In [1]:
import random
import numpy as np
import random
import keras
from sklearn.preprocessing import StandardScaler
import tensorflow as tf
from keras import models
from keras import layers
import matplotlib.pyplot as plt
%matplotlib inline
Using TensorFlow backend.
In [2]:
# number of characters in a word.
# for instance abccba has nb_chars = 6
nb_chars = 5
# number of possible characters used during the encoding.
# for instance abcde leads to 01234 has nb_letters = 5
nb_letters = 26
# number of words samples to be generated
nb_words = 10000
# percentage of words that will be used for validation
percentage_split = 0.60
# number of epochs for fitting the model training step
nb_epochs = 200
In [3]:
# total number of combinations
nb_letters**nb_chars
Out[3]:
11881376
In [4]:
def create_inputs(nb_words, nb_chars, nb_letters):
'''Create a numpy array of nb_words rows with nb_chars columns each element
being a random letter of nb_letters (a, b...)'''
words = np.zeros((nb_words, nb_chars), dtype=int)
for w in range(nb_words):
optim_tentative = False
if optim_tentative == True and w%10 != 0:
i = random.randint(0, nb_letters-1)
for c in range(nb_chars):
words[w, c] = ord('a') + i
else:
for c in range(nb_chars):
i = random.randint(0, nb_letters-1)
words[w, c] = ord('a') + i
return words
def encrypt(words, nb_words, nb_chars):
'''Encrypt each element of a numpy array of nb_words rows with nb_chars
columns each item with a secret algorithm'''
encrypted_words = words.copy()
encrypted_words_probs = np.zeros((nb_words, nb_chars, nb_chars))
#val_max = -1
for w in range(nb_words):
for c in range(nb_chars): # 0,1,2,3,4
encrypted_words[w,c] = int(words[w,c]) - 49
val = encrypted_words[w,c] - 48
#if val > val_max:
# val_max = val
# add entropy (i.e. mistakes in the encryption)
#epsilon = random.randint(0, 100)
#if epsilon == 5 and val != val_max:
#val +=1
#print('w:',w,', c:',c,', [wc]:', val)
#encrypted_words_probs[w, c, val ] = 1.0
encrypted_words[w,c] = val
return encrypted_words
In [5]:
nb_features = nb_chars
# This returns a tensor
inputs = layers.Input(shape=(nb_chars,), dtype='float32', name='main_input')
# a layer instance is callable on a tensor, and returns a tensor
x = layers.Dense(4096, activation='relu', name='hl_1')(inputs)
#x = layers.Dense(2048, activation='relu', name='hl_1')(inputs)
#x = layers.Dense(64, activation='relu', name='hl_2')(x)
outputs = []
losses = {}
for o in range(nb_chars):
name_i = 'output_'+str(o)
output_i = layers.Dense(nb_letters, activation='softmax', dtype='float32', name=name_i)(x)
outputs.append(output_i)
losses[name_i] = 'categorical_crossentropy'
coding_model = keras.models.Model(inputs=inputs, outputs=outputs)
coding_model.compile(optimizer='rmsprop',
loss=losses,
metrics=['accuracy'])
In [6]:
print(coding_model.summary())
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
main_input (InputLayer) (None, 5) 0
__________________________________________________________________________________________________
hl_1 (Dense) (None, 4096) 24576 main_input[0][0]
__________________________________________________________________________________________________
output_0 (Dense) (None, 26) 106522 hl_1[0][0]
__________________________________________________________________________________________________
output_1 (Dense) (None, 26) 106522 hl_1[0][0]
__________________________________________________________________________________________________
output_2 (Dense) (None, 26) 106522 hl_1[0][0]
__________________________________________________________________________________________________
output_3 (Dense) (None, 26) 106522 hl_1[0][0]
__________________________________________________________________________________________________
output_4 (Dense) (None, 26) 106522 hl_1[0][0]
==================================================================================================
Total params: 557,186
Trainable params: 557,186
Non-trainable params: 0
__________________________________________________________________________________________________
None
In [7]:
def print_readable_inputs(x):
words = []
for w in x:
word = ''
for c in w:
word += chr(c)
words.append(word)
print(words)
In [8]:
def print_readable_outputs_(outputs, nb_words, nb_chars):
# outputs are listed : first, per char, second by sample, third by letter probability
words = [''] * nb_words
c_i = 0
for char in outputs:
s_i = 0
for sample in char:
l_i = 0
best_value = -float('inf')
best_letter = -1
for letter_probs in sample:
if letter_probs > best_value:
best_value = letter_probs
best_letter = l_i
l_i += 1
words[s_i] += str(best_letter)
if c_i != nb_chars - 1:
words[s_i] += ' '
s_i += 1
c_i += 1
print(words)
In [9]:
def print_readable_outputs(outputs, nb_words, nb_chars):
# outputs are listed : first, per char, second by sample, third by letter probability
words = [''] * nb_words
c_i = 0
for char in outputs:
s_i = 0
for sample in char:
best_letter = np.argmax(sample)
words[s_i] += str(best_letter)
if c_i != nb_chars - 1:
words[s_i] += ' '
s_i += 1
c_i += 1
print(words)
In [10]:
x = create_inputs(nb_words, nb_chars, nb_letters)
print('x: (as readable inputs)')
first_n_samples = 4
print_readable_inputs(x[:first_n_samples])
print('x (partial):\n', x[:first_n_samples], 'out of ',len(x))
print()
# process the x data as useful ANN input data
scaler = StandardScaler()
x_train = scaler.fit_transform(x)
print('x_train:\n', x_train[:first_n_samples], 'out of ',len(x_train))
print()
# create output data for training
y = encrypt(x, nb_words, nb_chars)
print('y (readable):\n', y)
print()
# process the y data as useful ANN output data
y_train0 = keras.utils.to_categorical(y, nb_letters)
print('y (less readable):\n', y_train0[:first_n_samples], 'out of ',len(y_train0))
print('')
# process the y data as useful ANN multiple-outputs data
y_train = []
for c in range(nb_chars):
# extract each 'char' colomn from the global y_train0 tensor
# in order to have multiplue yi_train outputs tensors
yi_train = y_train0[:,c,:]
y_train.append(yi_train)
# Not really displayable, henced commented
#print('y_train):')
#print(y_train[:first_n_samples])
x: (as readable inputs)
['kfzlc', 'piusm', 'tsbil', 'hiixq']
x (partial):
[[107 102 122 108 99]
[112 105 117 115 109]
[116 115 98 105 108]
[104 105 105 120 113]] out of 10000
x_train:
[[-0.32792525 -0.99875678 1.65344344 -0.20652286 -1.397286 ]
[ 0.33959348 -0.60076114 0.98416862 0.72960357 -0.0735971 ]
[ 0.87360846 0.72589097 -1.55907569 -0.6077199 -0.20596599]
[-0.72843648 -0.60076114 -0.62209094 1.3982653 0.45587846]] out of 10000
y (readable):
[[10 5 25 11 2]
[15 8 20 18 12]
[19 18 1 8 11]
...
[16 15 7 17 9]
[12 4 10 11 14]
[ 4 23 16 9 17]]
y (less readable):
[[[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 1.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]]
[[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]]
[[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.
0. 0. 0.]
[0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]]
[[0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
1. 0. 0.]
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.
0. 0. 0.]]] out of 10000
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:595: DataConversionWarning: Data with input dtype int64 was converted to float64 by StandardScaler.
warnings.warn(msg, DataConversionWarning)
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:595: DataConversionWarning: Data with input dtype int64 was converted to float64 by StandardScaler.
warnings.warn(msg, DataConversionWarning)
In [11]:
history = coding_model.fit(x_train, y_train, validation_split=percentage_split, batch_size=32, epochs=nb_epochs, verbose=1)
Train on 4000 samples, validate on 6000 samples
Epoch 1/200
4000/4000 [==============================] - 1s 341us/step - loss: 13.1881 - output_0_loss: 2.6374 - output_1_loss: 2.6306 - output_2_loss: 2.6425 - output_3_loss: 2.6458 - output_4_loss: 2.6317 - output_0_acc: 0.1438 - output_1_acc: 0.1390 - output_2_acc: 0.1408 - output_3_acc: 0.1397 - output_4_acc: 0.1490 - val_loss: 11.7739 - val_output_0_loss: 2.3508 - val_output_1_loss: 2.3466 - val_output_2_loss: 2.3722 - val_output_3_loss: 2.3417 - val_output_4_loss: 2.3627 - val_output_0_acc: 0.1880 - val_output_1_acc: 0.1722 - val_output_2_acc: 0.1737 - val_output_3_acc: 0.1868 - val_output_4_acc: 0.1720
Epoch 2/200
4000/4000 [==============================] - 1s 216us/step - loss: 11.0172 - output_0_loss: 2.2014 - output_1_loss: 2.2061 - output_2_loss: 2.2083 - output_3_loss: 2.2110 - output_4_loss: 2.1903 - output_0_acc: 0.2245 - output_1_acc: 0.2102 - output_2_acc: 0.2142 - output_3_acc: 0.2093 - output_4_acc: 0.2198 - val_loss: 10.6746 - val_output_0_loss: 2.1230 - val_output_1_loss: 2.1451 - val_output_2_loss: 2.1326 - val_output_3_loss: 2.1587 - val_output_4_loss: 2.1151 - val_output_0_acc: 0.2158 - val_output_1_acc: 0.2137 - val_output_2_acc: 0.2157 - val_output_3_acc: 0.2123 - val_output_4_acc: 0.2220
Epoch 3/200
4000/4000 [==============================] - 1s 180us/step - loss: 10.1447 - output_0_loss: 2.0223 - output_1_loss: 2.0348 - output_2_loss: 2.0366 - output_3_loss: 2.0369 - output_4_loss: 2.0141 - output_0_acc: 0.2642 - output_1_acc: 0.2485 - output_2_acc: 0.2420 - output_3_acc: 0.2442 - output_4_acc: 0.2595 - val_loss: 9.9794 - val_output_0_loss: 2.0168 - val_output_1_loss: 1.9941 - val_output_2_loss: 1.9901 - val_output_3_loss: 1.9839 - val_output_4_loss: 1.9945 - val_output_0_acc: 0.2408 - val_output_1_acc: 0.2467 - val_output_2_acc: 0.2593 - val_output_3_acc: 0.2445 - val_output_4_acc: 0.2395
Epoch 4/200
4000/4000 [==============================] - 1s 180us/step - loss: 9.5551 - output_0_loss: 1.9083 - output_1_loss: 1.9260 - output_2_loss: 1.9085 - output_3_loss: 1.9152 - output_4_loss: 1.8971 - output_0_acc: 0.2830 - output_1_acc: 0.2700 - output_2_acc: 0.2883 - output_3_acc: 0.2767 - output_4_acc: 0.2830 - val_loss: 9.5679 - val_output_0_loss: 1.9372 - val_output_1_loss: 1.8968 - val_output_2_loss: 1.9127 - val_output_3_loss: 1.9376 - val_output_4_loss: 1.8837 - val_output_0_acc: 0.2438 - val_output_1_acc: 0.2595 - val_output_2_acc: 0.2568 - val_output_3_acc: 0.2698 - val_output_4_acc: 0.2802
Epoch 5/200
4000/4000 [==============================] - 1s 183us/step - loss: 9.0907 - output_0_loss: 1.8222 - output_1_loss: 1.8300 - output_2_loss: 1.8177 - output_3_loss: 1.8229 - output_4_loss: 1.7978 - output_0_acc: 0.3088 - output_1_acc: 0.3017 - output_2_acc: 0.3095 - output_3_acc: 0.3010 - output_4_acc: 0.3230 - val_loss: 9.1703 - val_output_0_loss: 1.8181 - val_output_1_loss: 1.8288 - val_output_2_loss: 1.8941 - val_output_3_loss: 1.7993 - val_output_4_loss: 1.8300 - val_output_0_acc: 0.2942 - val_output_1_acc: 0.2728 - val_output_2_acc: 0.2607 - val_output_3_acc: 0.2905 - val_output_4_acc: 0.2843
Epoch 6/200
4000/4000 [==============================] - 1s 188us/step - loss: 8.6595 - output_0_loss: 1.7203 - output_1_loss: 1.7490 - output_2_loss: 1.7375 - output_3_loss: 1.7334 - output_4_loss: 1.7192 - output_0_acc: 0.3392 - output_1_acc: 0.3188 - output_2_acc: 0.3280 - output_3_acc: 0.3340 - output_4_acc: 0.3402 - val_loss: 8.8132 - val_output_0_loss: 1.7986 - val_output_1_loss: 1.7783 - val_output_2_loss: 1.7640 - val_output_3_loss: 1.7330 - val_output_4_loss: 1.7392 - val_output_0_acc: 0.2928 - val_output_1_acc: 0.3260 - val_output_2_acc: 0.3070 - val_output_3_acc: 0.3267 - val_output_4_acc: 0.3145
Epoch 7/200
4000/4000 [==============================] - 1s 160us/step - loss: 8.3267 - output_0_loss: 1.6583 - output_1_loss: 1.6799 - output_2_loss: 1.6621 - output_3_loss: 1.6674 - output_4_loss: 1.6591 - output_0_acc: 0.3550 - output_1_acc: 0.3485 - output_2_acc: 0.3415 - output_3_acc: 0.3525 - output_4_acc: 0.3460 - val_loss: 8.5341 - val_output_0_loss: 1.6802 - val_output_1_loss: 1.7029 - val_output_2_loss: 1.6884 - val_output_3_loss: 1.7005 - val_output_4_loss: 1.7623 - val_output_0_acc: 0.3263 - val_output_1_acc: 0.3223 - val_output_2_acc: 0.3165 - val_output_3_acc: 0.3058 - val_output_4_acc: 0.3267
Epoch 8/200
4000/4000 [==============================] - 1s 162us/step - loss: 7.9822 - output_0_loss: 1.5928 - output_1_loss: 1.6156 - output_2_loss: 1.5955 - output_3_loss: 1.5974 - output_4_loss: 1.5810 - output_0_acc: 0.3832 - output_1_acc: 0.3623 - output_2_acc: 0.3695 - output_3_acc: 0.3728 - output_4_acc: 0.3832 - val_loss: 8.1554 - val_output_0_loss: 1.6856 - val_output_1_loss: 1.6192 - val_output_2_loss: 1.5576 - val_output_3_loss: 1.6494 - val_output_4_loss: 1.6437 - val_output_0_acc: 0.3225 - val_output_1_acc: 0.3583 - val_output_2_acc: 0.3825 - val_output_3_acc: 0.3260 - val_output_4_acc: 0.3588
Epoch 9/200
4000/4000 [==============================] - 1s 169us/step - loss: 7.6968 - output_0_loss: 1.5393 - output_1_loss: 1.5603 - output_2_loss: 1.5354 - output_3_loss: 1.5374 - output_4_loss: 1.5244 - output_0_acc: 0.4022 - output_1_acc: 0.3815 - output_2_acc: 0.3955 - output_3_acc: 0.3900 - output_4_acc: 0.3942 - val_loss: 7.9822 - val_output_0_loss: 1.5970 - val_output_1_loss: 1.5764 - val_output_2_loss: 1.6297 - val_output_3_loss: 1.6300 - val_output_4_loss: 1.5490 - val_output_0_acc: 0.3735 - val_output_1_acc: 0.3772 - val_output_2_acc: 0.3327 - val_output_3_acc: 0.3495 - val_output_4_acc: 0.3937
Epoch 10/200
4000/4000 [==============================] - 1s 166us/step - loss: 7.4401 - output_0_loss: 1.4873 - output_1_loss: 1.5055 - output_2_loss: 1.4710 - output_3_loss: 1.4962 - output_4_loss: 1.4801 - output_0_acc: 0.4150 - output_1_acc: 0.3972 - output_2_acc: 0.4312 - output_3_acc: 0.4108 - output_4_acc: 0.4160 - val_loss: 7.7654 - val_output_0_loss: 1.5389 - val_output_1_loss: 1.5070 - val_output_2_loss: 1.5630 - val_output_3_loss: 1.5511 - val_output_4_loss: 1.6054 - val_output_0_acc: 0.3635 - val_output_1_acc: 0.4015 - val_output_2_acc: 0.3645 - val_output_3_acc: 0.3790 - val_output_4_acc: 0.3612
Epoch 11/200
4000/4000 [==============================] - 1s 177us/step - loss: 7.1842 - output_0_loss: 1.4271 - output_1_loss: 1.4560 - output_2_loss: 1.4360 - output_3_loss: 1.4379 - output_4_loss: 1.4272 - output_0_acc: 0.4368 - output_1_acc: 0.4163 - output_2_acc: 0.4378 - output_3_acc: 0.4365 - output_4_acc: 0.4267 - val_loss: 7.5532 - val_output_0_loss: 1.4726 - val_output_1_loss: 1.5377 - val_output_2_loss: 1.5122 - val_output_3_loss: 1.5288 - val_output_4_loss: 1.5019 - val_output_0_acc: 0.3883 - val_output_1_acc: 0.3585 - val_output_2_acc: 0.3850 - val_output_3_acc: 0.3733 - val_output_4_acc: 0.3997
Epoch 12/200
4000/4000 [==============================] - 1s 165us/step - loss: 6.9615 - output_0_loss: 1.3912 - output_1_loss: 1.4129 - output_2_loss: 1.3940 - output_3_loss: 1.3944 - output_4_loss: 1.3690 - output_0_acc: 0.4465 - output_1_acc: 0.4365 - output_2_acc: 0.4515 - output_3_acc: 0.4462 - output_4_acc: 0.4557 - val_loss: 7.3401 - val_output_0_loss: 1.4685 - val_output_1_loss: 1.4361 - val_output_2_loss: 1.5056 - val_output_3_loss: 1.4428 - val_output_4_loss: 1.4872 - val_output_0_acc: 0.4033 - val_output_1_acc: 0.4505 - val_output_2_acc: 0.3750 - val_output_3_acc: 0.3953 - val_output_4_acc: 0.3793
Epoch 13/200
4000/4000 [==============================] - 1s 164us/step - loss: 6.7411 - output_0_loss: 1.3347 - output_1_loss: 1.3667 - output_2_loss: 1.3502 - output_3_loss: 1.3533 - output_4_loss: 1.3362 - output_0_acc: 0.4770 - output_1_acc: 0.4520 - output_2_acc: 0.4632 - output_3_acc: 0.4625 - output_4_acc: 0.4738 - val_loss: 7.1142 - val_output_0_loss: 1.4732 - val_output_1_loss: 1.4464 - val_output_2_loss: 1.3640 - val_output_3_loss: 1.3822 - val_output_4_loss: 1.4484 - val_output_0_acc: 0.3837 - val_output_1_acc: 0.4135 - val_output_2_acc: 0.4565 - val_output_3_acc: 0.4415 - val_output_4_acc: 0.4142
Epoch 14/200
4000/4000 [==============================] - 1s 162us/step - loss: 6.5364 - output_0_loss: 1.3047 - output_1_loss: 1.3204 - output_2_loss: 1.3039 - output_3_loss: 1.3134 - output_4_loss: 1.2940 - output_0_acc: 0.4768 - output_1_acc: 0.4850 - output_2_acc: 0.4778 - output_3_acc: 0.4793 - output_4_acc: 0.4847 - val_loss: 6.9679 - val_output_0_loss: 1.4419 - val_output_1_loss: 1.4042 - val_output_2_loss: 1.4371 - val_output_3_loss: 1.3242 - val_output_4_loss: 1.3606 - val_output_0_acc: 0.4267 - val_output_1_acc: 0.4197 - val_output_2_acc: 0.4210 - val_output_3_acc: 0.4737 - val_output_4_acc: 0.4422
Epoch 15/200
4000/4000 [==============================] - 1s 175us/step - loss: 6.3527 - output_0_loss: 1.2722 - output_1_loss: 1.2857 - output_2_loss: 1.2730 - output_3_loss: 1.2668 - output_4_loss: 1.2550 - output_0_acc: 0.4903 - output_1_acc: 0.4840 - output_2_acc: 0.4873 - output_3_acc: 0.4982 - output_4_acc: 0.5010 - val_loss: 6.8221 - val_output_0_loss: 1.4683 - val_output_1_loss: 1.3773 - val_output_2_loss: 1.3582 - val_output_3_loss: 1.2996 - val_output_4_loss: 1.3187 - val_output_0_acc: 0.4448 - val_output_1_acc: 0.4313 - val_output_2_acc: 0.4605 - val_output_3_acc: 0.4767 - val_output_4_acc: 0.4438
Epoch 16/200
4000/4000 [==============================] - 1s 170us/step - loss: 6.1924 - output_0_loss: 1.2399 - output_1_loss: 1.2582 - output_2_loss: 1.2314 - output_3_loss: 1.2406 - output_4_loss: 1.2223 - output_0_acc: 0.5067 - output_1_acc: 0.4988 - output_2_acc: 0.5065 - output_3_acc: 0.5135 - output_4_acc: 0.5190 - val_loss: 6.5932 - val_output_0_loss: 1.2500 - val_output_1_loss: 1.3540 - val_output_2_loss: 1.3713 - val_output_3_loss: 1.2643 - val_output_4_loss: 1.3535 - val_output_0_acc: 0.5047 - val_output_1_acc: 0.4638 - val_output_2_acc: 0.4267 - val_output_3_acc: 0.4842 - val_output_4_acc: 0.4648
Epoch 17/200
4000/4000 [==============================] - 1s 173us/step - loss: 6.0228 - output_0_loss: 1.1991 - output_1_loss: 1.2256 - output_2_loss: 1.1996 - output_3_loss: 1.2034 - output_4_loss: 1.1951 - output_0_acc: 0.5270 - output_1_acc: 0.5035 - output_2_acc: 0.5265 - output_3_acc: 0.5298 - output_4_acc: 0.5252 - val_loss: 6.6047 - val_output_0_loss: 1.4005 - val_output_1_loss: 1.3032 - val_output_2_loss: 1.3973 - val_output_3_loss: 1.2497 - val_output_4_loss: 1.2539 - val_output_0_acc: 0.4393 - val_output_1_acc: 0.4592 - val_output_2_acc: 0.4220 - val_output_3_acc: 0.5007 - val_output_4_acc: 0.4583
Epoch 18/200
4000/4000 [==============================] - 1s 174us/step - loss: 5.8550 - output_0_loss: 1.1719 - output_1_loss: 1.1866 - output_2_loss: 1.1647 - output_3_loss: 1.1715 - output_4_loss: 1.1603 - output_0_acc: 0.5298 - output_1_acc: 0.5275 - output_2_acc: 0.5335 - output_3_acc: 0.5367 - output_4_acc: 0.5333 - val_loss: 6.3743 - val_output_0_loss: 1.2539 - val_output_1_loss: 1.3202 - val_output_2_loss: 1.2718 - val_output_3_loss: 1.3160 - val_output_4_loss: 1.2125 - val_output_0_acc: 0.4992 - val_output_1_acc: 0.4687 - val_output_2_acc: 0.4825 - val_output_3_acc: 0.4722 - val_output_4_acc: 0.5132
Epoch 19/200
4000/4000 [==============================] - 1s 176us/step - loss: 5.7058 - output_0_loss: 1.1321 - output_1_loss: 1.1493 - output_2_loss: 1.1395 - output_3_loss: 1.1493 - output_4_loss: 1.1356 - output_0_acc: 0.5635 - output_1_acc: 0.5487 - output_2_acc: 0.5470 - output_3_acc: 0.5345 - output_4_acc: 0.5505 - val_loss: 6.0244 - val_output_0_loss: 1.1712 - val_output_1_loss: 1.2221 - val_output_2_loss: 1.1969 - val_output_3_loss: 1.2357 - val_output_4_loss: 1.1985 - val_output_0_acc: 0.5308 - val_output_1_acc: 0.5090 - val_output_2_acc: 0.5283 - val_output_3_acc: 0.4847 - val_output_4_acc: 0.5183
Epoch 20/200
4000/4000 [==============================] - 1s 185us/step - loss: 5.5692 - output_0_loss: 1.1170 - output_1_loss: 1.1280 - output_2_loss: 1.1064 - output_3_loss: 1.1137 - output_4_loss: 1.1041 - output_0_acc: 0.5523 - output_1_acc: 0.5480 - output_2_acc: 0.5590 - output_3_acc: 0.5530 - output_4_acc: 0.5533 - val_loss: 6.0317 - val_output_0_loss: 1.1908 - val_output_1_loss: 1.1992 - val_output_2_loss: 1.2793 - val_output_3_loss: 1.1982 - val_output_4_loss: 1.1643 - val_output_0_acc: 0.5180 - val_output_1_acc: 0.5112 - val_output_2_acc: 0.4663 - val_output_3_acc: 0.5077 - val_output_4_acc: 0.5213
Epoch 21/200
4000/4000 [==============================] - 1s 178us/step - loss: 5.4580 - output_0_loss: 1.0879 - output_1_loss: 1.1103 - output_2_loss: 1.0866 - output_3_loss: 1.0916 - output_4_loss: 1.0816 - output_0_acc: 0.5640 - output_1_acc: 0.5530 - output_2_acc: 0.5647 - output_3_acc: 0.5640 - output_4_acc: 0.5710 - val_loss: 5.9934 - val_output_0_loss: 1.1761 - val_output_1_loss: 1.1879 - val_output_2_loss: 1.2022 - val_output_3_loss: 1.2046 - val_output_4_loss: 1.2227 - val_output_0_acc: 0.5408 - val_output_1_acc: 0.5075 - val_output_2_acc: 0.4837 - val_output_3_acc: 0.4918 - val_output_4_acc: 0.4650
Epoch 22/200
4000/4000 [==============================] - 1s 180us/step - loss: 5.2921 - output_0_loss: 1.0586 - output_1_loss: 1.0815 - output_2_loss: 1.0477 - output_3_loss: 1.0576 - output_4_loss: 1.0467 - output_0_acc: 0.5885 - output_1_acc: 0.5720 - output_2_acc: 0.5895 - output_3_acc: 0.5803 - output_4_acc: 0.5850 - val_loss: 5.7767 - val_output_0_loss: 1.2078 - val_output_1_loss: 1.1619 - val_output_2_loss: 1.0955 - val_output_3_loss: 1.1482 - val_output_4_loss: 1.1634 - val_output_0_acc: 0.4853 - val_output_1_acc: 0.5055 - val_output_2_acc: 0.5553 - val_output_3_acc: 0.5247 - val_output_4_acc: 0.5203
Epoch 23/200
4000/4000 [==============================] - 1s 171us/step - loss: 5.2049 - output_0_loss: 1.0316 - output_1_loss: 1.0631 - output_2_loss: 1.0359 - output_3_loss: 1.0405 - output_4_loss: 1.0338 - output_0_acc: 0.5880 - output_1_acc: 0.5725 - output_2_acc: 0.5865 - output_3_acc: 0.5952 - output_4_acc: 0.5893 - val_loss: 5.8161 - val_output_0_loss: 1.0999 - val_output_1_loss: 1.1598 - val_output_2_loss: 1.2265 - val_output_3_loss: 1.1794 - val_output_4_loss: 1.1505 - val_output_0_acc: 0.5493 - val_output_1_acc: 0.5277 - val_output_2_acc: 0.5048 - val_output_3_acc: 0.4950 - val_output_4_acc: 0.5135
Epoch 24/200
4000/4000 [==============================] - 1s 173us/step - loss: 5.0537 - output_0_loss: 1.0055 - output_1_loss: 1.0330 - output_2_loss: 1.0047 - output_3_loss: 1.0146 - output_4_loss: 0.9959 - output_0_acc: 0.6020 - output_1_acc: 0.5883 - output_2_acc: 0.6033 - output_3_acc: 0.5958 - output_4_acc: 0.6120 - val_loss: 5.6674 - val_output_0_loss: 1.1370 - val_output_1_loss: 1.1894 - val_output_2_loss: 1.1045 - val_output_3_loss: 1.1094 - val_output_4_loss: 1.1270 - val_output_0_acc: 0.5420 - val_output_1_acc: 0.5140 - val_output_2_acc: 0.5348 - val_output_3_acc: 0.5580 - val_output_4_acc: 0.5270
Epoch 25/200
4000/4000 [==============================] - 1s 177us/step - loss: 4.9656 - output_0_loss: 0.9880 - output_1_loss: 1.0109 - output_2_loss: 0.9837 - output_3_loss: 0.9992 - output_4_loss: 0.9838 - output_0_acc: 0.6148 - output_1_acc: 0.5932 - output_2_acc: 0.6178 - output_3_acc: 0.6025 - output_4_acc: 0.6098 - val_loss: 5.4492 - val_output_0_loss: 1.0675 - val_output_1_loss: 1.1017 - val_output_2_loss: 1.1158 - val_output_3_loss: 1.0902 - val_output_4_loss: 1.0740 - val_output_0_acc: 0.5765 - val_output_1_acc: 0.5393 - val_output_2_acc: 0.5433 - val_output_3_acc: 0.5325 - val_output_4_acc: 0.5518
Epoch 26/200
4000/4000 [==============================] - 1s 165us/step - loss: 4.8478 - output_0_loss: 0.9667 - output_1_loss: 0.9911 - output_2_loss: 0.9577 - output_3_loss: 0.9743 - output_4_loss: 0.9580 - output_0_acc: 0.6162 - output_1_acc: 0.6065 - output_2_acc: 0.6285 - output_3_acc: 0.6120 - output_4_acc: 0.6162 - val_loss: 5.3908 - val_output_0_loss: 1.0939 - val_output_1_loss: 1.0495 - val_output_2_loss: 1.0610 - val_output_3_loss: 1.0923 - val_output_4_loss: 1.0941 - val_output_0_acc: 0.5608 - val_output_1_acc: 0.5752 - val_output_2_acc: 0.5675 - val_output_3_acc: 0.5517 - val_output_4_acc: 0.5733
Epoch 27/200
4000/4000 [==============================] - 1s 161us/step - loss: 4.7407 - output_0_loss: 0.9441 - output_1_loss: 0.9657 - output_2_loss: 0.9423 - output_3_loss: 0.9438 - output_4_loss: 0.9449 - output_0_acc: 0.6275 - output_1_acc: 0.6145 - output_2_acc: 0.6280 - output_3_acc: 0.6342 - output_4_acc: 0.6268 - val_loss: 5.2696 - val_output_0_loss: 1.0702 - val_output_1_loss: 1.0289 - val_output_2_loss: 1.0706 - val_output_3_loss: 1.0243 - val_output_4_loss: 1.0755 - val_output_0_acc: 0.5688 - val_output_1_acc: 0.6003 - val_output_2_acc: 0.5542 - val_output_3_acc: 0.5728 - val_output_4_acc: 0.5623
Epoch 28/200
4000/4000 [==============================] - 1s 164us/step - loss: 4.6139 - output_0_loss: 0.9198 - output_1_loss: 0.9439 - output_2_loss: 0.9147 - output_3_loss: 0.9218 - output_4_loss: 0.9138 - output_0_acc: 0.6370 - output_1_acc: 0.6255 - output_2_acc: 0.6390 - output_3_acc: 0.6445 - output_4_acc: 0.6470 - val_loss: 5.2352 - val_output_0_loss: 1.0568 - val_output_1_loss: 1.0556 - val_output_2_loss: 1.0388 - val_output_3_loss: 1.0573 - val_output_4_loss: 1.0268 - val_output_0_acc: 0.5633 - val_output_1_acc: 0.5653 - val_output_2_acc: 0.5702 - val_output_3_acc: 0.5362 - val_output_4_acc: 0.5853
Epoch 29/200
4000/4000 [==============================] - 1s 166us/step - loss: 4.5419 - output_0_loss: 0.9123 - output_1_loss: 0.9276 - output_2_loss: 0.9037 - output_3_loss: 0.9050 - output_4_loss: 0.8932 - output_0_acc: 0.6492 - output_1_acc: 0.6355 - output_2_acc: 0.6462 - output_3_acc: 0.6482 - output_4_acc: 0.6518 - val_loss: 5.1354 - val_output_0_loss: 1.0232 - val_output_1_loss: 1.0001 - val_output_2_loss: 1.0474 - val_output_3_loss: 1.0656 - val_output_4_loss: 0.9991 - val_output_0_acc: 0.5367 - val_output_1_acc: 0.6123 - val_output_2_acc: 0.5585 - val_output_3_acc: 0.5518 - val_output_4_acc: 0.5668
Epoch 30/200
4000/4000 [==============================] - 1s 178us/step - loss: 4.4303 - output_0_loss: 0.8881 - output_1_loss: 0.9036 - output_2_loss: 0.8859 - output_3_loss: 0.8873 - output_4_loss: 0.8654 - output_0_acc: 0.6485 - output_1_acc: 0.6482 - output_2_acc: 0.6568 - output_3_acc: 0.6515 - output_4_acc: 0.6610 - val_loss: 4.8040 - val_output_0_loss: 0.9366 - val_output_1_loss: 0.9997 - val_output_2_loss: 0.9564 - val_output_3_loss: 0.9190 - val_output_4_loss: 0.9923 - val_output_0_acc: 0.6318 - val_output_1_acc: 0.5933 - val_output_2_acc: 0.6103 - val_output_3_acc: 0.6705 - val_output_4_acc: 0.5793
Epoch 31/200
4000/4000 [==============================] - 1s 166us/step - loss: 4.3351 - output_0_loss: 0.8644 - output_1_loss: 0.8899 - output_2_loss: 0.8556 - output_3_loss: 0.8730 - output_4_loss: 0.8522 - output_0_acc: 0.6718 - output_1_acc: 0.6548 - output_2_acc: 0.6680 - output_3_acc: 0.6530 - output_4_acc: 0.6700 - val_loss: 4.8986 - val_output_0_loss: 1.0000 - val_output_1_loss: 0.9772 - val_output_2_loss: 0.9526 - val_output_3_loss: 0.9062 - val_output_4_loss: 1.0626 - val_output_0_acc: 0.5758 - val_output_1_acc: 0.5895 - val_output_2_acc: 0.6157 - val_output_3_acc: 0.6398 - val_output_4_acc: 0.5537
Epoch 32/200
4000/4000 [==============================] - 1s 174us/step - loss: 4.2476 - output_0_loss: 0.8513 - output_1_loss: 0.8600 - output_2_loss: 0.8483 - output_3_loss: 0.8456 - output_4_loss: 0.8423 - output_0_acc: 0.6667 - output_1_acc: 0.6703 - output_2_acc: 0.6695 - output_3_acc: 0.6665 - output_4_acc: 0.6753 - val_loss: 4.7627 - val_output_0_loss: 1.0025 - val_output_1_loss: 0.9464 - val_output_2_loss: 0.9112 - val_output_3_loss: 0.9471 - val_output_4_loss: 0.9554 - val_output_0_acc: 0.5972 - val_output_1_acc: 0.6217 - val_output_2_acc: 0.6405 - val_output_3_acc: 0.5987 - val_output_4_acc: 0.6395
Epoch 33/200
4000/4000 [==============================] - 1s 181us/step - loss: 4.1535 - output_0_loss: 0.8133 - output_1_loss: 0.8594 - output_2_loss: 0.8286 - output_3_loss: 0.8290 - output_4_loss: 0.8232 - output_0_acc: 0.6955 - output_1_acc: 0.6680 - output_2_acc: 0.6797 - output_3_acc: 0.6823 - output_4_acc: 0.6893 - val_loss: 4.9020 - val_output_0_loss: 1.0007 - val_output_1_loss: 0.9351 - val_output_2_loss: 0.9755 - val_output_3_loss: 0.9651 - val_output_4_loss: 1.0254 - val_output_0_acc: 0.6157 - val_output_1_acc: 0.6240 - val_output_2_acc: 0.5958 - val_output_3_acc: 0.6082 - val_output_4_acc: 0.5727
Epoch 34/200
4000/4000 [==============================] - 1s 183us/step - loss: 4.0875 - output_0_loss: 0.8184 - output_1_loss: 0.8285 - output_2_loss: 0.8195 - output_3_loss: 0.8124 - output_4_loss: 0.8086 - output_0_acc: 0.6937 - output_1_acc: 0.6825 - output_2_acc: 0.6810 - output_3_acc: 0.6945 - output_4_acc: 0.6910 - val_loss: 4.7392 - val_output_0_loss: 0.9661 - val_output_1_loss: 1.0244 - val_output_2_loss: 0.9035 - val_output_3_loss: 0.9186 - val_output_4_loss: 0.9266 - val_output_0_acc: 0.6205 - val_output_1_acc: 0.5828 - val_output_2_acc: 0.6380 - val_output_3_acc: 0.6045 - val_output_4_acc: 0.6255
Epoch 35/200
4000/4000 [==============================] - 1s 180us/step - loss: 3.9985 - output_0_loss: 0.8028 - output_1_loss: 0.8163 - output_2_loss: 0.7924 - output_3_loss: 0.7957 - output_4_loss: 0.7914 - output_0_acc: 0.6947 - output_1_acc: 0.6855 - output_2_acc: 0.6965 - output_3_acc: 0.7023 - output_4_acc: 0.7017 - val_loss: 4.6596 - val_output_0_loss: 0.9230 - val_output_1_loss: 0.9405 - val_output_2_loss: 0.8781 - val_output_3_loss: 0.9696 - val_output_4_loss: 0.9485 - val_output_0_acc: 0.6370 - val_output_1_acc: 0.6183 - val_output_2_acc: 0.6568 - val_output_3_acc: 0.5893 - val_output_4_acc: 0.6107
Epoch 36/200
4000/4000 [==============================] - 1s 171us/step - loss: 3.9392 - output_0_loss: 0.7855 - output_1_loss: 0.7974 - output_2_loss: 0.7822 - output_3_loss: 0.7926 - output_4_loss: 0.7814 - output_0_acc: 0.7025 - output_1_acc: 0.6935 - output_2_acc: 0.7043 - output_3_acc: 0.6950 - output_4_acc: 0.7013 - val_loss: 4.4502 - val_output_0_loss: 0.9155 - val_output_1_loss: 0.8998 - val_output_2_loss: 0.8504 - val_output_3_loss: 0.8838 - val_output_4_loss: 0.9007 - val_output_0_acc: 0.6242 - val_output_1_acc: 0.6290 - val_output_2_acc: 0.6592 - val_output_3_acc: 0.6443 - val_output_4_acc: 0.6500
Epoch 37/200
4000/4000 [==============================] - 1s 168us/step - loss: 3.8349 - output_0_loss: 0.7618 - output_1_loss: 0.7816 - output_2_loss: 0.7656 - output_3_loss: 0.7700 - output_4_loss: 0.7558 - output_0_acc: 0.7223 - output_1_acc: 0.7055 - output_2_acc: 0.7157 - output_3_acc: 0.7040 - output_4_acc: 0.7135 - val_loss: 4.4579 - val_output_0_loss: 0.8909 - val_output_1_loss: 0.9470 - val_output_2_loss: 0.8982 - val_output_3_loss: 0.8440 - val_output_4_loss: 0.8779 - val_output_0_acc: 0.6448 - val_output_1_acc: 0.6148 - val_output_2_acc: 0.6267 - val_output_3_acc: 0.6620 - val_output_4_acc: 0.6510
Epoch 38/200
4000/4000 [==============================] - 1s 164us/step - loss: 3.8022 - output_0_loss: 0.7589 - output_1_loss: 0.7720 - output_2_loss: 0.7624 - output_3_loss: 0.7557 - output_4_loss: 0.7532 - output_0_acc: 0.7137 - output_1_acc: 0.7092 - output_2_acc: 0.7075 - output_3_acc: 0.7152 - output_4_acc: 0.7175 - val_loss: 4.4245 - val_output_0_loss: 0.8817 - val_output_1_loss: 0.8862 - val_output_2_loss: 0.8716 - val_output_3_loss: 0.8382 - val_output_4_loss: 0.9469 - val_output_0_acc: 0.6363 - val_output_1_acc: 0.6555 - val_output_2_acc: 0.6482 - val_output_3_acc: 0.6650 - val_output_4_acc: 0.5880
Epoch 39/200
4000/4000 [==============================] - 1s 168us/step - loss: 3.7101 - output_0_loss: 0.7394 - output_1_loss: 0.7583 - output_2_loss: 0.7357 - output_3_loss: 0.7456 - output_4_loss: 0.7311 - output_0_acc: 0.7223 - output_1_acc: 0.7125 - output_2_acc: 0.7218 - output_3_acc: 0.7163 - output_4_acc: 0.7212 - val_loss: 4.2557 - val_output_0_loss: 0.8530 - val_output_1_loss: 0.8273 - val_output_2_loss: 0.8498 - val_output_3_loss: 0.8625 - val_output_4_loss: 0.8630 - val_output_0_acc: 0.6700 - val_output_1_acc: 0.6900 - val_output_2_acc: 0.6625 - val_output_3_acc: 0.6743 - val_output_4_acc: 0.6383
Epoch 40/200
4000/4000 [==============================] - 1s 172us/step - loss: 3.6365 - output_0_loss: 0.7315 - output_1_loss: 0.7389 - output_2_loss: 0.7259 - output_3_loss: 0.7233 - output_4_loss: 0.7169 - output_0_acc: 0.7255 - output_1_acc: 0.7250 - output_2_acc: 0.7318 - output_3_acc: 0.7340 - output_4_acc: 0.7240 - val_loss: 4.1628 - val_output_0_loss: 0.8495 - val_output_1_loss: 0.8375 - val_output_2_loss: 0.8739 - val_output_3_loss: 0.7739 - val_output_4_loss: 0.8280 - val_output_0_acc: 0.6637 - val_output_1_acc: 0.6478 - val_output_2_acc: 0.6525 - val_output_3_acc: 0.7075 - val_output_4_acc: 0.6832
Epoch 41/200
4000/4000 [==============================] - 1s 171us/step - loss: 3.5764 - output_0_loss: 0.7106 - output_1_loss: 0.7292 - output_2_loss: 0.7108 - output_3_loss: 0.7160 - output_4_loss: 0.7098 - output_0_acc: 0.7385 - output_1_acc: 0.7268 - output_2_acc: 0.7390 - output_3_acc: 0.7358 - output_4_acc: 0.7380 - val_loss: 4.1831 - val_output_0_loss: 0.8230 - val_output_1_loss: 0.8489 - val_output_2_loss: 0.8145 - val_output_3_loss: 0.8708 - val_output_4_loss: 0.8259 - val_output_0_acc: 0.6883 - val_output_1_acc: 0.6855 - val_output_2_acc: 0.6855 - val_output_3_acc: 0.6340 - val_output_4_acc: 0.6868
Epoch 42/200
4000/4000 [==============================] - 1s 172us/step - loss: 3.4898 - output_0_loss: 0.6980 - output_1_loss: 0.7135 - output_2_loss: 0.6967 - output_3_loss: 0.6960 - output_4_loss: 0.6855 - output_0_acc: 0.7448 - output_1_acc: 0.7410 - output_2_acc: 0.7365 - output_3_acc: 0.7433 - output_4_acc: 0.7440 - val_loss: 4.2469 - val_output_0_loss: 0.8409 - val_output_1_loss: 0.8333 - val_output_2_loss: 0.9127 - val_output_3_loss: 0.8248 - val_output_4_loss: 0.8351 - val_output_0_acc: 0.6290 - val_output_1_acc: 0.6582 - val_output_2_acc: 0.6490 - val_output_3_acc: 0.6618 - val_output_4_acc: 0.6527
Epoch 43/200
4000/4000 [==============================] - 1s 172us/step - loss: 3.4514 - output_0_loss: 0.6818 - output_1_loss: 0.7140 - output_2_loss: 0.6875 - output_3_loss: 0.6857 - output_4_loss: 0.6825 - output_0_acc: 0.7505 - output_1_acc: 0.7385 - output_2_acc: 0.7370 - output_3_acc: 0.7493 - output_4_acc: 0.7490 - val_loss: 4.0155 - val_output_0_loss: 0.8257 - val_output_1_loss: 0.8162 - val_output_2_loss: 0.7915 - val_output_3_loss: 0.7594 - val_output_4_loss: 0.8227 - val_output_0_acc: 0.6528 - val_output_1_acc: 0.6778 - val_output_2_acc: 0.6697 - val_output_3_acc: 0.7117 - val_output_4_acc: 0.6547
Epoch 44/200
4000/4000 [==============================] - 1s 180us/step - loss: 3.3898 - output_0_loss: 0.6776 - output_1_loss: 0.6931 - output_2_loss: 0.6742 - output_3_loss: 0.6706 - output_4_loss: 0.6745 - output_0_acc: 0.7532 - output_1_acc: 0.7452 - output_2_acc: 0.7535 - output_3_acc: 0.7612 - output_4_acc: 0.7485 - val_loss: 3.8810 - val_output_0_loss: 0.7460 - val_output_1_loss: 0.8365 - val_output_2_loss: 0.8143 - val_output_3_loss: 0.7477 - val_output_4_loss: 0.7365 - val_output_0_acc: 0.7400 - val_output_1_acc: 0.6815 - val_output_2_acc: 0.6682 - val_output_3_acc: 0.7172 - val_output_4_acc: 0.7367
Epoch 45/200
4000/4000 [==============================] - 1s 172us/step - loss: 3.3142 - output_0_loss: 0.6570 - output_1_loss: 0.6830 - output_2_loss: 0.6599 - output_3_loss: 0.6589 - output_4_loss: 0.6554 - output_0_acc: 0.7642 - output_1_acc: 0.7482 - output_2_acc: 0.7635 - output_3_acc: 0.7630 - output_4_acc: 0.7640 - val_loss: 3.9591 - val_output_0_loss: 0.7781 - val_output_1_loss: 0.8266 - val_output_2_loss: 0.7968 - val_output_3_loss: 0.7658 - val_output_4_loss: 0.7918 - val_output_0_acc: 0.6955 - val_output_1_acc: 0.6822 - val_output_2_acc: 0.6680 - val_output_3_acc: 0.7062 - val_output_4_acc: 0.6967
Epoch 46/200
4000/4000 [==============================] - 1s 160us/step - loss: 3.2577 - output_0_loss: 0.6467 - output_1_loss: 0.6659 - output_2_loss: 0.6465 - output_3_loss: 0.6504 - output_4_loss: 0.6482 - output_0_acc: 0.7690 - output_1_acc: 0.7585 - output_2_acc: 0.7635 - output_3_acc: 0.7605 - output_4_acc: 0.7698 - val_loss: 4.0005 - val_output_0_loss: 0.7493 - val_output_1_loss: 0.9295 - val_output_2_loss: 0.7326 - val_output_3_loss: 0.7401 - val_output_4_loss: 0.8489 - val_output_0_acc: 0.7123 - val_output_1_acc: 0.6208 - val_output_2_acc: 0.7410 - val_output_3_acc: 0.6982 - val_output_4_acc: 0.6402
Epoch 47/200
4000/4000 [==============================] - 1s 165us/step - loss: 3.1912 - output_0_loss: 0.6390 - output_1_loss: 0.6604 - output_2_loss: 0.6279 - output_3_loss: 0.6354 - output_4_loss: 0.6285 - output_0_acc: 0.7725 - output_1_acc: 0.7585 - output_2_acc: 0.7802 - output_3_acc: 0.7700 - output_4_acc: 0.7762 - val_loss: 3.9916 - val_output_0_loss: 0.8037 - val_output_1_loss: 0.8648 - val_output_2_loss: 0.8379 - val_output_3_loss: 0.7650 - val_output_4_loss: 0.7202 - val_output_0_acc: 0.6773 - val_output_1_acc: 0.6455 - val_output_2_acc: 0.6307 - val_output_3_acc: 0.6975 - val_output_4_acc: 0.7233
Epoch 48/200
4000/4000 [==============================] - 1s 170us/step - loss: 3.1502 - output_0_loss: 0.6288 - output_1_loss: 0.6375 - output_2_loss: 0.6297 - output_3_loss: 0.6286 - output_4_loss: 0.6257 - output_0_acc: 0.7750 - output_1_acc: 0.7670 - output_2_acc: 0.7755 - output_3_acc: 0.7770 - output_4_acc: 0.7855 - val_loss: 3.9092 - val_output_0_loss: 0.7843 - val_output_1_loss: 0.8130 - val_output_2_loss: 0.8081 - val_output_3_loss: 0.7623 - val_output_4_loss: 0.7414 - val_output_0_acc: 0.6710 - val_output_1_acc: 0.6697 - val_output_2_acc: 0.6603 - val_output_3_acc: 0.7032 - val_output_4_acc: 0.6945
Epoch 49/200
4000/4000 [==============================] - 1s 182us/step - loss: 3.0707 - output_0_loss: 0.6166 - output_1_loss: 0.6319 - output_2_loss: 0.6111 - output_3_loss: 0.6063 - output_4_loss: 0.6048 - output_0_acc: 0.7802 - output_1_acc: 0.7635 - output_2_acc: 0.7790 - output_3_acc: 0.7840 - output_4_acc: 0.7895 - val_loss: 3.8500 - val_output_0_loss: 0.8035 - val_output_1_loss: 0.7451 - val_output_2_loss: 0.7346 - val_output_3_loss: 0.7675 - val_output_4_loss: 0.7992 - val_output_0_acc: 0.6833 - val_output_1_acc: 0.7068 - val_output_2_acc: 0.7268 - val_output_3_acc: 0.7030 - val_output_4_acc: 0.6632
Epoch 50/200
4000/4000 [==============================] - 1s 178us/step - loss: 3.0436 - output_0_loss: 0.6041 - output_1_loss: 0.6248 - output_2_loss: 0.6029 - output_3_loss: 0.6085 - output_4_loss: 0.6033 - output_0_acc: 0.7855 - output_1_acc: 0.7795 - output_2_acc: 0.7910 - output_3_acc: 0.7777 - output_4_acc: 0.7917 - val_loss: 3.5668 - val_output_0_loss: 0.7293 - val_output_1_loss: 0.6980 - val_output_2_loss: 0.6859 - val_output_3_loss: 0.6683 - val_output_4_loss: 0.7852 - val_output_0_acc: 0.7212 - val_output_1_acc: 0.7585 - val_output_2_acc: 0.7547 - val_output_3_acc: 0.7727 - val_output_4_acc: 0.6793
Epoch 51/200
4000/4000 [==============================] - 1s 165us/step - loss: 2.9856 - output_0_loss: 0.5941 - output_1_loss: 0.6114 - output_2_loss: 0.5957 - output_3_loss: 0.5900 - output_4_loss: 0.5943 - output_0_acc: 0.7893 - output_1_acc: 0.7830 - output_2_acc: 0.7887 - output_3_acc: 0.7910 - output_4_acc: 0.7947 - val_loss: 3.5883 - val_output_0_loss: 0.7794 - val_output_1_loss: 0.7284 - val_output_2_loss: 0.6943 - val_output_3_loss: 0.7153 - val_output_4_loss: 0.6710 - val_output_0_acc: 0.6730 - val_output_1_acc: 0.7118 - val_output_2_acc: 0.7505 - val_output_3_acc: 0.6985 - val_output_4_acc: 0.7630
Epoch 52/200
4000/4000 [==============================] - 1s 164us/step - loss: 2.9313 - output_0_loss: 0.5862 - output_1_loss: 0.5999 - output_2_loss: 0.5806 - output_3_loss: 0.5854 - output_4_loss: 0.5793 - output_0_acc: 0.7973 - output_1_acc: 0.7920 - output_2_acc: 0.8010 - output_3_acc: 0.7867 - output_4_acc: 0.8058 - val_loss: 3.7293 - val_output_0_loss: 0.7993 - val_output_1_loss: 0.7392 - val_output_2_loss: 0.7148 - val_output_3_loss: 0.7166 - val_output_4_loss: 0.7595 - val_output_0_acc: 0.6783 - val_output_1_acc: 0.6910 - val_output_2_acc: 0.7297 - val_output_3_acc: 0.7148 - val_output_4_acc: 0.7148
Epoch 53/200
4000/4000 [==============================] - 1s 171us/step - loss: 2.8794 - output_0_loss: 0.5713 - output_1_loss: 0.5976 - output_2_loss: 0.5721 - output_3_loss: 0.5580 - output_4_loss: 0.5804 - output_0_acc: 0.8045 - output_1_acc: 0.7945 - output_2_acc: 0.7993 - output_3_acc: 0.8083 - output_4_acc: 0.7963 - val_loss: 3.7743 - val_output_0_loss: 0.7403 - val_output_1_loss: 0.7756 - val_output_2_loss: 0.7173 - val_output_3_loss: 0.8652 - val_output_4_loss: 0.6759 - val_output_0_acc: 0.7015 - val_output_1_acc: 0.6987 - val_output_2_acc: 0.7113 - val_output_3_acc: 0.6153 - val_output_4_acc: 0.7508
Epoch 54/200
4000/4000 [==============================] - 1s 180us/step - loss: 2.8393 - output_0_loss: 0.5631 - output_1_loss: 0.5842 - output_2_loss: 0.5628 - output_3_loss: 0.5737 - output_4_loss: 0.5556 - output_0_acc: 0.8060 - output_1_acc: 0.7977 - output_2_acc: 0.8053 - output_3_acc: 0.7940 - output_4_acc: 0.8150 - val_loss: 3.5624 - val_output_0_loss: 0.7406 - val_output_1_loss: 0.7351 - val_output_2_loss: 0.7743 - val_output_3_loss: 0.6773 - val_output_4_loss: 0.6352 - val_output_0_acc: 0.7100 - val_output_1_acc: 0.7145 - val_output_2_acc: 0.6888 - val_output_3_acc: 0.7390 - val_output_4_acc: 0.7882
Epoch 55/200
4000/4000 [==============================] - 1s 187us/step - loss: 2.7801 - output_0_loss: 0.5504 - output_1_loss: 0.5806 - output_2_loss: 0.5591 - output_3_loss: 0.5416 - output_4_loss: 0.5483 - output_0_acc: 0.8105 - output_1_acc: 0.7983 - output_2_acc: 0.8045 - output_3_acc: 0.8192 - output_4_acc: 0.8097 - val_loss: 3.4724 - val_output_0_loss: 0.6953 - val_output_1_loss: 0.7101 - val_output_2_loss: 0.6728 - val_output_3_loss: 0.6748 - val_output_4_loss: 0.7194 - val_output_0_acc: 0.7483 - val_output_1_acc: 0.7298 - val_output_2_acc: 0.7423 - val_output_3_acc: 0.7327 - val_output_4_acc: 0.7388
Epoch 56/200
4000/4000 [==============================] - 1s 190us/step - loss: 2.7289 - output_0_loss: 0.5474 - output_1_loss: 0.5650 - output_2_loss: 0.5448 - output_3_loss: 0.5348 - output_4_loss: 0.5369 - output_0_acc: 0.8123 - output_1_acc: 0.7987 - output_2_acc: 0.8157 - output_3_acc: 0.8180 - output_4_acc: 0.8143 - val_loss: 3.3738 - val_output_0_loss: 0.6483 - val_output_1_loss: 0.6948 - val_output_2_loss: 0.6684 - val_output_3_loss: 0.6710 - val_output_4_loss: 0.6912 - val_output_0_acc: 0.7795 - val_output_1_acc: 0.7393 - val_output_2_acc: 0.7517 - val_output_3_acc: 0.7457 - val_output_4_acc: 0.7557
Epoch 57/200
4000/4000 [==============================] - 1s 197us/step - loss: 2.6912 - output_0_loss: 0.5348 - output_1_loss: 0.5559 - output_2_loss: 0.5368 - output_3_loss: 0.5317 - output_4_loss: 0.5320 - output_0_acc: 0.8190 - output_1_acc: 0.8150 - output_2_acc: 0.8157 - output_3_acc: 0.8123 - output_4_acc: 0.8243 - val_loss: 3.4103 - val_output_0_loss: 0.6774 - val_output_1_loss: 0.6982 - val_output_2_loss: 0.6930 - val_output_3_loss: 0.6844 - val_output_4_loss: 0.6573 - val_output_0_acc: 0.7330 - val_output_1_acc: 0.7307 - val_output_2_acc: 0.7298 - val_output_3_acc: 0.7197 - val_output_4_acc: 0.7258
Epoch 58/200
4000/4000 [==============================] - 1s 194us/step - loss: 2.6459 - output_0_loss: 0.5249 - output_1_loss: 0.5481 - output_2_loss: 0.5297 - output_3_loss: 0.5172 - output_4_loss: 0.5260 - output_0_acc: 0.8227 - output_1_acc: 0.8100 - output_2_acc: 0.8170 - output_3_acc: 0.8250 - output_4_acc: 0.8220 - val_loss: 3.2560 - val_output_0_loss: 0.6502 - val_output_1_loss: 0.6849 - val_output_2_loss: 0.6327 - val_output_3_loss: 0.6609 - val_output_4_loss: 0.6273 - val_output_0_acc: 0.7613 - val_output_1_acc: 0.7417 - val_output_2_acc: 0.7548 - val_output_3_acc: 0.7527 - val_output_4_acc: 0.7685
Epoch 59/200
4000/4000 [==============================] - 1s 184us/step - loss: 2.5872 - output_0_loss: 0.5191 - output_1_loss: 0.5330 - output_2_loss: 0.5166 - output_3_loss: 0.5076 - output_4_loss: 0.5109 - output_0_acc: 0.8203 - output_1_acc: 0.8217 - output_2_acc: 0.8317 - output_3_acc: 0.8273 - output_4_acc: 0.8337 - val_loss: 3.3952 - val_output_0_loss: 0.6731 - val_output_1_loss: 0.6663 - val_output_2_loss: 0.6268 - val_output_3_loss: 0.6832 - val_output_4_loss: 0.7458 - val_output_0_acc: 0.7450 - val_output_1_acc: 0.7363 - val_output_2_acc: 0.7640 - val_output_3_acc: 0.7153 - val_output_4_acc: 0.6868
Epoch 60/200
4000/4000 [==============================] - 1s 165us/step - loss: 2.5614 - output_0_loss: 0.5062 - output_1_loss: 0.5383 - output_2_loss: 0.5128 - output_3_loss: 0.5028 - output_4_loss: 0.5014 - output_0_acc: 0.8293 - output_1_acc: 0.8170 - output_2_acc: 0.8297 - output_3_acc: 0.8375 - output_4_acc: 0.8303 - val_loss: 3.2217 - val_output_0_loss: 0.6139 - val_output_1_loss: 0.6452 - val_output_2_loss: 0.6625 - val_output_3_loss: 0.6416 - val_output_4_loss: 0.6584 - val_output_0_acc: 0.8025 - val_output_1_acc: 0.7728 - val_output_2_acc: 0.7357 - val_output_3_acc: 0.7512 - val_output_4_acc: 0.7372
Epoch 61/200
4000/4000 [==============================] - 1s 182us/step - loss: 2.5129 - output_0_loss: 0.4991 - output_1_loss: 0.5151 - output_2_loss: 0.4949 - output_3_loss: 0.5007 - output_4_loss: 0.5032 - output_0_acc: 0.8337 - output_1_acc: 0.8267 - output_2_acc: 0.8458 - output_3_acc: 0.8340 - output_4_acc: 0.8297 - val_loss: 3.1599 - val_output_0_loss: 0.6847 - val_output_1_loss: 0.6706 - val_output_2_loss: 0.6275 - val_output_3_loss: 0.6142 - val_output_4_loss: 0.5628 - val_output_0_acc: 0.7097 - val_output_1_acc: 0.7452 - val_output_2_acc: 0.7560 - val_output_3_acc: 0.7797 - val_output_4_acc: 0.8185
Epoch 62/200
4000/4000 [==============================] - 1s 184us/step - loss: 2.4620 - output_0_loss: 0.4983 - output_1_loss: 0.5119 - output_2_loss: 0.4867 - output_3_loss: 0.4810 - output_4_loss: 0.4841 - output_0_acc: 0.8397 - output_1_acc: 0.8305 - output_2_acc: 0.8455 - output_3_acc: 0.8423 - output_4_acc: 0.8460 - val_loss: 3.1291 - val_output_0_loss: 0.5817 - val_output_1_loss: 0.6578 - val_output_2_loss: 0.6526 - val_output_3_loss: 0.5421 - val_output_4_loss: 0.6949 - val_output_0_acc: 0.8012 - val_output_1_acc: 0.7627 - val_output_2_acc: 0.7510 - val_output_3_acc: 0.8460 - val_output_4_acc: 0.7045
Epoch 63/200
4000/4000 [==============================] - 1s 180us/step - loss: 2.4461 - output_0_loss: 0.4836 - output_1_loss: 0.5086 - output_2_loss: 0.4849 - output_3_loss: 0.4790 - output_4_loss: 0.4900 - output_0_acc: 0.8407 - output_1_acc: 0.8255 - output_2_acc: 0.8315 - output_3_acc: 0.8515 - output_4_acc: 0.8353 - val_loss: 3.1192 - val_output_0_loss: 0.6159 - val_output_1_loss: 0.6508 - val_output_2_loss: 0.5790 - val_output_3_loss: 0.6916 - val_output_4_loss: 0.5819 - val_output_0_acc: 0.7615 - val_output_1_acc: 0.7462 - val_output_2_acc: 0.7938 - val_output_3_acc: 0.6870 - val_output_4_acc: 0.7835
Epoch 64/200
4000/4000 [==============================] - 1s 166us/step - loss: 2.4022 - output_0_loss: 0.4786 - output_1_loss: 0.4940 - output_2_loss: 0.4797 - output_3_loss: 0.4764 - output_4_loss: 0.4735 - output_0_acc: 0.8423 - output_1_acc: 0.8380 - output_2_acc: 0.8445 - output_3_acc: 0.8432 - output_4_acc: 0.8470 - val_loss: 3.2712 - val_output_0_loss: 0.7097 - val_output_1_loss: 0.6571 - val_output_2_loss: 0.6389 - val_output_3_loss: 0.6239 - val_output_4_loss: 0.6416 - val_output_0_acc: 0.7072 - val_output_1_acc: 0.7498 - val_output_2_acc: 0.7440 - val_output_3_acc: 0.7455 - val_output_4_acc: 0.7323
Epoch 65/200
4000/4000 [==============================] - 1s 173us/step - loss: 2.3599 - output_0_loss: 0.4712 - output_1_loss: 0.4897 - output_2_loss: 0.4656 - output_3_loss: 0.4654 - output_4_loss: 0.4680 - output_0_acc: 0.8478 - output_1_acc: 0.8363 - output_2_acc: 0.8518 - output_3_acc: 0.8530 - output_4_acc: 0.8485 - val_loss: 3.2408 - val_output_0_loss: 0.6451 - val_output_1_loss: 0.6899 - val_output_2_loss: 0.7334 - val_output_3_loss: 0.6033 - val_output_4_loss: 0.5692 - val_output_0_acc: 0.7300 - val_output_1_acc: 0.7103 - val_output_2_acc: 0.7213 - val_output_3_acc: 0.7605 - val_output_4_acc: 0.8132
Epoch 66/200
4000/4000 [==============================] - 1s 170us/step - loss: 2.3146 - output_0_loss: 0.4676 - output_1_loss: 0.4771 - output_2_loss: 0.4642 - output_3_loss: 0.4501 - output_4_loss: 0.4556 - output_0_acc: 0.8445 - output_1_acc: 0.8520 - output_2_acc: 0.8518 - output_3_acc: 0.8548 - output_4_acc: 0.8580 - val_loss: 3.0680 - val_output_0_loss: 0.5719 - val_output_1_loss: 0.6248 - val_output_2_loss: 0.5807 - val_output_3_loss: 0.6486 - val_output_4_loss: 0.6420 - val_output_0_acc: 0.8033 - val_output_1_acc: 0.7587 - val_output_2_acc: 0.7825 - val_output_3_acc: 0.7515 - val_output_4_acc: 0.7432
Epoch 67/200
4000/4000 [==============================] - 1s 166us/step - loss: 2.2732 - output_0_loss: 0.4541 - output_1_loss: 0.4718 - output_2_loss: 0.4576 - output_3_loss: 0.4483 - output_4_loss: 0.4415 - output_0_acc: 0.8568 - output_1_acc: 0.8505 - output_2_acc: 0.8570 - output_3_acc: 0.8565 - output_4_acc: 0.8640 - val_loss: 3.0119 - val_output_0_loss: 0.5786 - val_output_1_loss: 0.5735 - val_output_2_loss: 0.5255 - val_output_3_loss: 0.5879 - val_output_4_loss: 0.7464 - val_output_0_acc: 0.7775 - val_output_1_acc: 0.8110 - val_output_2_acc: 0.8302 - val_output_3_acc: 0.7665 - val_output_4_acc: 0.6772
Epoch 68/200
4000/4000 [==============================] - 1s 188us/step - loss: 2.2309 - output_0_loss: 0.4389 - output_1_loss: 0.4674 - output_2_loss: 0.4439 - output_3_loss: 0.4375 - output_4_loss: 0.4432 - output_0_acc: 0.8668 - output_1_acc: 0.8518 - output_2_acc: 0.8600 - output_3_acc: 0.8623 - output_4_acc: 0.8560 - val_loss: 2.8753 - val_output_0_loss: 0.5937 - val_output_1_loss: 0.6087 - val_output_2_loss: 0.5039 - val_output_3_loss: 0.5790 - val_output_4_loss: 0.5899 - val_output_0_acc: 0.7862 - val_output_1_acc: 0.7620 - val_output_2_acc: 0.8518 - val_output_3_acc: 0.7958 - val_output_4_acc: 0.7887
Epoch 69/200
4000/4000 [==============================] - 1s 173us/step - loss: 2.1855 - output_0_loss: 0.4363 - output_1_loss: 0.4486 - output_2_loss: 0.4365 - output_3_loss: 0.4272 - output_4_loss: 0.4368 - output_0_acc: 0.8582 - output_1_acc: 0.8662 - output_2_acc: 0.8578 - output_3_acc: 0.8658 - output_4_acc: 0.8658 - val_loss: 2.9121 - val_output_0_loss: 0.6330 - val_output_1_loss: 0.5883 - val_output_2_loss: 0.5598 - val_output_3_loss: 0.5696 - val_output_4_loss: 0.5614 - val_output_0_acc: 0.7583 - val_output_1_acc: 0.7925 - val_output_2_acc: 0.8038 - val_output_3_acc: 0.7830 - val_output_4_acc: 0.7995
Epoch 70/200
4000/4000 [==============================] - 1s 170us/step - loss: 2.1668 - output_0_loss: 0.4350 - output_1_loss: 0.4533 - output_2_loss: 0.4282 - output_3_loss: 0.4201 - output_4_loss: 0.4302 - output_0_acc: 0.8665 - output_1_acc: 0.8605 - output_2_acc: 0.8665 - output_3_acc: 0.8690 - output_4_acc: 0.8600 - val_loss: 2.8465 - val_output_0_loss: 0.5689 - val_output_1_loss: 0.6309 - val_output_2_loss: 0.5492 - val_output_3_loss: 0.5376 - val_output_4_loss: 0.5598 - val_output_0_acc: 0.7932 - val_output_1_acc: 0.7590 - val_output_2_acc: 0.8088 - val_output_3_acc: 0.8095 - val_output_4_acc: 0.7953
Epoch 71/200
4000/4000 [==============================] - 1s 163us/step - loss: 2.1187 - output_0_loss: 0.4248 - output_1_loss: 0.4417 - output_2_loss: 0.4160 - output_3_loss: 0.4155 - output_4_loss: 0.4207 - output_0_acc: 0.8675 - output_1_acc: 0.8538 - output_2_acc: 0.8743 - output_3_acc: 0.8755 - output_4_acc: 0.8710 - val_loss: 2.9316 - val_output_0_loss: 0.5695 - val_output_1_loss: 0.5863 - val_output_2_loss: 0.5677 - val_output_3_loss: 0.6512 - val_output_4_loss: 0.5568 - val_output_0_acc: 0.7878 - val_output_1_acc: 0.7855 - val_output_2_acc: 0.7753 - val_output_3_acc: 0.7093 - val_output_4_acc: 0.7897
Epoch 72/200
4000/4000 [==============================] - 1s 166us/step - loss: 2.1019 - output_0_loss: 0.4151 - output_1_loss: 0.4357 - output_2_loss: 0.4268 - output_3_loss: 0.4100 - output_4_loss: 0.4142 - output_0_acc: 0.8760 - output_1_acc: 0.8635 - output_2_acc: 0.8665 - output_3_acc: 0.8710 - output_4_acc: 0.8713 - val_loss: 2.6757 - val_output_0_loss: 0.5864 - val_output_1_loss: 0.5690 - val_output_2_loss: 0.5205 - val_output_3_loss: 0.4865 - val_output_4_loss: 0.5133 - val_output_0_acc: 0.7803 - val_output_1_acc: 0.7885 - val_output_2_acc: 0.8212 - val_output_3_acc: 0.8520 - val_output_4_acc: 0.8222
Epoch 73/200
4000/4000 [==============================] - 1s 171us/step - loss: 2.0393 - output_0_loss: 0.4063 - output_1_loss: 0.4275 - output_2_loss: 0.4041 - output_3_loss: 0.3967 - output_4_loss: 0.4047 - output_0_acc: 0.8775 - output_1_acc: 0.8650 - output_2_acc: 0.8805 - output_3_acc: 0.8868 - output_4_acc: 0.8743 - val_loss: 2.8431 - val_output_0_loss: 0.5663 - val_output_1_loss: 0.6214 - val_output_2_loss: 0.5557 - val_output_3_loss: 0.5457 - val_output_4_loss: 0.5540 - val_output_0_acc: 0.8115 - val_output_1_acc: 0.7503 - val_output_2_acc: 0.7903 - val_output_3_acc: 0.7865 - val_output_4_acc: 0.7848
Epoch 74/200
4000/4000 [==============================] - 1s 164us/step - loss: 2.0118 - output_0_loss: 0.4050 - output_1_loss: 0.4172 - output_2_loss: 0.3956 - output_3_loss: 0.3979 - output_4_loss: 0.3961 - output_0_acc: 0.8840 - output_1_acc: 0.8665 - output_2_acc: 0.8795 - output_3_acc: 0.8822 - output_4_acc: 0.8772 - val_loss: 2.7221 - val_output_0_loss: 0.5525 - val_output_1_loss: 0.5427 - val_output_2_loss: 0.5653 - val_output_3_loss: 0.5130 - val_output_4_loss: 0.5485 - val_output_0_acc: 0.8147 - val_output_1_acc: 0.8085 - val_output_2_acc: 0.7715 - val_output_3_acc: 0.8255 - val_output_4_acc: 0.8027
Epoch 75/200
4000/4000 [==============================] - 1s 168us/step - loss: 2.0008 - output_0_loss: 0.4011 - output_1_loss: 0.4154 - output_2_loss: 0.3967 - output_3_loss: 0.3903 - output_4_loss: 0.3974 - output_0_acc: 0.8738 - output_1_acc: 0.8728 - output_2_acc: 0.8802 - output_3_acc: 0.8870 - output_4_acc: 0.8825 - val_loss: 2.4880 - val_output_0_loss: 0.5167 - val_output_1_loss: 0.5346 - val_output_2_loss: 0.4656 - val_output_3_loss: 0.4710 - val_output_4_loss: 0.5001 - val_output_0_acc: 0.8170 - val_output_1_acc: 0.8043 - val_output_2_acc: 0.8525 - val_output_3_acc: 0.8580 - val_output_4_acc: 0.8247
Epoch 76/200
4000/4000 [==============================] - 1s 178us/step - loss: 1.9470 - output_0_loss: 0.3854 - output_1_loss: 0.4048 - output_2_loss: 0.3818 - output_3_loss: 0.3850 - output_4_loss: 0.3900 - output_0_acc: 0.8877 - output_1_acc: 0.8780 - output_2_acc: 0.8847 - output_3_acc: 0.8912 - output_4_acc: 0.8888 - val_loss: 2.6739 - val_output_0_loss: 0.5305 - val_output_1_loss: 0.5422 - val_output_2_loss: 0.5671 - val_output_3_loss: 0.4845 - val_output_4_loss: 0.5496 - val_output_0_acc: 0.8188 - val_output_1_acc: 0.8010 - val_output_2_acc: 0.8057 - val_output_3_acc: 0.8397 - val_output_4_acc: 0.7880
Epoch 77/200
4000/4000 [==============================] - 1s 184us/step - loss: 1.9125 - output_0_loss: 0.3844 - output_1_loss: 0.3957 - output_2_loss: 0.3759 - output_3_loss: 0.3779 - output_4_loss: 0.3786 - output_0_acc: 0.8838 - output_1_acc: 0.8820 - output_2_acc: 0.8910 - output_3_acc: 0.8932 - output_4_acc: 0.8858 - val_loss: 2.6481 - val_output_0_loss: 0.5244 - val_output_1_loss: 0.5893 - val_output_2_loss: 0.5341 - val_output_3_loss: 0.4915 - val_output_4_loss: 0.5089 - val_output_0_acc: 0.8115 - val_output_1_acc: 0.7647 - val_output_2_acc: 0.7960 - val_output_3_acc: 0.8273 - val_output_4_acc: 0.8350
Epoch 78/200
4000/4000 [==============================] - 1s 166us/step - loss: 1.8980 - output_0_loss: 0.3768 - output_1_loss: 0.3930 - output_2_loss: 0.3751 - output_3_loss: 0.3701 - output_4_loss: 0.3829 - output_0_acc: 0.8902 - output_1_acc: 0.8812 - output_2_acc: 0.8930 - output_3_acc: 0.8918 - output_4_acc: 0.8860 - val_loss: 2.6608 - val_output_0_loss: 0.5890 - val_output_1_loss: 0.5524 - val_output_2_loss: 0.5042 - val_output_3_loss: 0.5015 - val_output_4_loss: 0.5137 - val_output_0_acc: 0.7835 - val_output_1_acc: 0.7878 - val_output_2_acc: 0.8222 - val_output_3_acc: 0.8223 - val_output_4_acc: 0.8268
Epoch 79/200
4000/4000 [==============================] - 1s 170us/step - loss: 1.8583 - output_0_loss: 0.3690 - output_1_loss: 0.3853 - output_2_loss: 0.3688 - output_3_loss: 0.3627 - output_4_loss: 0.3724 - output_0_acc: 0.8967 - output_1_acc: 0.8817 - output_2_acc: 0.8955 - output_3_acc: 0.8967 - output_4_acc: 0.8905 - val_loss: 2.5213 - val_output_0_loss: 0.4910 - val_output_1_loss: 0.5645 - val_output_2_loss: 0.4940 - val_output_3_loss: 0.4712 - val_output_4_loss: 0.5006 - val_output_0_acc: 0.8350 - val_output_1_acc: 0.7875 - val_output_2_acc: 0.8138 - val_output_3_acc: 0.8377 - val_output_4_acc: 0.8263
Epoch 80/200
4000/4000 [==============================] - 1s 162us/step - loss: 1.8265 - output_0_loss: 0.3589 - output_1_loss: 0.3860 - output_2_loss: 0.3610 - output_3_loss: 0.3557 - output_4_loss: 0.3648 - output_0_acc: 0.8942 - output_1_acc: 0.8860 - output_2_acc: 0.9000 - output_3_acc: 0.8922 - output_4_acc: 0.8892 - val_loss: 2.6847 - val_output_0_loss: 0.5893 - val_output_1_loss: 0.4975 - val_output_2_loss: 0.6133 - val_output_3_loss: 0.4639 - val_output_4_loss: 0.5207 - val_output_0_acc: 0.7568 - val_output_1_acc: 0.8332 - val_output_2_acc: 0.7302 - val_output_3_acc: 0.8490 - val_output_4_acc: 0.8057
Epoch 81/200
4000/4000 [==============================] - 1s 163us/step - loss: 1.7905 - output_0_loss: 0.3610 - output_1_loss: 0.3679 - output_2_loss: 0.3578 - output_3_loss: 0.3464 - output_4_loss: 0.3574 - output_0_acc: 0.9012 - output_1_acc: 0.8955 - output_2_acc: 0.8955 - output_3_acc: 0.9027 - output_4_acc: 0.8922 - val_loss: 2.5127 - val_output_0_loss: 0.4997 - val_output_1_loss: 0.5055 - val_output_2_loss: 0.5384 - val_output_3_loss: 0.5230 - val_output_4_loss: 0.4460 - val_output_0_acc: 0.8280 - val_output_1_acc: 0.8347 - val_output_2_acc: 0.7895 - val_output_3_acc: 0.8027 - val_output_4_acc: 0.8483
Epoch 82/200
4000/4000 [==============================] - 1s 169us/step - loss: 1.7644 - output_0_loss: 0.3469 - output_1_loss: 0.3737 - output_2_loss: 0.3481 - output_3_loss: 0.3437 - output_4_loss: 0.3520 - output_0_acc: 0.9040 - output_1_acc: 0.8948 - output_2_acc: 0.9040 - output_3_acc: 0.9012 - output_4_acc: 0.8935 - val_loss: 2.4432 - val_output_0_loss: 0.4942 - val_output_1_loss: 0.4763 - val_output_2_loss: 0.4896 - val_output_3_loss: 0.5054 - val_output_4_loss: 0.4776 - val_output_0_acc: 0.8313 - val_output_1_acc: 0.8472 - val_output_2_acc: 0.8265 - val_output_3_acc: 0.8340 - val_output_4_acc: 0.8330
Epoch 83/200
4000/4000 [==============================] - 1s 170us/step - loss: 1.7378 - output_0_loss: 0.3525 - output_1_loss: 0.3623 - output_2_loss: 0.3437 - output_3_loss: 0.3341 - output_4_loss: 0.3452 - output_0_acc: 0.8982 - output_1_acc: 0.8892 - output_2_acc: 0.9045 - output_3_acc: 0.9057 - output_4_acc: 0.8932 - val_loss: 2.4985 - val_output_0_loss: 0.5306 - val_output_1_loss: 0.4572 - val_output_2_loss: 0.5289 - val_output_3_loss: 0.4794 - val_output_4_loss: 0.5024 - val_output_0_acc: 0.8092 - val_output_1_acc: 0.8543 - val_output_2_acc: 0.8015 - val_output_3_acc: 0.8335 - val_output_4_acc: 0.8342
Epoch 84/200
4000/4000 [==============================] - 1s 165us/step - loss: 1.7285 - output_0_loss: 0.3425 - output_1_loss: 0.3599 - output_2_loss: 0.3465 - output_3_loss: 0.3391 - output_4_loss: 0.3405 - output_0_acc: 0.8995 - output_1_acc: 0.8962 - output_2_acc: 0.9038 - output_3_acc: 0.9095 - output_4_acc: 0.9040 - val_loss: 2.4068 - val_output_0_loss: 0.5058 - val_output_1_loss: 0.5651 - val_output_2_loss: 0.4216 - val_output_3_loss: 0.4302 - val_output_4_loss: 0.4842 - val_output_0_acc: 0.8088 - val_output_1_acc: 0.7742 - val_output_2_acc: 0.8737 - val_output_3_acc: 0.8682 - val_output_4_acc: 0.8293
Epoch 85/200
4000/4000 [==============================] - 1s 164us/step - loss: 1.6634 - output_0_loss: 0.3275 - output_1_loss: 0.3481 - output_2_loss: 0.3282 - output_3_loss: 0.3254 - output_4_loss: 0.3342 - output_0_acc: 0.9135 - output_1_acc: 0.8975 - output_2_acc: 0.9137 - output_3_acc: 0.9135 - output_4_acc: 0.9040 - val_loss: 2.2920 - val_output_0_loss: 0.4369 - val_output_1_loss: 0.5000 - val_output_2_loss: 0.4362 - val_output_3_loss: 0.4443 - val_output_4_loss: 0.4745 - val_output_0_acc: 0.8585 - val_output_1_acc: 0.8160 - val_output_2_acc: 0.8653 - val_output_3_acc: 0.8440 - val_output_4_acc: 0.8298
Epoch 86/200
4000/4000 [==============================] - 1s 165us/step - loss: 1.6478 - output_0_loss: 0.3252 - output_1_loss: 0.3467 - output_2_loss: 0.3288 - output_3_loss: 0.3198 - output_4_loss: 0.3273 - output_0_acc: 0.9120 - output_1_acc: 0.9022 - output_2_acc: 0.9093 - output_3_acc: 0.9087 - output_4_acc: 0.9105 - val_loss: 2.2418 - val_output_0_loss: 0.4383 - val_output_1_loss: 0.4718 - val_output_2_loss: 0.4276 - val_output_3_loss: 0.4587 - val_output_4_loss: 0.4453 - val_output_0_acc: 0.8485 - val_output_1_acc: 0.8405 - val_output_2_acc: 0.8773 - val_output_3_acc: 0.8377 - val_output_4_acc: 0.8555
Epoch 87/200
4000/4000 [==============================] - 1s 167us/step - loss: 1.6221 - output_0_loss: 0.3261 - output_1_loss: 0.3375 - output_2_loss: 0.3183 - output_3_loss: 0.3133 - output_4_loss: 0.3269 - output_0_acc: 0.9062 - output_1_acc: 0.9010 - output_2_acc: 0.9110 - output_3_acc: 0.9177 - output_4_acc: 0.9077 - val_loss: 2.2798 - val_output_0_loss: 0.4420 - val_output_1_loss: 0.4614 - val_output_2_loss: 0.4978 - val_output_3_loss: 0.4120 - val_output_4_loss: 0.4666 - val_output_0_acc: 0.8520 - val_output_1_acc: 0.8480 - val_output_2_acc: 0.8340 - val_output_3_acc: 0.8848 - val_output_4_acc: 0.8365
Epoch 88/200
4000/4000 [==============================] - 1s 168us/step - loss: 1.5965 - output_0_loss: 0.3157 - output_1_loss: 0.3375 - output_2_loss: 0.3148 - output_3_loss: 0.3116 - output_4_loss: 0.3168 - output_0_acc: 0.9145 - output_1_acc: 0.9025 - output_2_acc: 0.9145 - output_3_acc: 0.9160 - output_4_acc: 0.9125 - val_loss: 2.1682 - val_output_0_loss: 0.4445 - val_output_1_loss: 0.4375 - val_output_2_loss: 0.4517 - val_output_3_loss: 0.4485 - val_output_4_loss: 0.3860 - val_output_0_acc: 0.8520 - val_output_1_acc: 0.8647 - val_output_2_acc: 0.8400 - val_output_3_acc: 0.8423 - val_output_4_acc: 0.8962
Epoch 89/200
4000/4000 [==============================] - 1s 169us/step - loss: 1.5517 - output_0_loss: 0.3117 - output_1_loss: 0.3243 - output_2_loss: 0.3053 - output_3_loss: 0.3016 - output_4_loss: 0.3088 - output_0_acc: 0.9173 - output_1_acc: 0.9098 - output_2_acc: 0.9210 - output_3_acc: 0.9235 - output_4_acc: 0.9115 - val_loss: 2.2863 - val_output_0_loss: 0.4659 - val_output_1_loss: 0.5098 - val_output_2_loss: 0.4633 - val_output_3_loss: 0.4160 - val_output_4_loss: 0.4313 - val_output_0_acc: 0.8345 - val_output_1_acc: 0.8152 - val_output_2_acc: 0.8407 - val_output_3_acc: 0.8690 - val_output_4_acc: 0.8590
Epoch 90/200
4000/4000 [==============================] - 1s 164us/step - loss: 1.5393 - output_0_loss: 0.2998 - output_1_loss: 0.3212 - output_2_loss: 0.3036 - output_3_loss: 0.3043 - output_4_loss: 0.3103 - output_0_acc: 0.9207 - output_1_acc: 0.9120 - output_2_acc: 0.9195 - output_3_acc: 0.9170 - output_4_acc: 0.9150 - val_loss: 2.2685 - val_output_0_loss: 0.4341 - val_output_1_loss: 0.5582 - val_output_2_loss: 0.4279 - val_output_3_loss: 0.4195 - val_output_4_loss: 0.4289 - val_output_0_acc: 0.8563 - val_output_1_acc: 0.7843 - val_output_2_acc: 0.8638 - val_output_3_acc: 0.8602 - val_output_4_acc: 0.8585
Epoch 91/200
4000/4000 [==============================] - 1s 166us/step - loss: 1.5146 - output_0_loss: 0.3039 - output_1_loss: 0.3232 - output_2_loss: 0.2966 - output_3_loss: 0.2940 - output_4_loss: 0.2968 - output_0_acc: 0.9227 - output_1_acc: 0.9060 - output_2_acc: 0.9233 - output_3_acc: 0.9230 - output_4_acc: 0.9180 - val_loss: 2.1764 - val_output_0_loss: 0.4551 - val_output_1_loss: 0.4188 - val_output_2_loss: 0.4026 - val_output_3_loss: 0.4794 - val_output_4_loss: 0.4205 - val_output_0_acc: 0.8340 - val_output_1_acc: 0.8690 - val_output_2_acc: 0.8722 - val_output_3_acc: 0.8102 - val_output_4_acc: 0.8633
Epoch 92/200
4000/4000 [==============================] - 1s 166us/step - loss: 1.4949 - output_0_loss: 0.3000 - output_1_loss: 0.3158 - output_2_loss: 0.2897 - output_3_loss: 0.2929 - output_4_loss: 0.2966 - output_0_acc: 0.9167 - output_1_acc: 0.9077 - output_2_acc: 0.9290 - output_3_acc: 0.9207 - output_4_acc: 0.9213 - val_loss: 2.2638 - val_output_0_loss: 0.4462 - val_output_1_loss: 0.5101 - val_output_2_loss: 0.4049 - val_output_3_loss: 0.4022 - val_output_4_loss: 0.5004 - val_output_0_acc: 0.8425 - val_output_1_acc: 0.8052 - val_output_2_acc: 0.8767 - val_output_3_acc: 0.8697 - val_output_4_acc: 0.8088
Epoch 93/200
4000/4000 [==============================] - 1s 162us/step - loss: 1.4725 - output_0_loss: 0.2885 - output_1_loss: 0.3073 - output_2_loss: 0.2900 - output_3_loss: 0.2876 - output_4_loss: 0.2991 - output_0_acc: 0.9253 - output_1_acc: 0.9185 - output_2_acc: 0.9263 - output_3_acc: 0.9248 - output_4_acc: 0.9128 - val_loss: 2.0967 - val_output_0_loss: 0.4599 - val_output_1_loss: 0.4813 - val_output_2_loss: 0.4026 - val_output_3_loss: 0.3646 - val_output_4_loss: 0.3884 - val_output_0_acc: 0.8273 - val_output_1_acc: 0.8303 - val_output_2_acc: 0.8708 - val_output_3_acc: 0.9047 - val_output_4_acc: 0.8783
Epoch 94/200
4000/4000 [==============================] - 1s 162us/step - loss: 1.4298 - output_0_loss: 0.2804 - output_1_loss: 0.3030 - output_2_loss: 0.2854 - output_3_loss: 0.2785 - output_4_loss: 0.2826 - output_0_acc: 0.9275 - output_1_acc: 0.9155 - output_2_acc: 0.9267 - output_3_acc: 0.9287 - output_4_acc: 0.9225 - val_loss: 1.9606 - val_output_0_loss: 0.3933 - val_output_1_loss: 0.4021 - val_output_2_loss: 0.3638 - val_output_3_loss: 0.3828 - val_output_4_loss: 0.4187 - val_output_0_acc: 0.8797 - val_output_1_acc: 0.8792 - val_output_2_acc: 0.9075 - val_output_3_acc: 0.8883 - val_output_4_acc: 0.8753
Epoch 95/200
4000/4000 [==============================] - 1s 161us/step - loss: 1.4179 - output_0_loss: 0.2786 - output_1_loss: 0.3025 - output_2_loss: 0.2762 - output_3_loss: 0.2749 - output_4_loss: 0.2856 - output_0_acc: 0.9263 - output_1_acc: 0.9145 - output_2_acc: 0.9290 - output_3_acc: 0.9343 - output_4_acc: 0.9248 - val_loss: 2.1120 - val_output_0_loss: 0.4476 - val_output_1_loss: 0.4667 - val_output_2_loss: 0.4214 - val_output_3_loss: 0.3927 - val_output_4_loss: 0.3837 - val_output_0_acc: 0.8277 - val_output_1_acc: 0.8330 - val_output_2_acc: 0.8645 - val_output_3_acc: 0.8747 - val_output_4_acc: 0.8810
Epoch 96/200
4000/4000 [==============================] - 1s 178us/step - loss: 1.3953 - output_0_loss: 0.2783 - output_1_loss: 0.2954 - output_2_loss: 0.2729 - output_3_loss: 0.2715 - output_4_loss: 0.2772 - output_0_acc: 0.9245 - output_1_acc: 0.9215 - output_2_acc: 0.9315 - output_3_acc: 0.9343 - output_4_acc: 0.9255 - val_loss: 2.0468 - val_output_0_loss: 0.3820 - val_output_1_loss: 0.4563 - val_output_2_loss: 0.3561 - val_output_3_loss: 0.4139 - val_output_4_loss: 0.4386 - val_output_0_acc: 0.8817 - val_output_1_acc: 0.8362 - val_output_2_acc: 0.9077 - val_output_3_acc: 0.8563 - val_output_4_acc: 0.8405
Epoch 97/200
4000/4000 [==============================] - 1s 185us/step - loss: 1.3692 - output_0_loss: 0.2653 - output_1_loss: 0.2917 - output_2_loss: 0.2671 - output_3_loss: 0.2674 - output_4_loss: 0.2777 - output_0_acc: 0.9373 - output_1_acc: 0.9185 - output_2_acc: 0.9353 - output_3_acc: 0.9387 - output_4_acc: 0.9233 - val_loss: 2.0281 - val_output_0_loss: 0.4446 - val_output_1_loss: 0.4086 - val_output_2_loss: 0.3695 - val_output_3_loss: 0.4038 - val_output_4_loss: 0.4016 - val_output_0_acc: 0.8362 - val_output_1_acc: 0.8665 - val_output_2_acc: 0.9033 - val_output_3_acc: 0.8553 - val_output_4_acc: 0.8690
Epoch 98/200
4000/4000 [==============================] - 1s 193us/step - loss: 1.3477 - output_0_loss: 0.2609 - output_1_loss: 0.2868 - output_2_loss: 0.2614 - output_3_loss: 0.2657 - output_4_loss: 0.2728 - output_0_acc: 0.9347 - output_1_acc: 0.9218 - output_2_acc: 0.9407 - output_3_acc: 0.9343 - output_4_acc: 0.9278 - val_loss: 2.0534 - val_output_0_loss: 0.3831 - val_output_1_loss: 0.4252 - val_output_2_loss: 0.4038 - val_output_3_loss: 0.4202 - val_output_4_loss: 0.4211 - val_output_0_acc: 0.8773 - val_output_1_acc: 0.8607 - val_output_2_acc: 0.8635 - val_output_3_acc: 0.8608 - val_output_4_acc: 0.8505
Epoch 99/200
4000/4000 [==============================] - 1s 196us/step - loss: 1.3222 - output_0_loss: 0.2609 - output_1_loss: 0.2827 - output_2_loss: 0.2572 - output_3_loss: 0.2558 - output_4_loss: 0.2656 - output_0_acc: 0.9357 - output_1_acc: 0.9243 - output_2_acc: 0.9385 - output_3_acc: 0.9400 - output_4_acc: 0.9300 - val_loss: 1.9102 - val_output_0_loss: 0.4068 - val_output_1_loss: 0.4267 - val_output_2_loss: 0.3629 - val_output_3_loss: 0.3385 - val_output_4_loss: 0.3753 - val_output_0_acc: 0.8618 - val_output_1_acc: 0.8597 - val_output_2_acc: 0.8937 - val_output_3_acc: 0.9062 - val_output_4_acc: 0.8798
Epoch 100/200
4000/4000 [==============================] - 1s 188us/step - loss: 1.2938 - output_0_loss: 0.2553 - output_1_loss: 0.2723 - output_2_loss: 0.2518 - output_3_loss: 0.2516 - output_4_loss: 0.2628 - output_0_acc: 0.9368 - output_1_acc: 0.9278 - output_2_acc: 0.9338 - output_3_acc: 0.9407 - output_4_acc: 0.9313 - val_loss: 2.0960 - val_output_0_loss: 0.3857 - val_output_1_loss: 0.4852 - val_output_2_loss: 0.3920 - val_output_3_loss: 0.3921 - val_output_4_loss: 0.4410 - val_output_0_acc: 0.8807 - val_output_1_acc: 0.8125 - val_output_2_acc: 0.8867 - val_output_3_acc: 0.8692 - val_output_4_acc: 0.8412
Epoch 101/200
4000/4000 [==============================] - 1s 172us/step - loss: 1.2793 - output_0_loss: 0.2500 - output_1_loss: 0.2708 - output_2_loss: 0.2456 - output_3_loss: 0.2501 - output_4_loss: 0.2628 - output_0_acc: 0.9387 - output_1_acc: 0.9295 - output_2_acc: 0.9475 - output_3_acc: 0.9400 - output_4_acc: 0.9310 - val_loss: 1.9586 - val_output_0_loss: 0.4505 - val_output_1_loss: 0.3792 - val_output_2_loss: 0.3646 - val_output_3_loss: 0.4146 - val_output_4_loss: 0.3497 - val_output_0_acc: 0.8240 - val_output_1_acc: 0.8895 - val_output_2_acc: 0.8907 - val_output_3_acc: 0.8388 - val_output_4_acc: 0.9057
Epoch 102/200
4000/4000 [==============================] - 1s 169us/step - loss: 1.2573 - output_0_loss: 0.2482 - output_1_loss: 0.2689 - output_2_loss: 0.2464 - output_3_loss: 0.2393 - output_4_loss: 0.2545 - output_0_acc: 0.9395 - output_1_acc: 0.9273 - output_2_acc: 0.9417 - output_3_acc: 0.9437 - output_4_acc: 0.9368 - val_loss: 1.8702 - val_output_0_loss: 0.4371 - val_output_1_loss: 0.3755 - val_output_2_loss: 0.3360 - val_output_3_loss: 0.3462 - val_output_4_loss: 0.3755 - val_output_0_acc: 0.8368 - val_output_1_acc: 0.8837 - val_output_2_acc: 0.9113 - val_output_3_acc: 0.8982 - val_output_4_acc: 0.8838
Epoch 103/200
4000/4000 [==============================] - 1s 170us/step - loss: 1.2348 - output_0_loss: 0.2358 - output_1_loss: 0.2647 - output_2_loss: 0.2426 - output_3_loss: 0.2446 - output_4_loss: 0.2471 - output_0_acc: 0.9445 - output_1_acc: 0.9248 - output_2_acc: 0.9442 - output_3_acc: 0.9417 - output_4_acc: 0.9345 - val_loss: 1.8800 - val_output_0_loss: 0.3694 - val_output_1_loss: 0.4284 - val_output_2_loss: 0.3639 - val_output_3_loss: 0.3188 - val_output_4_loss: 0.3995 - val_output_0_acc: 0.8833 - val_output_1_acc: 0.8512 - val_output_2_acc: 0.8887 - val_output_3_acc: 0.9178 - val_output_4_acc: 0.8597
Epoch 104/200
4000/4000 [==============================] - 1s 169us/step - loss: 1.2162 - output_0_loss: 0.2421 - output_1_loss: 0.2561 - output_2_loss: 0.2370 - output_3_loss: 0.2364 - output_4_loss: 0.2447 - output_0_acc: 0.9382 - output_1_acc: 0.9353 - output_2_acc: 0.9435 - output_3_acc: 0.9430 - output_4_acc: 0.9377 - val_loss: 1.8364 - val_output_0_loss: 0.3745 - val_output_1_loss: 0.4107 - val_output_2_loss: 0.3676 - val_output_3_loss: 0.3273 - val_output_4_loss: 0.3562 - val_output_0_acc: 0.8775 - val_output_1_acc: 0.8695 - val_output_2_acc: 0.8863 - val_output_3_acc: 0.9040 - val_output_4_acc: 0.8835
Epoch 105/200
4000/4000 [==============================] - 1s 193us/step - loss: 1.1944 - output_0_loss: 0.2326 - output_1_loss: 0.2575 - output_2_loss: 0.2304 - output_3_loss: 0.2322 - output_4_loss: 0.2417 - output_0_acc: 0.9455 - output_1_acc: 0.9353 - output_2_acc: 0.9450 - output_3_acc: 0.9412 - output_4_acc: 0.9353 - val_loss: 1.8544 - val_output_0_loss: 0.3818 - val_output_1_loss: 0.3756 - val_output_2_loss: 0.3802 - val_output_3_loss: 0.2853 - val_output_4_loss: 0.4316 - val_output_0_acc: 0.8682 - val_output_1_acc: 0.8780 - val_output_2_acc: 0.8632 - val_output_3_acc: 0.9387 - val_output_4_acc: 0.8405
Epoch 106/200
4000/4000 [==============================] - 1s 193us/step - loss: 1.1781 - output_0_loss: 0.2306 - output_1_loss: 0.2499 - output_2_loss: 0.2320 - output_3_loss: 0.2227 - output_4_loss: 0.2429 - output_0_acc: 0.9417 - output_1_acc: 0.9340 - output_2_acc: 0.9483 - output_3_acc: 0.9490 - output_4_acc: 0.9375 - val_loss: 1.7630 - val_output_0_loss: 0.3356 - val_output_1_loss: 0.3810 - val_output_2_loss: 0.3234 - val_output_3_loss: 0.3157 - val_output_4_loss: 0.4073 - val_output_0_acc: 0.9013 - val_output_1_acc: 0.8773 - val_output_2_acc: 0.9082 - val_output_3_acc: 0.9150 - val_output_4_acc: 0.8522
Epoch 107/200
4000/4000 [==============================] - 1s 191us/step - loss: 1.1485 - output_0_loss: 0.2242 - output_1_loss: 0.2494 - output_2_loss: 0.2210 - output_3_loss: 0.2227 - output_4_loss: 0.2312 - output_0_acc: 0.9450 - output_1_acc: 0.9327 - output_2_acc: 0.9500 - output_3_acc: 0.9487 - output_4_acc: 0.9407 - val_loss: 1.8443 - val_output_0_loss: 0.3495 - val_output_1_loss: 0.4184 - val_output_2_loss: 0.3785 - val_output_3_loss: 0.3418 - val_output_4_loss: 0.3561 - val_output_0_acc: 0.8907 - val_output_1_acc: 0.8482 - val_output_2_acc: 0.8592 - val_output_3_acc: 0.8882 - val_output_4_acc: 0.8925
Epoch 108/200
4000/4000 [==============================] - 1s 183us/step - loss: 1.1246 - output_0_loss: 0.2250 - output_1_loss: 0.2402 - output_2_loss: 0.2166 - output_3_loss: 0.2098 - output_4_loss: 0.2330 - output_0_acc: 0.9437 - output_1_acc: 0.9390 - output_2_acc: 0.9520 - output_3_acc: 0.9532 - output_4_acc: 0.9390 - val_loss: 1.8168 - val_output_0_loss: 0.3268 - val_output_1_loss: 0.4072 - val_output_2_loss: 0.3538 - val_output_3_loss: 0.3358 - val_output_4_loss: 0.3932 - val_output_0_acc: 0.9043 - val_output_1_acc: 0.8497 - val_output_2_acc: 0.8852 - val_output_3_acc: 0.8898 - val_output_4_acc: 0.8592
Epoch 109/200
4000/4000 [==============================] - 1s 178us/step - loss: 1.1063 - output_0_loss: 0.2178 - output_1_loss: 0.2349 - output_2_loss: 0.2151 - output_3_loss: 0.2118 - output_4_loss: 0.2267 - output_0_acc: 0.9460 - output_1_acc: 0.9433 - output_2_acc: 0.9500 - output_3_acc: 0.9525 - output_4_acc: 0.9400 - val_loss: 1.8879 - val_output_0_loss: 0.3421 - val_output_1_loss: 0.3513 - val_output_2_loss: 0.4083 - val_output_3_loss: 0.4341 - val_output_4_loss: 0.3522 - val_output_0_acc: 0.9015 - val_output_1_acc: 0.8970 - val_output_2_acc: 0.8567 - val_output_3_acc: 0.8295 - val_output_4_acc: 0.8938
Epoch 110/200
4000/4000 [==============================] - 1s 164us/step - loss: 1.1065 - output_0_loss: 0.2165 - output_1_loss: 0.2331 - output_2_loss: 0.2218 - output_3_loss: 0.2166 - output_4_loss: 0.2184 - output_0_acc: 0.9480 - output_1_acc: 0.9393 - output_2_acc: 0.9425 - output_3_acc: 0.9527 - output_4_acc: 0.9470 - val_loss: 1.7594 - val_output_0_loss: 0.3331 - val_output_1_loss: 0.3615 - val_output_2_loss: 0.3322 - val_output_3_loss: 0.3123 - val_output_4_loss: 0.4204 - val_output_0_acc: 0.8947 - val_output_1_acc: 0.8782 - val_output_2_acc: 0.8982 - val_output_3_acc: 0.9162 - val_output_4_acc: 0.8485
Epoch 111/200
4000/4000 [==============================] - 1s 166us/step - loss: 1.0722 - output_0_loss: 0.2099 - output_1_loss: 0.2278 - output_2_loss: 0.2076 - output_3_loss: 0.2059 - output_4_loss: 0.2209 - output_0_acc: 0.9535 - output_1_acc: 0.9400 - output_2_acc: 0.9562 - output_3_acc: 0.9577 - output_4_acc: 0.9450 - val_loss: 1.7908 - val_output_0_loss: 0.4143 - val_output_1_loss: 0.3438 - val_output_2_loss: 0.3663 - val_output_3_loss: 0.3025 - val_output_4_loss: 0.3638 - val_output_0_acc: 0.8462 - val_output_1_acc: 0.8943 - val_output_2_acc: 0.8825 - val_output_3_acc: 0.9108 - val_output_4_acc: 0.8733
Epoch 112/200
4000/4000 [==============================] - 1s 175us/step - loss: 1.0650 - output_0_loss: 0.2087 - output_1_loss: 0.2291 - output_2_loss: 0.2021 - output_3_loss: 0.2083 - output_4_loss: 0.2168 - output_0_acc: 0.9497 - output_1_acc: 0.9405 - output_2_acc: 0.9585 - output_3_acc: 0.9550 - output_4_acc: 0.9410 - val_loss: 1.6923 - val_output_0_loss: 0.3515 - val_output_1_loss: 0.3114 - val_output_2_loss: 0.3433 - val_output_3_loss: 0.3119 - val_output_4_loss: 0.3743 - val_output_0_acc: 0.8910 - val_output_1_acc: 0.9113 - val_output_2_acc: 0.8903 - val_output_3_acc: 0.9057 - val_output_4_acc: 0.8690
Epoch 113/200
4000/4000 [==============================] - 1s 174us/step - loss: 1.0433 - output_0_loss: 0.2022 - output_1_loss: 0.2261 - output_2_loss: 0.2039 - output_3_loss: 0.1979 - output_4_loss: 0.2132 - output_0_acc: 0.9543 - output_1_acc: 0.9457 - output_2_acc: 0.9535 - output_3_acc: 0.9590 - output_4_acc: 0.9437 - val_loss: 1.6641 - val_output_0_loss: 0.2980 - val_output_1_loss: 0.3858 - val_output_2_loss: 0.3263 - val_output_3_loss: 0.3098 - val_output_4_loss: 0.3443 - val_output_0_acc: 0.9162 - val_output_1_acc: 0.8740 - val_output_2_acc: 0.8928 - val_output_3_acc: 0.9033 - val_output_4_acc: 0.8900
Epoch 114/200
4000/4000 [==============================] - 1s 176us/step - loss: 1.0292 - output_0_loss: 0.1985 - output_1_loss: 0.2229 - output_2_loss: 0.1943 - output_3_loss: 0.2008 - output_4_loss: 0.2127 - output_0_acc: 0.9580 - output_1_acc: 0.9440 - output_2_acc: 0.9603 - output_3_acc: 0.9550 - output_4_acc: 0.9465 - val_loss: 1.6615 - val_output_0_loss: 0.3009 - val_output_1_loss: 0.3467 - val_output_2_loss: 0.3647 - val_output_3_loss: 0.2852 - val_output_4_loss: 0.3639 - val_output_0_acc: 0.9097 - val_output_1_acc: 0.8960 - val_output_2_acc: 0.8707 - val_output_3_acc: 0.9260 - val_output_4_acc: 0.8820
Epoch 115/200
4000/4000 [==============================] - 1s 178us/step - loss: 1.0193 - output_0_loss: 0.2006 - output_1_loss: 0.2140 - output_2_loss: 0.2025 - output_3_loss: 0.1961 - output_4_loss: 0.2061 - output_0_acc: 0.9550 - output_1_acc: 0.9483 - output_2_acc: 0.9550 - output_3_acc: 0.9543 - output_4_acc: 0.9465 - val_loss: 1.5922 - val_output_0_loss: 0.3187 - val_output_1_loss: 0.3487 - val_output_2_loss: 0.2498 - val_output_3_loss: 0.3146 - val_output_4_loss: 0.3604 - val_output_0_acc: 0.9102 - val_output_1_acc: 0.8925 - val_output_2_acc: 0.9512 - val_output_3_acc: 0.8965 - val_output_4_acc: 0.8688
Epoch 116/200
4000/4000 [==============================] - 1s 180us/step - loss: 0.9856 - output_0_loss: 0.1945 - output_1_loss: 0.2179 - output_2_loss: 0.1881 - output_3_loss: 0.1848 - output_4_loss: 0.2002 - output_0_acc: 0.9587 - output_1_acc: 0.9447 - output_2_acc: 0.9625 - output_3_acc: 0.9620 - output_4_acc: 0.9513 - val_loss: 1.7453 - val_output_0_loss: 0.3409 - val_output_1_loss: 0.4405 - val_output_2_loss: 0.2912 - val_output_3_loss: 0.3171 - val_output_4_loss: 0.3555 - val_output_0_acc: 0.8882 - val_output_1_acc: 0.8322 - val_output_2_acc: 0.9162 - val_output_3_acc: 0.9007 - val_output_4_acc: 0.8868
Epoch 117/200
4000/4000 [==============================] - 1s 172us/step - loss: 0.9727 - output_0_loss: 0.1864 - output_1_loss: 0.2071 - output_2_loss: 0.1929 - output_3_loss: 0.1851 - output_4_loss: 0.2013 - output_0_acc: 0.9615 - output_1_acc: 0.9487 - output_2_acc: 0.9570 - output_3_acc: 0.9585 - output_4_acc: 0.9477 - val_loss: 1.6207 - val_output_0_loss: 0.3313 - val_output_1_loss: 0.3872 - val_output_2_loss: 0.3444 - val_output_3_loss: 0.2693 - val_output_4_loss: 0.2885 - val_output_0_acc: 0.8957 - val_output_1_acc: 0.8522 - val_output_2_acc: 0.8795 - val_output_3_acc: 0.9305 - val_output_4_acc: 0.9192
Epoch 118/200
4000/4000 [==============================] - 1s 164us/step - loss: 0.9646 - output_0_loss: 0.1869 - output_1_loss: 0.2093 - output_2_loss: 0.1894 - output_3_loss: 0.1830 - output_4_loss: 0.1960 - output_0_acc: 0.9575 - output_1_acc: 0.9470 - output_2_acc: 0.9585 - output_3_acc: 0.9610 - output_4_acc: 0.9513 - val_loss: 1.6489 - val_output_0_loss: 0.3491 - val_output_1_loss: 0.3532 - val_output_2_loss: 0.2828 - val_output_3_loss: 0.3048 - val_output_4_loss: 0.3589 - val_output_0_acc: 0.8855 - val_output_1_acc: 0.8863 - val_output_2_acc: 0.9202 - val_output_3_acc: 0.9012 - val_output_4_acc: 0.8690
Epoch 119/200
4000/4000 [==============================] - 1s 173us/step - loss: 0.9536 - output_0_loss: 0.1853 - output_1_loss: 0.2075 - output_2_loss: 0.1847 - output_3_loss: 0.1813 - output_4_loss: 0.1948 - output_0_acc: 0.9577 - output_1_acc: 0.9472 - output_2_acc: 0.9615 - output_3_acc: 0.9617 - output_4_acc: 0.9495 - val_loss: 1.6825 - val_output_0_loss: 0.3524 - val_output_1_loss: 0.3155 - val_output_2_loss: 0.3000 - val_output_3_loss: 0.4134 - val_output_4_loss: 0.3012 - val_output_0_acc: 0.8718 - val_output_1_acc: 0.9018 - val_output_2_acc: 0.8925 - val_output_3_acc: 0.8400 - val_output_4_acc: 0.9090
Epoch 120/200
4000/4000 [==============================] - 1s 180us/step - loss: 0.9319 - output_0_loss: 0.1816 - output_1_loss: 0.2002 - output_2_loss: 0.1777 - output_3_loss: 0.1832 - output_4_loss: 0.1891 - output_0_acc: 0.9585 - output_1_acc: 0.9500 - output_2_acc: 0.9617 - output_3_acc: 0.9567 - output_4_acc: 0.9523 - val_loss: 1.5461 - val_output_0_loss: 0.3332 - val_output_1_loss: 0.3485 - val_output_2_loss: 0.2870 - val_output_3_loss: 0.2804 - val_output_4_loss: 0.2969 - val_output_0_acc: 0.8810 - val_output_1_acc: 0.8805 - val_output_2_acc: 0.9182 - val_output_3_acc: 0.9165 - val_output_4_acc: 0.9132
Epoch 121/200
4000/4000 [==============================] - 1s 175us/step - loss: 0.9082 - output_0_loss: 0.1730 - output_1_loss: 0.2003 - output_2_loss: 0.1755 - output_3_loss: 0.1775 - output_4_loss: 0.1819 - output_0_acc: 0.9633 - output_1_acc: 0.9487 - output_2_acc: 0.9645 - output_3_acc: 0.9633 - output_4_acc: 0.9570 - val_loss: 1.6015 - val_output_0_loss: 0.2919 - val_output_1_loss: 0.3514 - val_output_2_loss: 0.3485 - val_output_3_loss: 0.2945 - val_output_4_loss: 0.3152 - val_output_0_acc: 0.9073 - val_output_1_acc: 0.8743 - val_output_2_acc: 0.8743 - val_output_3_acc: 0.9092 - val_output_4_acc: 0.9000
Epoch 122/200
4000/4000 [==============================] - 1s 164us/step - loss: 0.8884 - output_0_loss: 0.1739 - output_1_loss: 0.1949 - output_2_loss: 0.1684 - output_3_loss: 0.1719 - output_4_loss: 0.1793 - output_0_acc: 0.9603 - output_1_acc: 0.9537 - output_2_acc: 0.9652 - output_3_acc: 0.9675 - output_4_acc: 0.9590 - val_loss: 1.5069 - val_output_0_loss: 0.3278 - val_output_1_loss: 0.2865 - val_output_2_loss: 0.3086 - val_output_3_loss: 0.2579 - val_output_4_loss: 0.3262 - val_output_0_acc: 0.8863 - val_output_1_acc: 0.9205 - val_output_2_acc: 0.9078 - val_output_3_acc: 0.9283 - val_output_4_acc: 0.8903
Epoch 123/200
4000/4000 [==============================] - 1s 165us/step - loss: 0.8731 - output_0_loss: 0.1692 - output_1_loss: 0.1926 - output_2_loss: 0.1677 - output_3_loss: 0.1633 - output_4_loss: 0.1803 - output_0_acc: 0.9650 - output_1_acc: 0.9505 - output_2_acc: 0.9670 - output_3_acc: 0.9677 - output_4_acc: 0.9547 - val_loss: 1.4158 - val_output_0_loss: 0.2811 - val_output_1_loss: 0.3078 - val_output_2_loss: 0.2979 - val_output_3_loss: 0.2440 - val_output_4_loss: 0.2851 - val_output_0_acc: 0.9217 - val_output_1_acc: 0.9042 - val_output_2_acc: 0.8965 - val_output_3_acc: 0.9357 - val_output_4_acc: 0.9188
Epoch 124/200
4000/4000 [==============================] - 1s 175us/step - loss: 0.8634 - output_0_loss: 0.1683 - output_1_loss: 0.1877 - output_2_loss: 0.1629 - output_3_loss: 0.1690 - output_4_loss: 0.1755 - output_0_acc: 0.9615 - output_1_acc: 0.9532 - output_2_acc: 0.9670 - output_3_acc: 0.9607 - output_4_acc: 0.9592 - val_loss: 1.5615 - val_output_0_loss: 0.3441 - val_output_1_loss: 0.3378 - val_output_2_loss: 0.3271 - val_output_3_loss: 0.2626 - val_output_4_loss: 0.2898 - val_output_0_acc: 0.8788 - val_output_1_acc: 0.8813 - val_output_2_acc: 0.8970 - val_output_3_acc: 0.9267 - val_output_4_acc: 0.9060
Epoch 125/200
4000/4000 [==============================] - 1s 163us/step - loss: 0.8445 - output_0_loss: 0.1627 - output_1_loss: 0.1850 - output_2_loss: 0.1615 - output_3_loss: 0.1632 - output_4_loss: 0.1721 - output_0_acc: 0.9667 - output_1_acc: 0.9540 - output_2_acc: 0.9657 - output_3_acc: 0.9645 - output_4_acc: 0.9583 - val_loss: 1.4510 - val_output_0_loss: 0.3160 - val_output_1_loss: 0.3695 - val_output_2_loss: 0.2526 - val_output_3_loss: 0.2673 - val_output_4_loss: 0.2458 - val_output_0_acc: 0.8862 - val_output_1_acc: 0.8595 - val_output_2_acc: 0.9303 - val_output_3_acc: 0.9242 - val_output_4_acc: 0.9392
Epoch 126/200
4000/4000 [==============================] - 1s 167us/step - loss: 0.8393 - output_0_loss: 0.1571 - output_1_loss: 0.1819 - output_2_loss: 0.1620 - output_3_loss: 0.1622 - output_4_loss: 0.1760 - output_0_acc: 0.9708 - output_1_acc: 0.9585 - output_2_acc: 0.9677 - output_3_acc: 0.9688 - output_4_acc: 0.9555 - val_loss: 1.3227 - val_output_0_loss: 0.2903 - val_output_1_loss: 0.2907 - val_output_2_loss: 0.2229 - val_output_3_loss: 0.2231 - val_output_4_loss: 0.2955 - val_output_0_acc: 0.9057 - val_output_1_acc: 0.9157 - val_output_2_acc: 0.9522 - val_output_3_acc: 0.9523 - val_output_4_acc: 0.9115
Epoch 127/200
4000/4000 [==============================] - 1s 174us/step - loss: 0.8226 - output_0_loss: 0.1620 - output_1_loss: 0.1796 - output_2_loss: 0.1561 - output_3_loss: 0.1577 - output_4_loss: 0.1672 - output_0_acc: 0.9657 - output_1_acc: 0.9553 - output_2_acc: 0.9702 - output_3_acc: 0.9680 - output_4_acc: 0.9600 - val_loss: 1.3558 - val_output_0_loss: 0.2726 - val_output_1_loss: 0.3058 - val_output_2_loss: 0.2271 - val_output_3_loss: 0.2383 - val_output_4_loss: 0.3120 - val_output_0_acc: 0.9192 - val_output_1_acc: 0.9043 - val_output_2_acc: 0.9458 - val_output_3_acc: 0.9368 - val_output_4_acc: 0.9050
Epoch 128/200
4000/4000 [==============================] - 1s 164us/step - loss: 0.8060 - output_0_loss: 0.1548 - output_1_loss: 0.1734 - output_2_loss: 0.1578 - output_3_loss: 0.1568 - output_4_loss: 0.1632 - output_0_acc: 0.9708 - output_1_acc: 0.9605 - output_2_acc: 0.9663 - output_3_acc: 0.9685 - output_4_acc: 0.9620 - val_loss: 1.4040 - val_output_0_loss: 0.2701 - val_output_1_loss: 0.3364 - val_output_2_loss: 0.2463 - val_output_3_loss: 0.2726 - val_output_4_loss: 0.2786 - val_output_0_acc: 0.9138 - val_output_1_acc: 0.8742 - val_output_2_acc: 0.9340 - val_output_3_acc: 0.9113 - val_output_4_acc: 0.9088
Epoch 129/200
4000/4000 [==============================] - 1s 163us/step - loss: 0.7888 - output_0_loss: 0.1534 - output_1_loss: 0.1724 - output_2_loss: 0.1502 - output_3_loss: 0.1522 - output_4_loss: 0.1606 - output_0_acc: 0.9693 - output_1_acc: 0.9573 - output_2_acc: 0.9665 - output_3_acc: 0.9695 - output_4_acc: 0.9655 - val_loss: 1.5896 - val_output_0_loss: 0.3242 - val_output_1_loss: 0.3357 - val_output_2_loss: 0.3568 - val_output_3_loss: 0.2622 - val_output_4_loss: 0.3107 - val_output_0_acc: 0.8920 - val_output_1_acc: 0.8823 - val_output_2_acc: 0.8737 - val_output_3_acc: 0.9235 - val_output_4_acc: 0.8905
Epoch 130/200
4000/4000 [==============================] - 1s 175us/step - loss: 0.7870 - output_0_loss: 0.1520 - output_1_loss: 0.1715 - output_2_loss: 0.1488 - output_3_loss: 0.1506 - output_4_loss: 0.1642 - output_0_acc: 0.9645 - output_1_acc: 0.9567 - output_2_acc: 0.9715 - output_3_acc: 0.9728 - output_4_acc: 0.9607 - val_loss: 1.2812 - val_output_0_loss: 0.2543 - val_output_1_loss: 0.3141 - val_output_2_loss: 0.2475 - val_output_3_loss: 0.2165 - val_output_4_loss: 0.2488 - val_output_0_acc: 0.9320 - val_output_1_acc: 0.9002 - val_output_2_acc: 0.9285 - val_output_3_acc: 0.9503 - val_output_4_acc: 0.9295
Epoch 131/200
4000/4000 [==============================] - 1s 174us/step - loss: 0.7633 - output_0_loss: 0.1496 - output_1_loss: 0.1674 - output_2_loss: 0.1493 - output_3_loss: 0.1422 - output_4_loss: 0.1547 - output_0_acc: 0.9682 - output_1_acc: 0.9633 - output_2_acc: 0.9715 - output_3_acc: 0.9753 - output_4_acc: 0.9635 - val_loss: 1.4223 - val_output_0_loss: 0.2502 - val_output_1_loss: 0.3645 - val_output_2_loss: 0.2849 - val_output_3_loss: 0.2499 - val_output_4_loss: 0.2729 - val_output_0_acc: 0.9278 - val_output_1_acc: 0.8665 - val_output_2_acc: 0.9210 - val_output_3_acc: 0.9230 - val_output_4_acc: 0.9198
Epoch 132/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.7509 - output_0_loss: 0.1474 - output_1_loss: 0.1598 - output_2_loss: 0.1413 - output_3_loss: 0.1465 - output_4_loss: 0.1559 - output_0_acc: 0.9680 - output_1_acc: 0.9607 - output_2_acc: 0.9720 - output_3_acc: 0.9700 - output_4_acc: 0.9627 - val_loss: 1.2604 - val_output_0_loss: 0.2210 - val_output_1_loss: 0.2938 - val_output_2_loss: 0.2374 - val_output_3_loss: 0.2169 - val_output_4_loss: 0.2913 - val_output_0_acc: 0.9472 - val_output_1_acc: 0.9045 - val_output_2_acc: 0.9347 - val_output_3_acc: 0.9510 - val_output_4_acc: 0.9030
Epoch 133/200
4000/4000 [==============================] - 1s 179us/step - loss: 0.7431 - output_0_loss: 0.1428 - output_1_loss: 0.1616 - output_2_loss: 0.1448 - output_3_loss: 0.1433 - output_4_loss: 0.1505 - output_0_acc: 0.9718 - output_1_acc: 0.9617 - output_2_acc: 0.9680 - output_3_acc: 0.9720 - output_4_acc: 0.9643 - val_loss: 1.2562 - val_output_0_loss: 0.2755 - val_output_1_loss: 0.2702 - val_output_2_loss: 0.2126 - val_output_3_loss: 0.2506 - val_output_4_loss: 0.2474 - val_output_0_acc: 0.9125 - val_output_1_acc: 0.9087 - val_output_2_acc: 0.9445 - val_output_3_acc: 0.9265 - val_output_4_acc: 0.9267
Epoch 134/200
4000/4000 [==============================] - 1s 183us/step - loss: 0.7279 - output_0_loss: 0.1381 - output_1_loss: 0.1630 - output_2_loss: 0.1374 - output_3_loss: 0.1375 - output_4_loss: 0.1519 - output_0_acc: 0.9712 - output_1_acc: 0.9605 - output_2_acc: 0.9702 - output_3_acc: 0.9762 - output_4_acc: 0.9647 - val_loss: 1.2451 - val_output_0_loss: 0.2262 - val_output_1_loss: 0.2816 - val_output_2_loss: 0.2355 - val_output_3_loss: 0.2281 - val_output_4_loss: 0.2737 - val_output_0_acc: 0.9362 - val_output_1_acc: 0.9100 - val_output_2_acc: 0.9352 - val_output_3_acc: 0.9423 - val_output_4_acc: 0.9062
Epoch 135/200
4000/4000 [==============================] - 1s 166us/step - loss: 0.7153 - output_0_loss: 0.1403 - output_1_loss: 0.1580 - output_2_loss: 0.1379 - output_3_loss: 0.1350 - output_4_loss: 0.1442 - output_0_acc: 0.9730 - output_1_acc: 0.9580 - output_2_acc: 0.9735 - output_3_acc: 0.9735 - output_4_acc: 0.9695 - val_loss: 1.2470 - val_output_0_loss: 0.1917 - val_output_1_loss: 0.2462 - val_output_2_loss: 0.2382 - val_output_3_loss: 0.2777 - val_output_4_loss: 0.2932 - val_output_0_acc: 0.9558 - val_output_1_acc: 0.9337 - val_output_2_acc: 0.9320 - val_output_3_acc: 0.9150 - val_output_4_acc: 0.9090
Epoch 136/200
4000/4000 [==============================] - 1s 174us/step - loss: 0.6954 - output_0_loss: 0.1339 - output_1_loss: 0.1592 - output_2_loss: 0.1295 - output_3_loss: 0.1288 - output_4_loss: 0.1439 - output_0_acc: 0.9720 - output_1_acc: 0.9615 - output_2_acc: 0.9748 - output_3_acc: 0.9768 - output_4_acc: 0.9680 - val_loss: 1.3566 - val_output_0_loss: 0.2429 - val_output_1_loss: 0.3050 - val_output_2_loss: 0.2715 - val_output_3_loss: 0.2214 - val_output_4_loss: 0.3159 - val_output_0_acc: 0.9325 - val_output_1_acc: 0.8957 - val_output_2_acc: 0.9098 - val_output_3_acc: 0.9403 - val_output_4_acc: 0.8942
Epoch 137/200
4000/4000 [==============================] - 1s 177us/step - loss: 0.6836 - output_0_loss: 0.1299 - output_1_loss: 0.1527 - output_2_loss: 0.1287 - output_3_loss: 0.1299 - output_4_loss: 0.1424 - output_0_acc: 0.9753 - output_1_acc: 0.9635 - output_2_acc: 0.9772 - output_3_acc: 0.9762 - output_4_acc: 0.9698 - val_loss: 1.2494 - val_output_0_loss: 0.2626 - val_output_1_loss: 0.2664 - val_output_2_loss: 0.2180 - val_output_3_loss: 0.2857 - val_output_4_loss: 0.2167 - val_output_0_acc: 0.9163 - val_output_1_acc: 0.9123 - val_output_2_acc: 0.9465 - val_output_3_acc: 0.9020 - val_output_4_acc: 0.9437
Epoch 138/200
4000/4000 [==============================] - 1s 171us/step - loss: 0.6768 - output_0_loss: 0.1294 - output_1_loss: 0.1515 - output_2_loss: 0.1277 - output_3_loss: 0.1283 - output_4_loss: 0.1399 - output_0_acc: 0.9750 - output_1_acc: 0.9645 - output_2_acc: 0.9760 - output_3_acc: 0.9765 - output_4_acc: 0.9677 - val_loss: 1.2052 - val_output_0_loss: 0.2393 - val_output_1_loss: 0.2856 - val_output_2_loss: 0.2429 - val_output_3_loss: 0.2116 - val_output_4_loss: 0.2258 - val_output_0_acc: 0.9325 - val_output_1_acc: 0.9088 - val_output_2_acc: 0.9233 - val_output_3_acc: 0.9443 - val_output_4_acc: 0.9373
Epoch 139/200
4000/4000 [==============================] - 1s 174us/step - loss: 0.6688 - output_0_loss: 0.1288 - output_1_loss: 0.1495 - output_2_loss: 0.1248 - output_3_loss: 0.1256 - output_4_loss: 0.1402 - output_0_acc: 0.9742 - output_1_acc: 0.9647 - output_2_acc: 0.9742 - output_3_acc: 0.9783 - output_4_acc: 0.9667 - val_loss: 1.0945 - val_output_0_loss: 0.2471 - val_output_1_loss: 0.2548 - val_output_2_loss: 0.2101 - val_output_3_loss: 0.1873 - val_output_4_loss: 0.1951 - val_output_0_acc: 0.9315 - val_output_1_acc: 0.9158 - val_output_2_acc: 0.9457 - val_output_3_acc: 0.9612 - val_output_4_acc: 0.9548
Epoch 140/200
4000/4000 [==============================] - 1s 167us/step - loss: 0.6492 - output_0_loss: 0.1245 - output_1_loss: 0.1465 - output_2_loss: 0.1228 - output_3_loss: 0.1217 - output_4_loss: 0.1337 - output_0_acc: 0.9738 - output_1_acc: 0.9652 - output_2_acc: 0.9783 - output_3_acc: 0.9788 - output_4_acc: 0.9690 - val_loss: 1.3169 - val_output_0_loss: 0.2251 - val_output_1_loss: 0.3101 - val_output_2_loss: 0.2196 - val_output_3_loss: 0.3335 - val_output_4_loss: 0.2286 - val_output_0_acc: 0.9368 - val_output_1_acc: 0.8955 - val_output_2_acc: 0.9480 - val_output_3_acc: 0.8683 - val_output_4_acc: 0.9368
Epoch 141/200
4000/4000 [==============================] - 1s 163us/step - loss: 0.6454 - output_0_loss: 0.1235 - output_1_loss: 0.1409 - output_2_loss: 0.1228 - output_3_loss: 0.1240 - output_4_loss: 0.1342 - output_0_acc: 0.9730 - output_1_acc: 0.9682 - output_2_acc: 0.9772 - output_3_acc: 0.9765 - output_4_acc: 0.9673 - val_loss: 1.2748 - val_output_0_loss: 0.2606 - val_output_1_loss: 0.2798 - val_output_2_loss: 0.2415 - val_output_3_loss: 0.2430 - val_output_4_loss: 0.2498 - val_output_0_acc: 0.9167 - val_output_1_acc: 0.9047 - val_output_2_acc: 0.9192 - val_output_3_acc: 0.9217 - val_output_4_acc: 0.9163
Epoch 142/200
4000/4000 [==============================] - 1s 176us/step - loss: 0.6290 - output_0_loss: 0.1174 - output_1_loss: 0.1419 - output_2_loss: 0.1192 - output_3_loss: 0.1177 - output_4_loss: 0.1328 - output_0_acc: 0.9813 - output_1_acc: 0.9667 - output_2_acc: 0.9755 - output_3_acc: 0.9798 - output_4_acc: 0.9695 - val_loss: 1.1362 - val_output_0_loss: 0.1928 - val_output_1_loss: 0.2963 - val_output_2_loss: 0.2173 - val_output_3_loss: 0.1915 - val_output_4_loss: 0.2383 - val_output_0_acc: 0.9493 - val_output_1_acc: 0.8950 - val_output_2_acc: 0.9380 - val_output_3_acc: 0.9585 - val_output_4_acc: 0.9233
Epoch 143/200
4000/4000 [==============================] - 1s 191us/step - loss: 0.6290 - output_0_loss: 0.1200 - output_1_loss: 0.1397 - output_2_loss: 0.1189 - output_3_loss: 0.1205 - output_4_loss: 0.1300 - output_0_acc: 0.9753 - output_1_acc: 0.9647 - output_2_acc: 0.9775 - output_3_acc: 0.9768 - output_4_acc: 0.9682 - val_loss: 1.1672 - val_output_0_loss: 0.2686 - val_output_1_loss: 0.2391 - val_output_2_loss: 0.2256 - val_output_3_loss: 0.2123 - val_output_4_loss: 0.2216 - val_output_0_acc: 0.9083 - val_output_1_acc: 0.9292 - val_output_2_acc: 0.9280 - val_output_3_acc: 0.9360 - val_output_4_acc: 0.9333
Epoch 144/200
4000/4000 [==============================] - 1s 180us/step - loss: 0.6101 - output_0_loss: 0.1159 - output_1_loss: 0.1393 - output_2_loss: 0.1110 - output_3_loss: 0.1160 - output_4_loss: 0.1278 - output_0_acc: 0.9770 - output_1_acc: 0.9700 - output_2_acc: 0.9828 - output_3_acc: 0.9772 - output_4_acc: 0.9698 - val_loss: 1.0931 - val_output_0_loss: 0.1861 - val_output_1_loss: 0.2859 - val_output_2_loss: 0.2076 - val_output_3_loss: 0.1860 - val_output_4_loss: 0.2274 - val_output_0_acc: 0.9545 - val_output_1_acc: 0.9072 - val_output_2_acc: 0.9435 - val_output_3_acc: 0.9542 - val_output_4_acc: 0.9367
Epoch 145/200
4000/4000 [==============================] - 1s 190us/step - loss: 0.5983 - output_0_loss: 0.1151 - output_1_loss: 0.1317 - output_2_loss: 0.1089 - output_3_loss: 0.1152 - output_4_loss: 0.1275 - output_0_acc: 0.9772 - output_1_acc: 0.9730 - output_2_acc: 0.9820 - output_3_acc: 0.9805 - output_4_acc: 0.9735 - val_loss: 1.2095 - val_output_0_loss: 0.2097 - val_output_1_loss: 0.2797 - val_output_2_loss: 0.2411 - val_output_3_loss: 0.2159 - val_output_4_loss: 0.2631 - val_output_0_acc: 0.9445 - val_output_1_acc: 0.9085 - val_output_2_acc: 0.9272 - val_output_3_acc: 0.9422 - val_output_4_acc: 0.9067
Epoch 146/200
4000/4000 [==============================] - 1s 173us/step - loss: 0.5830 - output_0_loss: 0.1112 - output_1_loss: 0.1286 - output_2_loss: 0.1124 - output_3_loss: 0.1112 - output_4_loss: 0.1196 - output_0_acc: 0.9753 - output_1_acc: 0.9715 - output_2_acc: 0.9790 - output_3_acc: 0.9802 - output_4_acc: 0.9748 - val_loss: 0.9870 - val_output_0_loss: 0.1827 - val_output_1_loss: 0.2172 - val_output_2_loss: 0.1887 - val_output_3_loss: 0.2002 - val_output_4_loss: 0.1982 - val_output_0_acc: 0.9533 - val_output_1_acc: 0.9413 - val_output_2_acc: 0.9535 - val_output_3_acc: 0.9520 - val_output_4_acc: 0.9463
Epoch 147/200
4000/4000 [==============================] - 1s 160us/step - loss: 0.5763 - output_0_loss: 0.1114 - output_1_loss: 0.1305 - output_2_loss: 0.1048 - output_3_loss: 0.1075 - output_4_loss: 0.1220 - output_0_acc: 0.9778 - output_1_acc: 0.9695 - output_2_acc: 0.9830 - output_3_acc: 0.9848 - output_4_acc: 0.9725 - val_loss: 1.0419 - val_output_0_loss: 0.1821 - val_output_1_loss: 0.2338 - val_output_2_loss: 0.1913 - val_output_3_loss: 0.2035 - val_output_4_loss: 0.2311 - val_output_0_acc: 0.9615 - val_output_1_acc: 0.9265 - val_output_2_acc: 0.9520 - val_output_3_acc: 0.9437 - val_output_4_acc: 0.9298
Epoch 148/200
4000/4000 [==============================] - 1s 163us/step - loss: 0.5679 - output_0_loss: 0.1052 - output_1_loss: 0.1301 - output_2_loss: 0.1079 - output_3_loss: 0.1074 - output_4_loss: 0.1174 - output_0_acc: 0.9813 - output_1_acc: 0.9735 - output_2_acc: 0.9818 - output_3_acc: 0.9828 - output_4_acc: 0.9740 - val_loss: 1.2033 - val_output_0_loss: 0.2106 - val_output_1_loss: 0.3039 - val_output_2_loss: 0.2128 - val_output_3_loss: 0.2581 - val_output_4_loss: 0.2178 - val_output_0_acc: 0.9357 - val_output_1_acc: 0.8922 - val_output_2_acc: 0.9383 - val_output_3_acc: 0.9098 - val_output_4_acc: 0.9342
Epoch 149/200
4000/4000 [==============================] - 1s 179us/step - loss: 0.5717 - output_0_loss: 0.1108 - output_1_loss: 0.1347 - output_2_loss: 0.1023 - output_3_loss: 0.1029 - output_4_loss: 0.1211 - output_0_acc: 0.9785 - output_1_acc: 0.9663 - output_2_acc: 0.9825 - output_3_acc: 0.9838 - output_4_acc: 0.9710 - val_loss: 0.9982 - val_output_0_loss: 0.1788 - val_output_1_loss: 0.2432 - val_output_2_loss: 0.1802 - val_output_3_loss: 0.1512 - val_output_4_loss: 0.2448 - val_output_0_acc: 0.9597 - val_output_1_acc: 0.9242 - val_output_2_acc: 0.9577 - val_output_3_acc: 0.9718 - val_output_4_acc: 0.9178
Epoch 150/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.5469 - output_0_loss: 0.1019 - output_1_loss: 0.1228 - output_2_loss: 0.1040 - output_3_loss: 0.1017 - output_4_loss: 0.1166 - output_0_acc: 0.9843 - output_1_acc: 0.9740 - output_2_acc: 0.9838 - output_3_acc: 0.9820 - output_4_acc: 0.9732 - val_loss: 0.9722 - val_output_0_loss: 0.1939 - val_output_1_loss: 0.2173 - val_output_2_loss: 0.1843 - val_output_3_loss: 0.1931 - val_output_4_loss: 0.1836 - val_output_0_acc: 0.9525 - val_output_1_acc: 0.9313 - val_output_2_acc: 0.9505 - val_output_3_acc: 0.9432 - val_output_4_acc: 0.9592
Epoch 151/200
4000/4000 [==============================] - 1s 164us/step - loss: 0.5430 - output_0_loss: 0.1035 - output_1_loss: 0.1243 - output_2_loss: 0.1005 - output_3_loss: 0.1000 - output_4_loss: 0.1147 - output_0_acc: 0.9813 - output_1_acc: 0.9710 - output_2_acc: 0.9828 - output_3_acc: 0.9835 - output_4_acc: 0.9778 - val_loss: 1.0024 - val_output_0_loss: 0.2044 - val_output_1_loss: 0.2170 - val_output_2_loss: 0.2007 - val_output_3_loss: 0.1869 - val_output_4_loss: 0.1933 - val_output_0_acc: 0.9432 - val_output_1_acc: 0.9345 - val_output_2_acc: 0.9442 - val_output_3_acc: 0.9520 - val_output_4_acc: 0.9470
Epoch 152/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.5306 - output_0_loss: 0.0984 - output_1_loss: 0.1160 - output_2_loss: 0.1048 - output_3_loss: 0.0992 - output_4_loss: 0.1123 - output_0_acc: 0.9805 - output_1_acc: 0.9738 - output_2_acc: 0.9808 - output_3_acc: 0.9845 - output_4_acc: 0.9760 - val_loss: 1.1554 - val_output_0_loss: 0.2617 - val_output_1_loss: 0.2573 - val_output_2_loss: 0.1659 - val_output_3_loss: 0.2363 - val_output_4_loss: 0.2342 - val_output_0_acc: 0.9052 - val_output_1_acc: 0.9148 - val_output_2_acc: 0.9627 - val_output_3_acc: 0.9250 - val_output_4_acc: 0.9182
Epoch 153/200
4000/4000 [==============================] - 1s 170us/step - loss: 0.5203 - output_0_loss: 0.0994 - output_1_loss: 0.1204 - output_2_loss: 0.0950 - output_3_loss: 0.0985 - output_4_loss: 0.1068 - output_0_acc: 0.9822 - output_1_acc: 0.9715 - output_2_acc: 0.9830 - output_3_acc: 0.9818 - output_4_acc: 0.9778 - val_loss: 1.0831 - val_output_0_loss: 0.1757 - val_output_1_loss: 0.2278 - val_output_2_loss: 0.2406 - val_output_3_loss: 0.1652 - val_output_4_loss: 0.2739 - val_output_0_acc: 0.9527 - val_output_1_acc: 0.9310 - val_output_2_acc: 0.9162 - val_output_3_acc: 0.9627 - val_output_4_acc: 0.9117
Epoch 154/200
4000/4000 [==============================] - 1s 166us/step - loss: 0.5032 - output_0_loss: 0.0933 - output_1_loss: 0.1154 - output_2_loss: 0.0959 - output_3_loss: 0.0931 - output_4_loss: 0.1056 - output_0_acc: 0.9848 - output_1_acc: 0.9775 - output_2_acc: 0.9818 - output_3_acc: 0.9870 - output_4_acc: 0.9775 - val_loss: 1.0921 - val_output_0_loss: 0.2644 - val_output_1_loss: 0.2778 - val_output_2_loss: 0.1772 - val_output_3_loss: 0.1634 - val_output_4_loss: 0.2094 - val_output_0_acc: 0.9092 - val_output_1_acc: 0.9112 - val_output_2_acc: 0.9575 - val_output_3_acc: 0.9625 - val_output_4_acc: 0.9372
Epoch 155/200
4000/4000 [==============================] - 1s 169us/step - loss: 0.5035 - output_0_loss: 0.0961 - output_1_loss: 0.1110 - output_2_loss: 0.0940 - output_3_loss: 0.0959 - output_4_loss: 0.1065 - output_0_acc: 0.9815 - output_1_acc: 0.9750 - output_2_acc: 0.9822 - output_3_acc: 0.9828 - output_4_acc: 0.9768 - val_loss: 0.9744 - val_output_0_loss: 0.1728 - val_output_1_loss: 0.2255 - val_output_2_loss: 0.2044 - val_output_3_loss: 0.1555 - val_output_4_loss: 0.2161 - val_output_0_acc: 0.9592 - val_output_1_acc: 0.9280 - val_output_2_acc: 0.9367 - val_output_3_acc: 0.9692 - val_output_4_acc: 0.9318
Epoch 156/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.4942 - output_0_loss: 0.0918 - output_1_loss: 0.1157 - output_2_loss: 0.0928 - output_3_loss: 0.0898 - output_4_loss: 0.1041 - output_0_acc: 0.9858 - output_1_acc: 0.9723 - output_2_acc: 0.9845 - output_3_acc: 0.9862 - output_4_acc: 0.9772 - val_loss: 1.0266 - val_output_0_loss: 0.1794 - val_output_1_loss: 0.2386 - val_output_2_loss: 0.2123 - val_output_3_loss: 0.1978 - val_output_4_loss: 0.1986 - val_output_0_acc: 0.9480 - val_output_1_acc: 0.9225 - val_output_2_acc: 0.9302 - val_output_3_acc: 0.9402 - val_output_4_acc: 0.9367
Epoch 157/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.4790 - output_0_loss: 0.0874 - output_1_loss: 0.1074 - output_2_loss: 0.0920 - output_3_loss: 0.0899 - output_4_loss: 0.1022 - output_0_acc: 0.9873 - output_1_acc: 0.9765 - output_2_acc: 0.9865 - output_3_acc: 0.9850 - output_4_acc: 0.9778 - val_loss: 1.1430 - val_output_0_loss: 0.2079 - val_output_1_loss: 0.2416 - val_output_2_loss: 0.2620 - val_output_3_loss: 0.1592 - val_output_4_loss: 0.2723 - val_output_0_acc: 0.9345 - val_output_1_acc: 0.9273 - val_output_2_acc: 0.9067 - val_output_3_acc: 0.9627 - val_output_4_acc: 0.9062
Epoch 158/200
4000/4000 [==============================] - 1s 194us/step - loss: 0.4776 - output_0_loss: 0.0877 - output_1_loss: 0.1097 - output_2_loss: 0.0887 - output_3_loss: 0.0903 - output_4_loss: 0.1011 - output_0_acc: 0.9835 - output_1_acc: 0.9745 - output_2_acc: 0.9848 - output_3_acc: 0.9860 - output_4_acc: 0.9765 - val_loss: 0.9951 - val_output_0_loss: 0.1755 - val_output_1_loss: 0.2090 - val_output_2_loss: 0.2225 - val_output_3_loss: 0.1770 - val_output_4_loss: 0.2111 - val_output_0_acc: 0.9547 - val_output_1_acc: 0.9425 - val_output_2_acc: 0.9258 - val_output_3_acc: 0.9517 - val_output_4_acc: 0.9312
Epoch 159/200
4000/4000 [==============================] - 1s 195us/step - loss: 0.4639 - output_0_loss: 0.0872 - output_1_loss: 0.1015 - output_2_loss: 0.0870 - output_3_loss: 0.0861 - output_4_loss: 0.1021 - output_0_acc: 0.9865 - output_1_acc: 0.9795 - output_2_acc: 0.9848 - output_3_acc: 0.9850 - output_4_acc: 0.9783 - val_loss: 1.0107 - val_output_0_loss: 0.1996 - val_output_1_loss: 0.2902 - val_output_2_loss: 0.1697 - val_output_3_loss: 0.1769 - val_output_4_loss: 0.1744 - val_output_0_acc: 0.9422 - val_output_1_acc: 0.8995 - val_output_2_acc: 0.9607 - val_output_3_acc: 0.9480 - val_output_4_acc: 0.9547
Epoch 160/200
4000/4000 [==============================] - 1s 182us/step - loss: 0.4600 - output_0_loss: 0.0868 - output_1_loss: 0.1087 - output_2_loss: 0.0832 - output_3_loss: 0.0822 - output_4_loss: 0.0991 - output_0_acc: 0.9845 - output_1_acc: 0.9760 - output_2_acc: 0.9850 - output_3_acc: 0.9888 - output_4_acc: 0.9795 - val_loss: 0.9570 - val_output_0_loss: 0.1846 - val_output_1_loss: 0.1911 - val_output_2_loss: 0.2005 - val_output_3_loss: 0.1763 - val_output_4_loss: 0.2046 - val_output_0_acc: 0.9460 - val_output_1_acc: 0.9452 - val_output_2_acc: 0.9485 - val_output_3_acc: 0.9512 - val_output_4_acc: 0.9365
Epoch 161/200
4000/4000 [==============================] - 1s 193us/step - loss: 0.4636 - output_0_loss: 0.0844 - output_1_loss: 0.1019 - output_2_loss: 0.0894 - output_3_loss: 0.0893 - output_4_loss: 0.0986 - output_0_acc: 0.9875 - output_1_acc: 0.9800 - output_2_acc: 0.9843 - output_3_acc: 0.9830 - output_4_acc: 0.9772 - val_loss: 0.9398 - val_output_0_loss: 0.1867 - val_output_1_loss: 0.2238 - val_output_2_loss: 0.1626 - val_output_3_loss: 0.1560 - val_output_4_loss: 0.2106 - val_output_0_acc: 0.9442 - val_output_1_acc: 0.9262 - val_output_2_acc: 0.9547 - val_output_3_acc: 0.9593 - val_output_4_acc: 0.9362
Epoch 162/200
4000/4000 [==============================] - 1s 189us/step - loss: 0.4394 - output_0_loss: 0.0803 - output_1_loss: 0.1015 - output_2_loss: 0.0811 - output_3_loss: 0.0819 - output_4_loss: 0.0946 - output_0_acc: 0.9865 - output_1_acc: 0.9790 - output_2_acc: 0.9850 - output_3_acc: 0.9870 - output_4_acc: 0.9795 - val_loss: 0.9716 - val_output_0_loss: 0.1596 - val_output_1_loss: 0.2617 - val_output_2_loss: 0.1338 - val_output_3_loss: 0.2176 - val_output_4_loss: 0.1989 - val_output_0_acc: 0.9595 - val_output_1_acc: 0.9148 - val_output_2_acc: 0.9743 - val_output_3_acc: 0.9335 - val_output_4_acc: 0.9415
Epoch 163/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.4373 - output_0_loss: 0.0804 - output_1_loss: 0.0995 - output_2_loss: 0.0803 - output_3_loss: 0.0823 - output_4_loss: 0.0947 - output_0_acc: 0.9862 - output_1_acc: 0.9783 - output_2_acc: 0.9865 - output_3_acc: 0.9845 - output_4_acc: 0.9800 - val_loss: 0.9450 - val_output_0_loss: 0.1895 - val_output_1_loss: 0.2694 - val_output_2_loss: 0.1487 - val_output_3_loss: 0.1681 - val_output_4_loss: 0.1694 - val_output_0_acc: 0.9385 - val_output_1_acc: 0.9017 - val_output_2_acc: 0.9627 - val_output_3_acc: 0.9588 - val_output_4_acc: 0.9555
Epoch 164/200
4000/4000 [==============================] - 1s 190us/step - loss: 0.4348 - output_0_loss: 0.0813 - output_1_loss: 0.1008 - output_2_loss: 0.0816 - output_3_loss: 0.0790 - output_4_loss: 0.0919 - output_0_acc: 0.9855 - output_1_acc: 0.9780 - output_2_acc: 0.9870 - output_3_acc: 0.9898 - output_4_acc: 0.9813 - val_loss: 0.9237 - val_output_0_loss: 0.1760 - val_output_1_loss: 0.2141 - val_output_2_loss: 0.1532 - val_output_3_loss: 0.1607 - val_output_4_loss: 0.2196 - val_output_0_acc: 0.9510 - val_output_1_acc: 0.9242 - val_output_2_acc: 0.9682 - val_output_3_acc: 0.9595 - val_output_4_acc: 0.9270
Epoch 165/200
4000/4000 [==============================] - 1s 199us/step - loss: 0.4173 - output_0_loss: 0.0781 - output_1_loss: 0.0950 - output_2_loss: 0.0782 - output_3_loss: 0.0770 - output_4_loss: 0.0890 - output_0_acc: 0.9873 - output_1_acc: 0.9785 - output_2_acc: 0.9882 - output_3_acc: 0.9880 - output_4_acc: 0.9813 - val_loss: 0.8666 - val_output_0_loss: 0.1991 - val_output_1_loss: 0.1994 - val_output_2_loss: 0.1566 - val_output_3_loss: 0.1321 - val_output_4_loss: 0.1795 - val_output_0_acc: 0.9310 - val_output_1_acc: 0.9438 - val_output_2_acc: 0.9645 - val_output_3_acc: 0.9773 - val_output_4_acc: 0.9443
Epoch 166/200
4000/4000 [==============================] - 1s 203us/step - loss: 0.4115 - output_0_loss: 0.0777 - output_1_loss: 0.0975 - output_2_loss: 0.0770 - output_3_loss: 0.0747 - output_4_loss: 0.0845 - output_0_acc: 0.9862 - output_1_acc: 0.9783 - output_2_acc: 0.9882 - output_3_acc: 0.9882 - output_4_acc: 0.9843 - val_loss: 0.8750 - val_output_0_loss: 0.2272 - val_output_1_loss: 0.1769 - val_output_2_loss: 0.1579 - val_output_3_loss: 0.1251 - val_output_4_loss: 0.1880 - val_output_0_acc: 0.9167 - val_output_1_acc: 0.9538 - val_output_2_acc: 0.9612 - val_output_3_acc: 0.9752 - val_output_4_acc: 0.9483
Epoch 167/200
4000/4000 [==============================] - 1s 193us/step - loss: 0.4040 - output_0_loss: 0.0732 - output_1_loss: 0.0922 - output_2_loss: 0.0751 - output_3_loss: 0.0775 - output_4_loss: 0.0860 - output_0_acc: 0.9888 - output_1_acc: 0.9822 - output_2_acc: 0.9888 - output_3_acc: 0.9895 - output_4_acc: 0.9838 - val_loss: 0.9429 - val_output_0_loss: 0.1836 - val_output_1_loss: 0.2401 - val_output_2_loss: 0.1946 - val_output_3_loss: 0.1615 - val_output_4_loss: 0.1631 - val_output_0_acc: 0.9477 - val_output_1_acc: 0.9165 - val_output_2_acc: 0.9413 - val_output_3_acc: 0.9580 - val_output_4_acc: 0.9482
Epoch 168/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.3955 - output_0_loss: 0.0715 - output_1_loss: 0.0928 - output_2_loss: 0.0732 - output_3_loss: 0.0729 - output_4_loss: 0.0850 - output_0_acc: 0.9878 - output_1_acc: 0.9810 - output_2_acc: 0.9875 - output_3_acc: 0.9882 - output_4_acc: 0.9825 - val_loss: 0.8541 - val_output_0_loss: 0.1527 - val_output_1_loss: 0.1931 - val_output_2_loss: 0.1371 - val_output_3_loss: 0.1808 - val_output_4_loss: 0.1904 - val_output_0_acc: 0.9615 - val_output_1_acc: 0.9390 - val_output_2_acc: 0.9668 - val_output_3_acc: 0.9385 - val_output_4_acc: 0.9377
Epoch 169/200
4000/4000 [==============================] - 1s 190us/step - loss: 0.3913 - output_0_loss: 0.0731 - output_1_loss: 0.0898 - output_2_loss: 0.0718 - output_3_loss: 0.0709 - output_4_loss: 0.0858 - output_0_acc: 0.9875 - output_1_acc: 0.9840 - output_2_acc: 0.9905 - output_3_acc: 0.9912 - output_4_acc: 0.9808 - val_loss: 0.8552 - val_output_0_loss: 0.1670 - val_output_1_loss: 0.1917 - val_output_2_loss: 0.1406 - val_output_3_loss: 0.1600 - val_output_4_loss: 0.1958 - val_output_0_acc: 0.9522 - val_output_1_acc: 0.9465 - val_output_2_acc: 0.9645 - val_output_3_acc: 0.9578 - val_output_4_acc: 0.9412
Epoch 170/200
4000/4000 [==============================] - 1s 197us/step - loss: 0.3831 - output_0_loss: 0.0704 - output_1_loss: 0.0909 - output_2_loss: 0.0684 - output_3_loss: 0.0737 - output_4_loss: 0.0797 - output_0_acc: 0.9873 - output_1_acc: 0.9795 - output_2_acc: 0.9900 - output_3_acc: 0.9888 - output_4_acc: 0.9852 - val_loss: 0.8597 - val_output_0_loss: 0.1716 - val_output_1_loss: 0.2196 - val_output_2_loss: 0.1509 - val_output_3_loss: 0.1467 - val_output_4_loss: 0.1708 - val_output_0_acc: 0.9555 - val_output_1_acc: 0.9277 - val_output_2_acc: 0.9548 - val_output_3_acc: 0.9652 - val_output_4_acc: 0.9468
Epoch 171/200
4000/4000 [==============================] - 1s 176us/step - loss: 0.3833 - output_0_loss: 0.0706 - output_1_loss: 0.0871 - output_2_loss: 0.0729 - output_3_loss: 0.0690 - output_4_loss: 0.0837 - output_0_acc: 0.9862 - output_1_acc: 0.9843 - output_2_acc: 0.9865 - output_3_acc: 0.9900 - output_4_acc: 0.9798 - val_loss: 0.8132 - val_output_0_loss: 0.1346 - val_output_1_loss: 0.2269 - val_output_2_loss: 0.1526 - val_output_3_loss: 0.1426 - val_output_4_loss: 0.1564 - val_output_0_acc: 0.9618 - val_output_1_acc: 0.9237 - val_output_2_acc: 0.9543 - val_output_3_acc: 0.9633 - val_output_4_acc: 0.9562
Epoch 172/200
4000/4000 [==============================] - 1s 171us/step - loss: 0.3726 - output_0_loss: 0.0635 - output_1_loss: 0.0864 - output_2_loss: 0.0687 - output_3_loss: 0.0710 - output_4_loss: 0.0829 - output_0_acc: 0.9930 - output_1_acc: 0.9830 - output_2_acc: 0.9905 - output_3_acc: 0.9888 - output_4_acc: 0.9835 - val_loss: 0.8192 - val_output_0_loss: 0.1492 - val_output_1_loss: 0.2224 - val_output_2_loss: 0.1557 - val_output_3_loss: 0.1558 - val_output_4_loss: 0.1361 - val_output_0_acc: 0.9630 - val_output_1_acc: 0.9227 - val_output_2_acc: 0.9533 - val_output_3_acc: 0.9552 - val_output_4_acc: 0.9633
Epoch 173/200
4000/4000 [==============================] - 1s 177us/step - loss: 0.3593 - output_0_loss: 0.0652 - output_1_loss: 0.0856 - output_2_loss: 0.0661 - output_3_loss: 0.0647 - output_4_loss: 0.0777 - output_0_acc: 0.9900 - output_1_acc: 0.9835 - output_2_acc: 0.9895 - output_3_acc: 0.9922 - output_4_acc: 0.9870 - val_loss: 0.8209 - val_output_0_loss: 0.1988 - val_output_1_loss: 0.1857 - val_output_2_loss: 0.1428 - val_output_3_loss: 0.1367 - val_output_4_loss: 0.1569 - val_output_0_acc: 0.9298 - val_output_1_acc: 0.9487 - val_output_2_acc: 0.9688 - val_output_3_acc: 0.9683 - val_output_4_acc: 0.9532
Epoch 174/200
4000/4000 [==============================] - 1s 180us/step - loss: 0.3621 - output_0_loss: 0.0673 - output_1_loss: 0.0841 - output_2_loss: 0.0659 - output_3_loss: 0.0680 - output_4_loss: 0.0768 - output_0_acc: 0.9908 - output_1_acc: 0.9830 - output_2_acc: 0.9885 - output_3_acc: 0.9898 - output_4_acc: 0.9840 - val_loss: 0.8155 - val_output_0_loss: 0.1219 - val_output_1_loss: 0.2454 - val_output_2_loss: 0.1182 - val_output_3_loss: 0.1526 - val_output_4_loss: 0.1773 - val_output_0_acc: 0.9732 - val_output_1_acc: 0.9038 - val_output_2_acc: 0.9772 - val_output_3_acc: 0.9482 - val_output_4_acc: 0.9463
Epoch 175/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.3523 - output_0_loss: 0.0646 - output_1_loss: 0.0838 - output_2_loss: 0.0625 - output_3_loss: 0.0657 - output_4_loss: 0.0756 - output_0_acc: 0.9903 - output_1_acc: 0.9810 - output_2_acc: 0.9908 - output_3_acc: 0.9905 - output_4_acc: 0.9843 - val_loss: 0.7948 - val_output_0_loss: 0.2032 - val_output_1_loss: 0.1800 - val_output_2_loss: 0.1297 - val_output_3_loss: 0.1255 - val_output_4_loss: 0.1564 - val_output_0_acc: 0.9350 - val_output_1_acc: 0.9495 - val_output_2_acc: 0.9697 - val_output_3_acc: 0.9763 - val_output_4_acc: 0.9505
Epoch 176/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.3411 - output_0_loss: 0.0627 - output_1_loss: 0.0795 - output_2_loss: 0.0623 - output_3_loss: 0.0638 - output_4_loss: 0.0729 - output_0_acc: 0.9915 - output_1_acc: 0.9838 - output_2_acc: 0.9915 - output_3_acc: 0.9903 - output_4_acc: 0.9880 - val_loss: 0.7277 - val_output_0_loss: 0.1149 - val_output_1_loss: 0.1650 - val_output_2_loss: 0.1440 - val_output_3_loss: 0.1362 - val_output_4_loss: 0.1675 - val_output_0_acc: 0.9758 - val_output_1_acc: 0.9500 - val_output_2_acc: 0.9618 - val_output_3_acc: 0.9623 - val_output_4_acc: 0.9440
Epoch 177/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.3360 - output_0_loss: 0.0598 - output_1_loss: 0.0824 - output_2_loss: 0.0620 - output_3_loss: 0.0619 - output_4_loss: 0.0698 - output_0_acc: 0.9908 - output_1_acc: 0.9800 - output_2_acc: 0.9928 - output_3_acc: 0.9895 - output_4_acc: 0.9860 - val_loss: 0.8782 - val_output_0_loss: 0.1589 - val_output_1_loss: 0.1876 - val_output_2_loss: 0.1807 - val_output_3_loss: 0.1418 - val_output_4_loss: 0.2092 - val_output_0_acc: 0.9550 - val_output_1_acc: 0.9422 - val_output_2_acc: 0.9367 - val_output_3_acc: 0.9663 - val_output_4_acc: 0.9270
Epoch 178/200
4000/4000 [==============================] - 1s 195us/step - loss: 0.3297 - output_0_loss: 0.0615 - output_1_loss: 0.0755 - output_2_loss: 0.0603 - output_3_loss: 0.0622 - output_4_loss: 0.0702 - output_0_acc: 0.9910 - output_1_acc: 0.9860 - output_2_acc: 0.9908 - output_3_acc: 0.9930 - output_4_acc: 0.9873 - val_loss: 0.7487 - val_output_0_loss: 0.1303 - val_output_1_loss: 0.1977 - val_output_2_loss: 0.1145 - val_output_3_loss: 0.1436 - val_output_4_loss: 0.1626 - val_output_0_acc: 0.9643 - val_output_1_acc: 0.9362 - val_output_2_acc: 0.9763 - val_output_3_acc: 0.9583 - val_output_4_acc: 0.9583
Epoch 179/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.3254 - output_0_loss: 0.0607 - output_1_loss: 0.0770 - output_2_loss: 0.0599 - output_3_loss: 0.0602 - output_4_loss: 0.0676 - output_0_acc: 0.9903 - output_1_acc: 0.9870 - output_2_acc: 0.9903 - output_3_acc: 0.9930 - output_4_acc: 0.9880 - val_loss: 0.8331 - val_output_0_loss: 0.1272 - val_output_1_loss: 0.1809 - val_output_2_loss: 0.1386 - val_output_3_loss: 0.1651 - val_output_4_loss: 0.2213 - val_output_0_acc: 0.9687 - val_output_1_acc: 0.9460 - val_output_2_acc: 0.9663 - val_output_3_acc: 0.9480 - val_output_4_acc: 0.9142
Epoch 180/200
4000/4000 [==============================] - 1s 185us/step - loss: 0.3236 - output_0_loss: 0.0589 - output_1_loss: 0.0721 - output_2_loss: 0.0643 - output_3_loss: 0.0592 - output_4_loss: 0.0690 - output_0_acc: 0.9888 - output_1_acc: 0.9868 - output_2_acc: 0.9880 - output_3_acc: 0.9920 - output_4_acc: 0.9873 - val_loss: 0.7394 - val_output_0_loss: 0.1425 - val_output_1_loss: 0.2251 - val_output_2_loss: 0.1202 - val_output_3_loss: 0.1203 - val_output_4_loss: 0.1314 - val_output_0_acc: 0.9563 - val_output_1_acc: 0.9252 - val_output_2_acc: 0.9700 - val_output_3_acc: 0.9665 - val_output_4_acc: 0.9643
Epoch 181/200
4000/4000 [==============================] - 1s 197us/step - loss: 0.3112 - output_0_loss: 0.0578 - output_1_loss: 0.0750 - output_2_loss: 0.0551 - output_3_loss: 0.0584 - output_4_loss: 0.0649 - output_0_acc: 0.9900 - output_1_acc: 0.9820 - output_2_acc: 0.9928 - output_3_acc: 0.9933 - output_4_acc: 0.9885 - val_loss: 0.7499 - val_output_0_loss: 0.1260 - val_output_1_loss: 0.1771 - val_output_2_loss: 0.1483 - val_output_3_loss: 0.1543 - val_output_4_loss: 0.1442 - val_output_0_acc: 0.9692 - val_output_1_acc: 0.9465 - val_output_2_acc: 0.9607 - val_output_3_acc: 0.9430 - val_output_4_acc: 0.9653
Epoch 182/200
4000/4000 [==============================] - 1s 193us/step - loss: 0.3033 - output_0_loss: 0.0540 - output_1_loss: 0.0708 - output_2_loss: 0.0571 - output_3_loss: 0.0558 - output_4_loss: 0.0656 - output_0_acc: 0.9920 - output_1_acc: 0.9890 - output_2_acc: 0.9888 - output_3_acc: 0.9928 - output_4_acc: 0.9870 - val_loss: 0.6842 - val_output_0_loss: 0.1481 - val_output_1_loss: 0.1837 - val_output_2_loss: 0.1071 - val_output_3_loss: 0.1154 - val_output_4_loss: 0.1300 - val_output_0_acc: 0.9558 - val_output_1_acc: 0.9428 - val_output_2_acc: 0.9760 - val_output_3_acc: 0.9738 - val_output_4_acc: 0.9647
Epoch 183/200
4000/4000 [==============================] - 1s 188us/step - loss: 0.3028 - output_0_loss: 0.0556 - output_1_loss: 0.0713 - output_2_loss: 0.0543 - output_3_loss: 0.0565 - output_4_loss: 0.0651 - output_0_acc: 0.9925 - output_1_acc: 0.9845 - output_2_acc: 0.9942 - output_3_acc: 0.9922 - output_4_acc: 0.9868 - val_loss: 0.7205 - val_output_0_loss: 0.1465 - val_output_1_loss: 0.2323 - val_output_2_loss: 0.1048 - val_output_3_loss: 0.1148 - val_output_4_loss: 0.1220 - val_output_0_acc: 0.9550 - val_output_1_acc: 0.9142 - val_output_2_acc: 0.9778 - val_output_3_acc: 0.9732 - val_output_4_acc: 0.9708
Epoch 184/200
4000/4000 [==============================] - 1s 189us/step - loss: 0.2986 - output_0_loss: 0.0550 - output_1_loss: 0.0687 - output_2_loss: 0.0574 - output_3_loss: 0.0542 - output_4_loss: 0.0633 - output_0_acc: 0.9928 - output_1_acc: 0.9875 - output_2_acc: 0.9900 - output_3_acc: 0.9925 - output_4_acc: 0.9888 - val_loss: 0.8299 - val_output_0_loss: 0.1670 - val_output_1_loss: 0.1669 - val_output_2_loss: 0.2111 - val_output_3_loss: 0.1290 - val_output_4_loss: 0.1559 - val_output_0_acc: 0.9412 - val_output_1_acc: 0.9433 - val_output_2_acc: 0.9190 - val_output_3_acc: 0.9662 - val_output_4_acc: 0.9472
Epoch 185/200
4000/4000 [==============================] - 1s 187us/step - loss: 0.2945 - output_0_loss: 0.0521 - output_1_loss: 0.0715 - output_2_loss: 0.0541 - output_3_loss: 0.0529 - output_4_loss: 0.0637 - output_0_acc: 0.9925 - output_1_acc: 0.9855 - output_2_acc: 0.9912 - output_3_acc: 0.9920 - output_4_acc: 0.9875 - val_loss: 0.6525 - val_output_0_loss: 0.1203 - val_output_1_loss: 0.1340 - val_output_2_loss: 0.1009 - val_output_3_loss: 0.1586 - val_output_4_loss: 0.1387 - val_output_0_acc: 0.9683 - val_output_1_acc: 0.9655 - val_output_2_acc: 0.9800 - val_output_3_acc: 0.9452 - val_output_4_acc: 0.9603
Epoch 186/200
4000/4000 [==============================] - 1s 189us/step - loss: 0.2857 - output_0_loss: 0.0529 - output_1_loss: 0.0682 - output_2_loss: 0.0514 - output_3_loss: 0.0518 - output_4_loss: 0.0614 - output_0_acc: 0.9925 - output_1_acc: 0.9862 - output_2_acc: 0.9928 - output_3_acc: 0.9928 - output_4_acc: 0.9890 - val_loss: 0.7752 - val_output_0_loss: 0.1586 - val_output_1_loss: 0.2121 - val_output_2_loss: 0.1061 - val_output_3_loss: 0.1070 - val_output_4_loss: 0.1914 - val_output_0_acc: 0.9440 - val_output_1_acc: 0.9317 - val_output_2_acc: 0.9752 - val_output_3_acc: 0.9812 - val_output_4_acc: 0.9370
Epoch 187/200
4000/4000 [==============================] - 1s 198us/step - loss: 0.2789 - output_0_loss: 0.0518 - output_1_loss: 0.0663 - output_2_loss: 0.0501 - output_3_loss: 0.0505 - output_4_loss: 0.0602 - output_0_acc: 0.9915 - output_1_acc: 0.9868 - output_2_acc: 0.9925 - output_3_acc: 0.9945 - output_4_acc: 0.9888 - val_loss: 0.6074 - val_output_0_loss: 0.1296 - val_output_1_loss: 0.1460 - val_output_2_loss: 0.0999 - val_output_3_loss: 0.0960 - val_output_4_loss: 0.1358 - val_output_0_acc: 0.9660 - val_output_1_acc: 0.9590 - val_output_2_acc: 0.9790 - val_output_3_acc: 0.9837 - val_output_4_acc: 0.9620
Epoch 188/200
4000/4000 [==============================] - 1s 201us/step - loss: 0.2763 - output_0_loss: 0.0501 - output_1_loss: 0.0641 - output_2_loss: 0.0526 - output_3_loss: 0.0524 - output_4_loss: 0.0571 - output_0_acc: 0.9952 - output_1_acc: 0.9870 - output_2_acc: 0.9918 - output_3_acc: 0.9940 - output_4_acc: 0.9905 - val_loss: 0.6371 - val_output_0_loss: 0.1014 - val_output_1_loss: 0.1502 - val_output_2_loss: 0.1623 - val_output_3_loss: 0.1066 - val_output_4_loss: 0.1168 - val_output_0_acc: 0.9775 - val_output_1_acc: 0.9560 - val_output_2_acc: 0.9430 - val_output_3_acc: 0.9767 - val_output_4_acc: 0.9680
Epoch 189/200
4000/4000 [==============================] - 1s 201us/step - loss: 0.2717 - output_0_loss: 0.0494 - output_1_loss: 0.0633 - output_2_loss: 0.0494 - output_3_loss: 0.0506 - output_4_loss: 0.0591 - output_0_acc: 0.9922 - output_1_acc: 0.9885 - output_2_acc: 0.9922 - output_3_acc: 0.9933 - output_4_acc: 0.9890 - val_loss: 0.6604 - val_output_0_loss: 0.1281 - val_output_1_loss: 0.1261 - val_output_2_loss: 0.1311 - val_output_3_loss: 0.0881 - val_output_4_loss: 0.1870 - val_output_0_acc: 0.9660 - val_output_1_acc: 0.9697 - val_output_2_acc: 0.9665 - val_output_3_acc: 0.9833 - val_output_4_acc: 0.9375
Epoch 190/200
4000/4000 [==============================] - 1s 198us/step - loss: 0.2680 - output_0_loss: 0.0456 - output_1_loss: 0.0654 - output_2_loss: 0.0471 - output_3_loss: 0.0507 - output_4_loss: 0.0591 - output_0_acc: 0.9942 - output_1_acc: 0.9878 - output_2_acc: 0.9948 - output_3_acc: 0.9940 - output_4_acc: 0.9888 - val_loss: 0.7075 - val_output_0_loss: 0.1372 - val_output_1_loss: 0.1613 - val_output_2_loss: 0.1190 - val_output_3_loss: 0.1347 - val_output_4_loss: 0.1553 - val_output_0_acc: 0.9640 - val_output_1_acc: 0.9522 - val_output_2_acc: 0.9727 - val_output_3_acc: 0.9627 - val_output_4_acc: 0.9465
Epoch 191/200
4000/4000 [==============================] - 1s 179us/step - loss: 0.2562 - output_0_loss: 0.0467 - output_1_loss: 0.0594 - output_2_loss: 0.0478 - output_3_loss: 0.0482 - output_4_loss: 0.0541 - output_0_acc: 0.9942 - output_1_acc: 0.9900 - output_2_acc: 0.9925 - output_3_acc: 0.9940 - output_4_acc: 0.9908 - val_loss: 0.7061 - val_output_0_loss: 0.1100 - val_output_1_loss: 0.1768 - val_output_2_loss: 0.1552 - val_output_3_loss: 0.1131 - val_output_4_loss: 0.1510 - val_output_0_acc: 0.9705 - val_output_1_acc: 0.9372 - val_output_2_acc: 0.9530 - val_output_3_acc: 0.9737 - val_output_4_acc: 0.9558
Epoch 192/200
4000/4000 [==============================] - 1s 180us/step - loss: 0.2572 - output_0_loss: 0.0450 - output_1_loss: 0.0601 - output_2_loss: 0.0476 - output_3_loss: 0.0466 - output_4_loss: 0.0578 - output_0_acc: 0.9935 - output_1_acc: 0.9885 - output_2_acc: 0.9933 - output_3_acc: 0.9942 - output_4_acc: 0.9892 - val_loss: 0.7481 - val_output_0_loss: 0.1417 - val_output_1_loss: 0.1514 - val_output_2_loss: 0.1485 - val_output_3_loss: 0.1802 - val_output_4_loss: 0.1262 - val_output_0_acc: 0.9517 - val_output_1_acc: 0.9553 - val_output_2_acc: 0.9517 - val_output_3_acc: 0.9322 - val_output_4_acc: 0.9632
Epoch 193/200
4000/4000 [==============================] - 1s 172us/step - loss: 0.2498 - output_0_loss: 0.0449 - output_1_loss: 0.0581 - output_2_loss: 0.0453 - output_3_loss: 0.0472 - output_4_loss: 0.0543 - output_0_acc: 0.9948 - output_1_acc: 0.9895 - output_2_acc: 0.9948 - output_3_acc: 0.9928 - output_4_acc: 0.9903 - val_loss: 0.6439 - val_output_0_loss: 0.1131 - val_output_1_loss: 0.1685 - val_output_2_loss: 0.0963 - val_output_3_loss: 0.1199 - val_output_4_loss: 0.1462 - val_output_0_acc: 0.9702 - val_output_1_acc: 0.9515 - val_output_2_acc: 0.9822 - val_output_3_acc: 0.9645 - val_output_4_acc: 0.9560
Epoch 194/200
4000/4000 [==============================] - 1s 163us/step - loss: 0.2456 - output_0_loss: 0.0463 - output_1_loss: 0.0587 - output_2_loss: 0.0437 - output_3_loss: 0.0447 - output_4_loss: 0.0523 - output_0_acc: 0.9910 - output_1_acc: 0.9903 - output_2_acc: 0.9945 - output_3_acc: 0.9960 - output_4_acc: 0.9910 - val_loss: 0.6336 - val_output_0_loss: 0.1297 - val_output_1_loss: 0.1592 - val_output_2_loss: 0.1346 - val_output_3_loss: 0.1052 - val_output_4_loss: 0.1049 - val_output_0_acc: 0.9633 - val_output_1_acc: 0.9525 - val_output_2_acc: 0.9662 - val_output_3_acc: 0.9758 - val_output_4_acc: 0.9737
Epoch 195/200
4000/4000 [==============================] - 1s 165us/step - loss: 0.2374 - output_0_loss: 0.0415 - output_1_loss: 0.0562 - output_2_loss: 0.0418 - output_3_loss: 0.0448 - output_4_loss: 0.0531 - output_0_acc: 0.9945 - output_1_acc: 0.9915 - output_2_acc: 0.9950 - output_3_acc: 0.9963 - output_4_acc: 0.9905 - val_loss: 0.6546 - val_output_0_loss: 0.1222 - val_output_1_loss: 0.1862 - val_output_2_loss: 0.1263 - val_output_3_loss: 0.0925 - val_output_4_loss: 0.1274 - val_output_0_acc: 0.9670 - val_output_1_acc: 0.9288 - val_output_2_acc: 0.9637 - val_output_3_acc: 0.9827 - val_output_4_acc: 0.9613
Epoch 196/200
4000/4000 [==============================] - 1s 170us/step - loss: 0.2370 - output_0_loss: 0.0415 - output_1_loss: 0.0590 - output_2_loss: 0.0426 - output_3_loss: 0.0455 - output_4_loss: 0.0485 - output_0_acc: 0.9933 - output_1_acc: 0.9870 - output_2_acc: 0.9945 - output_3_acc: 0.9942 - output_4_acc: 0.9925 - val_loss: 0.6560 - val_output_0_loss: 0.1080 - val_output_1_loss: 0.1326 - val_output_2_loss: 0.1175 - val_output_3_loss: 0.1313 - val_output_4_loss: 0.1665 - val_output_0_acc: 0.9735 - val_output_1_acc: 0.9612 - val_output_2_acc: 0.9715 - val_output_3_acc: 0.9627 - val_output_4_acc: 0.9392
Epoch 197/200
4000/4000 [==============================] - 1s 160us/step - loss: 0.2369 - output_0_loss: 0.0423 - output_1_loss: 0.0552 - output_2_loss: 0.0412 - output_3_loss: 0.0444 - output_4_loss: 0.0538 - output_0_acc: 0.9940 - output_1_acc: 0.9903 - output_2_acc: 0.9963 - output_3_acc: 0.9955 - output_4_acc: 0.9898 - val_loss: 0.5583 - val_output_0_loss: 0.0913 - val_output_1_loss: 0.1636 - val_output_2_loss: 0.0994 - val_output_3_loss: 0.1009 - val_output_4_loss: 0.1031 - val_output_0_acc: 0.9800 - val_output_1_acc: 0.9502 - val_output_2_acc: 0.9732 - val_output_3_acc: 0.9792 - val_output_4_acc: 0.9762
Epoch 198/200
4000/4000 [==============================] - 1s 182us/step - loss: 0.2242 - output_0_loss: 0.0396 - output_1_loss: 0.0565 - output_2_loss: 0.0398 - output_3_loss: 0.0429 - output_4_loss: 0.0453 - output_0_acc: 0.9952 - output_1_acc: 0.9882 - output_2_acc: 0.9965 - output_3_acc: 0.9948 - output_4_acc: 0.9930 - val_loss: 0.7061 - val_output_0_loss: 0.1454 - val_output_1_loss: 0.1301 - val_output_2_loss: 0.1379 - val_output_3_loss: 0.1383 - val_output_4_loss: 0.1544 - val_output_0_acc: 0.9492 - val_output_1_acc: 0.9600 - val_output_2_acc: 0.9553 - val_output_3_acc: 0.9593 - val_output_4_acc: 0.9460
Epoch 199/200
4000/4000 [==============================] - 1s 178us/step - loss: 0.2291 - output_0_loss: 0.0439 - output_1_loss: 0.0542 - output_2_loss: 0.0396 - output_3_loss: 0.0420 - output_4_loss: 0.0494 - output_0_acc: 0.9925 - output_1_acc: 0.9885 - output_2_acc: 0.9958 - output_3_acc: 0.9952 - output_4_acc: 0.9915 - val_loss: 0.6422 - val_output_0_loss: 0.1283 - val_output_1_loss: 0.1808 - val_output_2_loss: 0.1068 - val_output_3_loss: 0.0923 - val_output_4_loss: 0.1340 - val_output_0_acc: 0.9610 - val_output_1_acc: 0.9398 - val_output_2_acc: 0.9795 - val_output_3_acc: 0.9772 - val_output_4_acc: 0.9605
Epoch 200/200
4000/4000 [==============================] - 1s 179us/step - loss: 0.2206 - output_0_loss: 0.0399 - output_1_loss: 0.0515 - output_2_loss: 0.0405 - output_3_loss: 0.0398 - output_4_loss: 0.0488 - output_0_acc: 0.9940 - output_1_acc: 0.9920 - output_2_acc: 0.9940 - output_3_acc: 0.9952 - output_4_acc: 0.9905 - val_loss: 0.5999 - val_output_0_loss: 0.1317 - val_output_1_loss: 0.1252 - val_output_2_loss: 0.0995 - val_output_3_loss: 0.0779 - val_output_4_loss: 0.1657 - val_output_0_acc: 0.9613 - val_output_1_acc: 0.9623 - val_output_2_acc: 0.9785 - val_output_3_acc: 0.9857 - val_output_4_acc: 0.9405
In [12]:
# Plot training & validation accuracy values (of first char only)
plt.plot(history.history['output_0_acc'])
plt.plot(history.history['val_output_0_acc'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
# Plot training & validation loss values (of first char only)
plt.plot(history.history['output_0_loss'])
plt.plot(history.history['val_output_0_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
In [13]:
nb_words_to_test = 100000
x_test = create_inputs(nb_words_to_test, nb_chars, nb_letters)
x_test_scaled = scaler.transform(x_test)
y_test_raw = encrypt(x_test, nb_words_to_test, nb_chars)
y_test_raw_cate = keras.utils.to_categorical(y_test_raw, nb_letters)
# process the y data as useful ANN multiple-outputs data
y_test = []
for c in range(nb_chars):
# extract each 'char' colomn from the global y_train0 tensor
# in order to have multiplue yi_train outputs tensors
yi_test = y_test_raw_cate[:,c,:]
y_test.append(yi_test)
print('\n# Evaluate on test data')
results = coding_model.evaluate(x_test_scaled, y_test, batch_size=128)
for r in range(len(results)):
print(coding_model.metrics_names[r],':',results[r])
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:595: DataConversionWarning: Data with input dtype int64 was converted to float64 by StandardScaler.
warnings.warn(msg, DataConversionWarning)
# Evaluate on test data
100000/100000 [==============================] - 2s 20us/step
loss : 0.608820455379486
output_0_loss : 0.12797732916116714
output_1_loss : 0.12703988629817964
output_2_loss : 0.1091656908273697
output_3_loss : 0.08084671403884888
output_4_loss : 0.1637908337879181
output_0_acc : 0.96503
output_1_acc : 0.96431
output_2_acc : 0.97602
output_3_acc : 0.98366
output_4_acc : 0.94185
In [14]:
nb_words_to_test = 3
x_test = create_inputs(nb_words_to_test, nb_chars, nb_letters)
print_readable_inputs(x_test)
print("x_test=\n", x_test)
x_test_scaled = scaler.transform(x_test)
print("x_test_scaled=\n", x_test_scaled)
print('-->')
prediction = coding_model.predict(x_test_scaled)
#print(prediction)
print('prediction')
print_readable_outputs(prediction, nb_words_to_test, nb_chars)
print('check prediction')
y_test = encrypt(x_test, nb_words_to_test, nb_chars)
print("y_test=\n", y_test)
['llhak', 'zupys', 'ymymg']
x_test=
[[108 108 104 97 107]
[122 117 112 121 115]
[121 109 121 109 103]]
x_test_scaled=
[[-0.1944215 -0.20276551 -0.75594591 -1.67757868 -0.33833488]
[ 1.67463093 0.99122139 0.3148938 1.53199765 0.72061624]
[ 1.54112718 -0.0701003 1.51958848 -0.07279052 -0.86781044]]
-->
prediction
['11 11 7 0 10', '25 20 15 24 18', '24 12 24 12 6']
check prediction
y_test=
[[11 11 7 0 10]
[25 20 15 24 18]
[24 12 24 12 6]]
/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:595: DataConversionWarning: Data with input dtype int64 was converted to float64 by StandardScaler.
warnings.warn(msg, DataConversionWarning)
Content source: marxav/hello-world
Similar notebooks: