epoch 477 of 500 took 0.004s
training loss: 0.003943
(target: 0.002212, side: 0.001731 (absolute))
validation loss: 0.059616
validation accuracy: 98.00 %
side validation loss: 0.164123
side validation accuracy: nan %
Epoch 478 of 500 took 0.004s
training loss: 0.003909
(target: 0.002036, side: 0.001873 (absolute))
validation loss: 0.059080
validation accuracy: 98.00 %
side validation loss: 0.162756
side validation accuracy: nan %
Epoch 479 of 500 took 0.005s
training loss: 0.004005
(target: 0.002171, side: 0.001834 (absolute))
validation loss: 0.059042
validation accuracy: 98.00 %
side validation loss: 0.160957
side validation accuracy: nan %
Epoch 480 of 500 took 0.005s
training loss: 0.003820
(target: 0.002080, side: 0.001740 (absolute))
validation loss: 0.059125
validation accuracy: 98.00 %
side validation loss: 0.162545
side validation accuracy: nan %
Epoch 481 of 500 took 0.004s
training loss: 0.003734
(target: 0.002105, side: 0.001629 (absolute))
validation loss: 0.059301
validation accuracy: 98.00 %
side validation loss: 0.162071
side validation accuracy: nan %
Epoch 482 of 500 took 0.004s
training loss: 0.003924
(target: 0.002076, side: 0.001849 (absolute))
validation loss: 0.058934
validation accuracy: 98.00 %
side validation loss: 0.162197
side validation accuracy: nan %
Epoch 483 of 500 took 0.004s
training loss: 0.003788
(target: 0.002121, side: 0.001667 (absolute))
validation loss: 0.058786
validation accuracy: 98.00 %
side validation loss: 0.161627
side validation accuracy: nan %
Epoch 484 of 500 took 0.004s
training loss: 0.003791
(target: 0.002075, side: 0.001716 (absolute))
validation loss: 0.058737
validation accuracy: 98.00 %
side validation loss: 0.161628
side validation accuracy: nan %
Epoch 485 of 500 took 0.004s
training loss: 0.003773
(target: 0.002062, side: 0.001712 (absolute))
validation loss: 0.059580
validation accuracy: 98.00 %
side validation loss: 0.160740
side validation accuracy: nan %
Epoch 486 of 500 took 0.004s
training loss: 0.003838
(target: 0.002144, side: 0.001694 (absolute))
validation loss: 0.058535
validation accuracy: 98.00 %
side validation loss: 0.162468
side validation accuracy: nan %
Epoch 487 of 500 took 0.004s
training loss: 0.003801
(target: 0.002082, side: 0.001719 (absolute))
validation loss: 0.059511
validation accuracy: 98.00 %
side validation loss: 0.161823
side validation accuracy: nan %
Epoch 488 of 500 took 0.004s
training loss: 0.003779
(target: 0.002068, side: 0.001711 (absolute))
validation loss: 0.058275
validation accuracy: 98.00 %
side validation loss: 0.161468
side validation accuracy: nan %
Epoch 489 of 500 took 0.004s
training loss: 0.003811
(target: 0.002075, side: 0.001736 (absolute))
validation loss: 0.058410
validation accuracy: 98.00 %
side validation loss: 0.160279
side validation accuracy: nan %
Epoch 490 of 500 took 0.004s
training loss: 0.003797
(target: 0.002128, side: 0.001668 (absolute))
validation loss: 0.059112
validation accuracy: 98.00 %
side validation loss: 0.161064
side validation accuracy: nan %
Epoch 491 of 500 took 0.004s
training loss: 0.003794
(target: 0.002002, side: 0.001793 (absolute))
validation loss: 0.058642
validation accuracy: 98.00 %
side validation loss: 0.160398
side validation accuracy: nan %
Epoch 492 of 500 took 0.005s
training loss: 0.003819
(target: 0.002112, side: 0.001707 (absolute))
validation loss: 0.058983
validation accuracy: 98.00 %
side validation loss: 0.160900
side validation accuracy: nan %
Epoch 493 of 500 took 0.004s
training loss: 0.003808
(target: 0.002038, side: 0.001770 (absolute))
validation loss: 0.058744
validation accuracy: 98.00 %
side validation loss: 0.160203
side validation accuracy: nan %
Epoch 494 of 500 took 0.004s
training loss: 0.003737
(target: 0.002055, side: 0.001682 (absolute))
validation loss: 0.060337
validation accuracy: 98.00 %
side validation loss: 0.161312
side validation accuracy: nan %
Epoch 495 of 500 took 0.004s
training loss: 0.003763
(target: 0.001984, side: 0.001778 (absolute))
validation loss: 0.058235
validation accuracy: 98.00 %
side validation loss: 0.160249
side validation accuracy: nan %
Epoch 496 of 500 took 0.004s
training loss: 0.003748
(target: 0.002076, side: 0.001673 (absolute))
validation loss: 0.059169
validation accuracy: 98.00 %
side validation loss: 0.160728
side validation accuracy: nan %
Epoch 497 of 500 took 0.004s
training loss: 0.003815
(target: 0.002027, side: 0.001788 (absolute))
validation loss: 0.058794
validation accuracy: 98.00 %
side validation loss: 0.159533
side validation accuracy: nan %
Epoch 498 of 500 took 0.004s
training loss: 0.003759
(target: 0.002073, side: 0.001686 (absolute))
validation loss: 0.058611
validation accuracy: 98.00 %
side validation loss: 0.160658
side validation accuracy: nan %
Epoch 499 of 500 took 0.004s
training loss: 0.003715
(target: 0.001953, side: 0.001762 (absolute))
validation loss: 0.059512
validation accuracy: 98.00 %
side validation loss: 0.159068
side validation accuracy: nan %
Epoch 500 of 500 took 0.004s
training loss: 0.003780
(target: 0.002110, side: 0.001670 (absolute))
validation loss: 0.058076
validation accuracy: 98.00 %
side validation loss: 0.158164
side validation accuracy: nan %
Some statistics: Test score.
In [32]:
trainer.score(X_test, y_test, verbose=True)
pass
Score:
loss: 0.029269
accuracy: 99.00 %
We can also compute a test score for the side loss:
In [33]:
trainer.score_side([X_test, Z_test], verbose=True)
pass
Score for side:
loss: 0.127295
accuracy: nan %
You can then also query the prediction output, similar to the scikit-learn API:
In [34]:
trainer.predict(X_test)
Out[34]:
array([0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0,
1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 0, 1, 0,
0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1,
0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1,
0, 1, 0, 1, 0, 1, 1, 0])
Content source: tu-rbo/concarne
Similar notebooks: