In [89]:
%matplotlib inline
from keras.models import Sequential
from keras.layers import Dense
from keras.optimizers import SGD, Adam
import numpy as np
import matplotlib.pyplot as plt
In [58]:
# ランダムシードは固定するべし
# 結果を再現・アルゴリズムを比較するため
np.random.seed(7)
In [59]:
!wget http://archive.ics.uci.edu/ml/machine-learning-databases/pima-indians-diabetes/pima-indians-diabetes.data
--2017-05-26 17:02:49-- http://archive.ics.uci.edu/ml/machine-learning-databases/pima-indians-diabetes/pima-indians-diabetes.data
Resolving archive.ics.uci.edu... 128.195.10.249
Connecting to archive.ics.uci.edu|128.195.10.249|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 23279 (23K) [text/plain]
Saving to: ‘pima-indians-diabetes.data.2’
pima-indians-diabet 100%[===================>] 22.73K --.-KB/s in 0.1s
2017-05-26 17:02:49 (174 KB/s) - ‘pima-indians-diabetes.data.2’ saved [23279/23279]
In [60]:
dataset = np.loadtxt('pima-indians-diabetes.data', delimiter=',')
In [61]:
dataset
Out[61]:
array([[ 6. , 148. , 72. , ..., 0.627, 50. , 1. ],
[ 1. , 85. , 66. , ..., 0.351, 31. , 0. ],
[ 8. , 183. , 64. , ..., 0.672, 32. , 1. ],
...,
[ 5. , 121. , 72. , ..., 0.245, 30. , 0. ],
[ 1. , 126. , 60. , ..., 0.349, 47. , 1. ],
[ 1. , 93. , 70. , ..., 0.315, 23. , 0. ]])
In [119]:
X = dataset[:, 0:8]
Y = dataset[:, 8]
In [133]:
from sklearn import preprocessing
X_scaled = preprocessing.scale(X)
print(X_scaled.shape)
print(X_scaled.mean(axis=0))
print(X_scaled.std(axis=0))
(768, 8)
[ -7.74843153e-17 3.61400724e-18 -1.32724416e-17 7.76288755e-17
-5.49329101e-17 2.97273780e-15 1.92438658e-15 2.19297959e-16]
[ 1. 1. 1. 1. 1. 1. 1. 1.]
In [176]:
model = Sequential()
model.add(Dense(12, input_dim=8, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(1, activation='sigmoid'))
model.summary()
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
history = model.fit(X_scaled, Y, epochs=150, batch_size=10)
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_88 (Dense) (None, 12) 108
_________________________________________________________________
dense_89 (Dense) (None, 8) 104
_________________________________________________________________
dense_90 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
Epoch 1/150
768/768 [==============================] - 1s - loss: 0.6448 - acc: 0.6719
Epoch 2/150
768/768 [==============================] - 0s - loss: 0.5721 - acc: 0.7253
Epoch 3/150
768/768 [==============================] - 0s - loss: 0.5276 - acc: 0.7500
Epoch 4/150
768/768 [==============================] - 0s - loss: 0.5052 - acc: 0.7526
Epoch 5/150
768/768 [==============================] - 0s - loss: 0.4934 - acc: 0.7461
Epoch 6/150
768/768 [==============================] - 0s - loss: 0.4866 - acc: 0.7552
Epoch 7/150
768/768 [==============================] - 0s - loss: 0.4821 - acc: 0.7630
Epoch 8/150
768/768 [==============================] - 0s - loss: 0.4775 - acc: 0.7604
Epoch 9/150
768/768 [==============================] - 0s - loss: 0.4754 - acc: 0.7643
Epoch 10/150
768/768 [==============================] - 0s - loss: 0.4721 - acc: 0.7682
Epoch 11/150
768/768 [==============================] - 0s - loss: 0.4700 - acc: 0.7695
Epoch 12/150
768/768 [==============================] - 0s - loss: 0.4676 - acc: 0.7669
Epoch 13/150
768/768 [==============================] - 0s - loss: 0.4667 - acc: 0.7682
Epoch 14/150
768/768 [==============================] - 0s - loss: 0.4643 - acc: 0.7682
Epoch 15/150
768/768 [==============================] - 0s - loss: 0.4631 - acc: 0.7682
Epoch 16/150
768/768 [==============================] - 0s - loss: 0.4615 - acc: 0.7656
Epoch 17/150
768/768 [==============================] - 0s - loss: 0.4607 - acc: 0.7747
Epoch 18/150
768/768 [==============================] - 0s - loss: 0.4604 - acc: 0.7786
Epoch 19/150
768/768 [==============================] - 0s - loss: 0.4576 - acc: 0.7773
Epoch 20/150
768/768 [==============================] - 0s - loss: 0.4566 - acc: 0.7812
Epoch 21/150
768/768 [==============================] - 0s - loss: 0.4553 - acc: 0.7826
Epoch 22/150
768/768 [==============================] - 0s - loss: 0.4549 - acc: 0.7773
Epoch 23/150
768/768 [==============================] - 0s - loss: 0.4531 - acc: 0.7839
Epoch 24/150
768/768 [==============================] - 0s - loss: 0.4528 - acc: 0.7826
Epoch 25/150
768/768 [==============================] - 0s - loss: 0.4515 - acc: 0.7839
Epoch 26/150
768/768 [==============================] - 0s - loss: 0.4515 - acc: 0.7839
Epoch 27/150
768/768 [==============================] - 0s - loss: 0.4502 - acc: 0.7852
Epoch 28/150
768/768 [==============================] - 0s - loss: 0.4489 - acc: 0.7891
Epoch 29/150
768/768 [==============================] - 0s - loss: 0.4484 - acc: 0.7878
Epoch 30/150
768/768 [==============================] - 0s - loss: 0.4474 - acc: 0.7852
Epoch 31/150
768/768 [==============================] - 0s - loss: 0.4466 - acc: 0.7891
Epoch 32/150
768/768 [==============================] - 0s - loss: 0.4460 - acc: 0.7865
Epoch 33/150
768/768 [==============================] - 0s - loss: 0.4450 - acc: 0.7865
Epoch 34/150
768/768 [==============================] - 0s - loss: 0.4437 - acc: 0.7852
Epoch 35/150
768/768 [==============================] - 0s - loss: 0.4428 - acc: 0.7891
Epoch 36/150
768/768 [==============================] - 0s - loss: 0.4421 - acc: 0.7878
Epoch 37/150
768/768 [==============================] - 0s - loss: 0.4415 - acc: 0.7826
Epoch 38/150
768/768 [==============================] - 0s - loss: 0.4411 - acc: 0.7865
Epoch 39/150
768/768 [==============================] - 0s - loss: 0.4403 - acc: 0.7891
Epoch 40/150
768/768 [==============================] - 0s - loss: 0.4399 - acc: 0.7878
Epoch 41/150
768/768 [==============================] - 0s - loss: 0.4392 - acc: 0.7930
Epoch 42/150
768/768 [==============================] - 0s - loss: 0.4386 - acc: 0.7878
Epoch 43/150
768/768 [==============================] - 0s - loss: 0.4381 - acc: 0.7930
Epoch 44/150
768/768 [==============================] - 0s - loss: 0.4374 - acc: 0.7930
Epoch 45/150
768/768 [==============================] - 0s - loss: 0.4375 - acc: 0.7904
Epoch 46/150
768/768 [==============================] - 0s - loss: 0.4364 - acc: 0.7904
Epoch 47/150
768/768 [==============================] - 0s - loss: 0.4352 - acc: 0.7969
Epoch 48/150
768/768 [==============================] - 0s - loss: 0.4343 - acc: 0.7943
Epoch 49/150
768/768 [==============================] - 0s - loss: 0.4345 - acc: 0.7969
Epoch 50/150
768/768 [==============================] - 0s - loss: 0.4335 - acc: 0.7943
Epoch 51/150
768/768 [==============================] - 0s - loss: 0.4331 - acc: 0.7982
Epoch 52/150
768/768 [==============================] - 0s - loss: 0.4330 - acc: 0.7969
Epoch 53/150
768/768 [==============================] - 0s - loss: 0.4326 - acc: 0.7891
Epoch 54/150
768/768 [==============================] - 0s - loss: 0.4318 - acc: 0.7904
Epoch 55/150
768/768 [==============================] - 0s - loss: 0.4316 - acc: 0.7995
Epoch 56/150
768/768 [==============================] - 0s - loss: 0.4313 - acc: 0.7943
Epoch 57/150
768/768 [==============================] - 0s - loss: 0.4312 - acc: 0.7917
Epoch 58/150
768/768 [==============================] - 0s - loss: 0.4304 - acc: 0.7995
Epoch 59/150
768/768 [==============================] - 0s - loss: 0.4296 - acc: 0.7995
Epoch 60/150
768/768 [==============================] - 0s - loss: 0.4295 - acc: 0.8021
Epoch 61/150
768/768 [==============================] - 0s - loss: 0.4289 - acc: 0.7995
Epoch 62/150
768/768 [==============================] - 0s - loss: 0.4290 - acc: 0.7943
Epoch 63/150
768/768 [==============================] - 0s - loss: 0.4296 - acc: 0.8008
Epoch 64/150
768/768 [==============================] - 0s - loss: 0.4279 - acc: 0.7917
Epoch 65/150
768/768 [==============================] - 0s - loss: 0.4279 - acc: 0.7982
Epoch 66/150
768/768 [==============================] - 0s - loss: 0.4276 - acc: 0.7982
Epoch 67/150
768/768 [==============================] - 0s - loss: 0.4270 - acc: 0.7969
Epoch 68/150
768/768 [==============================] - 0s - loss: 0.4269 - acc: 0.7982
Epoch 69/150
768/768 [==============================] - 0s - loss: 0.4260 - acc: 0.7982
Epoch 70/150
768/768 [==============================] - 0s - loss: 0.4257 - acc: 0.7956
Epoch 71/150
768/768 [==============================] - 0s - loss: 0.4254 - acc: 0.8021
Epoch 72/150
768/768 [==============================] - 0s - loss: 0.4264 - acc: 0.7982
Epoch 73/150
768/768 [==============================] - 0s - loss: 0.4259 - acc: 0.7969
Epoch 74/150
768/768 [==============================] - 0s - loss: 0.4247 - acc: 0.8060
Epoch 75/150
768/768 [==============================] - 0s - loss: 0.4258 - acc: 0.7995
Epoch 76/150
768/768 [==============================] - 0s - loss: 0.4246 - acc: 0.7995
Epoch 77/150
768/768 [==============================] - 0s - loss: 0.4248 - acc: 0.7956
Epoch 78/150
768/768 [==============================] - 0s - loss: 0.4240 - acc: 0.8047
Epoch 79/150
768/768 [==============================] - 0s - loss: 0.4235 - acc: 0.7969
Epoch 80/150
768/768 [==============================] - 0s - loss: 0.4232 - acc: 0.7995
Epoch 81/150
768/768 [==============================] - 0s - loss: 0.4234 - acc: 0.7982
Epoch 82/150
768/768 [==============================] - 0s - loss: 0.4238 - acc: 0.8021
Epoch 83/150
768/768 [==============================] - 0s - loss: 0.4227 - acc: 0.7969
Epoch 84/150
768/768 [==============================] - 0s - loss: 0.4222 - acc: 0.8008
Epoch 85/150
768/768 [==============================] - 0s - loss: 0.4218 - acc: 0.7995
Epoch 86/150
768/768 [==============================] - 0s - loss: 0.4220 - acc: 0.7956
Epoch 87/150
768/768 [==============================] - 0s - loss: 0.4215 - acc: 0.7969
Epoch 88/150
768/768 [==============================] - 0s - loss: 0.4214 - acc: 0.7982
Epoch 89/150
768/768 [==============================] - 0s - loss: 0.4212 - acc: 0.8008
Epoch 90/150
768/768 [==============================] - 0s - loss: 0.4213 - acc: 0.7995
Epoch 91/150
768/768 [==============================] - 0s - loss: 0.4203 - acc: 0.7969
Epoch 92/150
768/768 [==============================] - 0s - loss: 0.4197 - acc: 0.7943
Epoch 93/150
768/768 [==============================] - 0s - loss: 0.4210 - acc: 0.8021
Epoch 94/150
768/768 [==============================] - 0s - loss: 0.4200 - acc: 0.7969
Epoch 95/150
768/768 [==============================] - 0s - loss: 0.4192 - acc: 0.8008
Epoch 96/150
768/768 [==============================] - 0s - loss: 0.4196 - acc: 0.7982
Epoch 97/150
768/768 [==============================] - 0s - loss: 0.4189 - acc: 0.7969
Epoch 98/150
768/768 [==============================] - 0s - loss: 0.4182 - acc: 0.7995
Epoch 99/150
768/768 [==============================] - 0s - loss: 0.4176 - acc: 0.7982
Epoch 100/150
768/768 [==============================] - 0s - loss: 0.4173 - acc: 0.8021
Epoch 101/150
768/768 [==============================] - 0s - loss: 0.4171 - acc: 0.7969
Epoch 102/150
768/768 [==============================] - 0s - loss: 0.4171 - acc: 0.8021
Epoch 103/150
768/768 [==============================] - 0s - loss: 0.4170 - acc: 0.7995
Epoch 104/150
768/768 [==============================] - 0s - loss: 0.4155 - acc: 0.8060
Epoch 105/150
768/768 [==============================] - 0s - loss: 0.4161 - acc: 0.8008
Epoch 106/150
768/768 [==============================] - 0s - loss: 0.4160 - acc: 0.8047
Epoch 107/150
768/768 [==============================] - 0s - loss: 0.4150 - acc: 0.8021
Epoch 108/150
768/768 [==============================] - 0s - loss: 0.4152 - acc: 0.8034
Epoch 109/150
768/768 [==============================] - 0s - loss: 0.4147 - acc: 0.8021
Epoch 110/150
768/768 [==============================] - 0s - loss: 0.4140 - acc: 0.8021
Epoch 111/150
768/768 [==============================] - 0s - loss: 0.4137 - acc: 0.8034
Epoch 112/150
768/768 [==============================] - 0s - loss: 0.4143 - acc: 0.8008
Epoch 113/150
768/768 [==============================] - 0s - loss: 0.4130 - acc: 0.8008
Epoch 114/150
768/768 [==============================] - 0s - loss: 0.4125 - acc: 0.7995
Epoch 115/150
768/768 [==============================] - 0s - loss: 0.4129 - acc: 0.8021
Epoch 116/150
768/768 [==============================] - 0s - loss: 0.4117 - acc: 0.8008
Epoch 117/150
768/768 [==============================] - 0s - loss: 0.4120 - acc: 0.7956
Epoch 118/150
768/768 [==============================] - 0s - loss: 0.4112 - acc: 0.7969
Epoch 119/150
768/768 [==============================] - 0s - loss: 0.4108 - acc: 0.8034
Epoch 120/150
768/768 [==============================] - 0s - loss: 0.4103 - acc: 0.7995
Epoch 121/150
768/768 [==============================] - 0s - loss: 0.4110 - acc: 0.8021
Epoch 122/150
768/768 [==============================] - 0s - loss: 0.4107 - acc: 0.7943
Epoch 123/150
768/768 [==============================] - 0s - loss: 0.4095 - acc: 0.8060
Epoch 124/150
768/768 [==============================] - 0s - loss: 0.4086 - acc: 0.7982
Epoch 125/150
768/768 [==============================] - 0s - loss: 0.4102 - acc: 0.8073
Epoch 126/150
768/768 [==============================] - 0s - loss: 0.4079 - acc: 0.8034
Epoch 127/150
768/768 [==============================] - 0s - loss: 0.4082 - acc: 0.8047
Epoch 128/150
768/768 [==============================] - 0s - loss: 0.4078 - acc: 0.7995
Epoch 129/150
768/768 [==============================] - 0s - loss: 0.4080 - acc: 0.8034
Epoch 130/150
768/768 [==============================] - 0s - loss: 0.4065 - acc: 0.8034
Epoch 131/150
768/768 [==============================] - 0s - loss: 0.4062 - acc: 0.8021
Epoch 132/150
768/768 [==============================] - 0s - loss: 0.4064 - acc: 0.8073
Epoch 133/150
768/768 [==============================] - 0s - loss: 0.4059 - acc: 0.8021
Epoch 134/150
768/768 [==============================] - 0s - loss: 0.4061 - acc: 0.8008
Epoch 135/150
768/768 [==============================] - 0s - loss: 0.4048 - acc: 0.8047
Epoch 136/150
768/768 [==============================] - 0s - loss: 0.4051 - acc: 0.8034
Epoch 137/150
768/768 [==============================] - 0s - loss: 0.4043 - acc: 0.7995
Epoch 138/150
768/768 [==============================] - 0s - loss: 0.4034 - acc: 0.8086
Epoch 139/150
768/768 [==============================] - 0s - loss: 0.4038 - acc: 0.8060
Epoch 140/150
768/768 [==============================] - 0s - loss: 0.4040 - acc: 0.8034
Epoch 141/150
768/768 [==============================] - 0s - loss: 0.4040 - acc: 0.8073
Epoch 142/150
768/768 [==============================] - 0s - loss: 0.4029 - acc: 0.8086
Epoch 143/150
768/768 [==============================] - 0s - loss: 0.4028 - acc: 0.8099
Epoch 144/150
768/768 [==============================] - 0s - loss: 0.4042 - acc: 0.8099
Epoch 145/150
768/768 [==============================] - 0s - loss: 0.4018 - acc: 0.8099
Epoch 146/150
768/768 [==============================] - 0s - loss: 0.4018 - acc: 0.8112
Epoch 147/150
768/768 [==============================] - 0s - loss: 0.4018 - acc: 0.8060
Epoch 148/150
768/768 [==============================] - 0s - loss: 0.4020 - acc: 0.8125
Epoch 149/150
768/768 [==============================] - 0s - loss: 0.4009 - acc: 0.8086
Epoch 150/150
768/768 [==============================] - 0s - loss: 0.4002 - acc: 0.8151
In [138]:
scores = model.evaluate(X_scaled, Y)
32/768 [>.............................] - ETA: 0s
In [139]:
print('%s: %.2f%%' % (model.metrics_names[1], scores[1] * 100))
acc: 83.72%
In [143]:
predictions = model.predict(X_scaled)
In [144]:
rounded = [round(x[0]) for x in predictions]
In [148]:
print(rounded[:10])
print(Y[:10])
[1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, 1.0]
[ 1. 0. 1. 0. 1. 0. 1. 0. 1. 1.]
In [157]:
from keras.layers import Input, Dense
from keras.models import Model
inputs = Input(shape=(8, ))
x = Dense(12, activation='relu')(inputs)
x = Dense(8, activation='relu')(x)
predictions = Dense(1, activation='sigmoid')(x)
model = Model(inputs=inputs, outputs=predictions)
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_9 (InputLayer) (None, 8) 0
_________________________________________________________________
dense_85 (Dense) (None, 12) 108
_________________________________________________________________
dense_86 (Dense) (None, 8) 104
_________________________________________________________________
dense_87 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
In [159]:
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
history = model.fit(X_scaled, Y, epochs=150, batch_size=10)
Epoch 1/150
768/768 [==============================] - 1s - loss: 0.4722 - acc: 0.7799
Epoch 2/150
768/768 [==============================] - 0s - loss: 0.4649 - acc: 0.7826
Epoch 3/150
768/768 [==============================] - 0s - loss: 0.4604 - acc: 0.7852
Epoch 4/150
768/768 [==============================] - 0s - loss: 0.4575 - acc: 0.7839
Epoch 5/150
768/768 [==============================] - 0s - loss: 0.4551 - acc: 0.7865
Epoch 6/150
768/768 [==============================] - 0s - loss: 0.4518 - acc: 0.7852
Epoch 7/150
768/768 [==============================] - 0s - loss: 0.4497 - acc: 0.7839
Epoch 8/150
768/768 [==============================] - 0s - loss: 0.4485 - acc: 0.7799
Epoch 9/150
768/768 [==============================] - 0s - loss: 0.4467 - acc: 0.7812
Epoch 10/150
768/768 [==============================] - 0s - loss: 0.4451 - acc: 0.7786
Epoch 11/150
768/768 [==============================] - 0s - loss: 0.4432 - acc: 0.7839
Epoch 12/150
768/768 [==============================] - 0s - loss: 0.4423 - acc: 0.7852
Epoch 13/150
768/768 [==============================] - 0s - loss: 0.4405 - acc: 0.7865
Epoch 14/150
768/768 [==============================] - 0s - loss: 0.4393 - acc: 0.7852
Epoch 15/150
768/768 [==============================] - 0s - loss: 0.4386 - acc: 0.7891
Epoch 16/150
768/768 [==============================] - 0s - loss: 0.4375 - acc: 0.7930
Epoch 17/150
768/768 [==============================] - 0s - loss: 0.4372 - acc: 0.7865
Epoch 18/150
768/768 [==============================] - 0s - loss: 0.4351 - acc: 0.7930
Epoch 19/150
768/768 [==============================] - 0s - loss: 0.4341 - acc: 0.7917
Epoch 20/150
768/768 [==============================] - 0s - loss: 0.4333 - acc: 0.7917
Epoch 21/150
768/768 [==============================] - 0s - loss: 0.4319 - acc: 0.7930
Epoch 22/150
768/768 [==============================] - 0s - loss: 0.4311 - acc: 0.7956
Epoch 23/150
768/768 [==============================] - 0s - loss: 0.4306 - acc: 0.7917
Epoch 24/150
768/768 [==============================] - 0s - loss: 0.4304 - acc: 0.7943
Epoch 25/150
768/768 [==============================] - 0s - loss: 0.4286 - acc: 0.7956
Epoch 26/150
768/768 [==============================] - 0s - loss: 0.4283 - acc: 0.7995
Epoch 27/150
768/768 [==============================] - 0s - loss: 0.4273 - acc: 0.7943
Epoch 28/150
768/768 [==============================] - 0s - loss: 0.4267 - acc: 0.7969
Epoch 29/150
768/768 [==============================] - 0s - loss: 0.4256 - acc: 0.7969
Epoch 30/150
768/768 [==============================] - 0s - loss: 0.4249 - acc: 0.7956
Epoch 31/150
768/768 [==============================] - 0s - loss: 0.4239 - acc: 0.7956
Epoch 32/150
768/768 [==============================] - 0s - loss: 0.4239 - acc: 0.8021
Epoch 33/150
768/768 [==============================] - 0s - loss: 0.4218 - acc: 0.8008
Epoch 34/150
768/768 [==============================] - 0s - loss: 0.4211 - acc: 0.7969
Epoch 35/150
768/768 [==============================] - 0s - loss: 0.4203 - acc: 0.7969
Epoch 36/150
768/768 [==============================] - 0s - loss: 0.4194 - acc: 0.7982
Epoch 37/150
768/768 [==============================] - 0s - loss: 0.4188 - acc: 0.7995
Epoch 38/150
768/768 [==============================] - 0s - loss: 0.4186 - acc: 0.8047
Epoch 39/150
768/768 [==============================] - 0s - loss: 0.4188 - acc: 0.8047
Epoch 40/150
768/768 [==============================] - 0s - loss: 0.4173 - acc: 0.8021
Epoch 41/150
768/768 [==============================] - 0s - loss: 0.4165 - acc: 0.8060
Epoch 42/150
768/768 [==============================] - 0s - loss: 0.4156 - acc: 0.8060
Epoch 43/150
768/768 [==============================] - 0s - loss: 0.4151 - acc: 0.8047
Epoch 44/150
768/768 [==============================] - 0s - loss: 0.4148 - acc: 0.8034
Epoch 45/150
768/768 [==============================] - 0s - loss: 0.4145 - acc: 0.8060
Epoch 46/150
768/768 [==============================] - 0s - loss: 0.4136 - acc: 0.8086
Epoch 47/150
768/768 [==============================] - 0s - loss: 0.4131 - acc: 0.8125
Epoch 48/150
768/768 [==============================] - 0s - loss: 0.4124 - acc: 0.8099
Epoch 49/150
768/768 [==============================] - 0s - loss: 0.4123 - acc: 0.8099
Epoch 50/150
768/768 [==============================] - 0s - loss: 0.4112 - acc: 0.8060
Epoch 51/150
768/768 [==============================] - 0s - loss: 0.4108 - acc: 0.8125
Epoch 52/150
768/768 [==============================] - 0s - loss: 0.4112 - acc: 0.8138
Epoch 53/150
768/768 [==============================] - 0s - loss: 0.4101 - acc: 0.8151
Epoch 54/150
768/768 [==============================] - 0s - loss: 0.4105 - acc: 0.8086
Epoch 55/150
768/768 [==============================] - 0s - loss: 0.4086 - acc: 0.8138
Epoch 56/150
768/768 [==============================] - 0s - loss: 0.4099 - acc: 0.8112
Epoch 57/150
768/768 [==============================] - 0s - loss: 0.4080 - acc: 0.8125
Epoch 58/150
768/768 [==============================] - 0s - loss: 0.4073 - acc: 0.8177
Epoch 59/150
768/768 [==============================] - 0s - loss: 0.4069 - acc: 0.8164
Epoch 60/150
768/768 [==============================] - 0s - loss: 0.4066 - acc: 0.8099
Epoch 61/150
768/768 [==============================] - 0s - loss: 0.4072 - acc: 0.8099
Epoch 62/150
768/768 [==============================] - 0s - loss: 0.4064 - acc: 0.8125
Epoch 63/150
768/768 [==============================] - 0s - loss: 0.4054 - acc: 0.8125
Epoch 64/150
768/768 [==============================] - 0s - loss: 0.4055 - acc: 0.8164
Epoch 65/150
768/768 [==============================] - 0s - loss: 0.4048 - acc: 0.8125
Epoch 66/150
768/768 [==============================] - 0s - loss: 0.4039 - acc: 0.8151
Epoch 67/150
768/768 [==============================] - 0s - loss: 0.4044 - acc: 0.8151
Epoch 68/150
768/768 [==============================] - 0s - loss: 0.4039 - acc: 0.8164
Epoch 69/150
768/768 [==============================] - 0s - loss: 0.4039 - acc: 0.8203
Epoch 70/150
768/768 [==============================] - 0s - loss: 0.4023 - acc: 0.8190
Epoch 71/150
768/768 [==============================] - 0s - loss: 0.4022 - acc: 0.8164
Epoch 72/150
768/768 [==============================] - 0s - loss: 0.4018 - acc: 0.8151
Epoch 73/150
768/768 [==============================] - 0s - loss: 0.4021 - acc: 0.8151
Epoch 74/150
768/768 [==============================] - 0s - loss: 0.4016 - acc: 0.8177
Epoch 75/150
768/768 [==============================] - 0s - loss: 0.4017 - acc: 0.8151
Epoch 76/150
768/768 [==============================] - 0s - loss: 0.4005 - acc: 0.8112
Epoch 77/150
768/768 [==============================] - 0s - loss: 0.3997 - acc: 0.8203
Epoch 78/150
768/768 [==============================] - 0s - loss: 0.3994 - acc: 0.8177
Epoch 79/150
768/768 [==============================] - 0s - loss: 0.4003 - acc: 0.8190
Epoch 80/150
768/768 [==============================] - 0s - loss: 0.3986 - acc: 0.8177
Epoch 81/150
768/768 [==============================] - 0s - loss: 0.3987 - acc: 0.8190
Epoch 82/150
768/768 [==============================] - 0s - loss: 0.3982 - acc: 0.8164
Epoch 83/150
768/768 [==============================] - 0s - loss: 0.3998 - acc: 0.8164
Epoch 84/150
768/768 [==============================] - 0s - loss: 0.3980 - acc: 0.8177
Epoch 85/150
768/768 [==============================] - 0s - loss: 0.3978 - acc: 0.8164
Epoch 86/150
768/768 [==============================] - 0s - loss: 0.3962 - acc: 0.8190
Epoch 87/150
768/768 [==============================] - 0s - loss: 0.3964 - acc: 0.8203
Epoch 88/150
768/768 [==============================] - 0s - loss: 0.3960 - acc: 0.8099
Epoch 89/150
768/768 [==============================] - 0s - loss: 0.3962 - acc: 0.8216
Epoch 90/150
768/768 [==============================] - 0s - loss: 0.3947 - acc: 0.8151
Epoch 91/150
768/768 [==============================] - 0s - loss: 0.3950 - acc: 0.8177
Epoch 92/150
768/768 [==============================] - 0s - loss: 0.3962 - acc: 0.8177
Epoch 93/150
768/768 [==============================] - 0s - loss: 0.3939 - acc: 0.8216
Epoch 94/150
768/768 [==============================] - 0s - loss: 0.3932 - acc: 0.8190
Epoch 95/150
768/768 [==============================] - 0s - loss: 0.3922 - acc: 0.8190
Epoch 96/150
768/768 [==============================] - 0s - loss: 0.3920 - acc: 0.8177
Epoch 97/150
768/768 [==============================] - 0s - loss: 0.3922 - acc: 0.8164
Epoch 98/150
768/768 [==============================] - 0s - loss: 0.3921 - acc: 0.8177
Epoch 99/150
768/768 [==============================] - 0s - loss: 0.3907 - acc: 0.8190
Epoch 100/150
768/768 [==============================] - 0s - loss: 0.3916 - acc: 0.8190
Epoch 101/150
768/768 [==============================] - 0s - loss: 0.3900 - acc: 0.8203
Epoch 102/150
768/768 [==============================] - 0s - loss: 0.3896 - acc: 0.8177
Epoch 103/150
768/768 [==============================] - 0s - loss: 0.3899 - acc: 0.8177
Epoch 104/150
768/768 [==============================] - 0s - loss: 0.3883 - acc: 0.8229
Epoch 105/150
768/768 [==============================] - 0s - loss: 0.3876 - acc: 0.8216
Epoch 106/150
768/768 [==============================] - 0s - loss: 0.3877 - acc: 0.8203
Epoch 107/150
768/768 [==============================] - 0s - loss: 0.3872 - acc: 0.8229
Epoch 108/150
768/768 [==============================] - 0s - loss: 0.3872 - acc: 0.8190
Epoch 109/150
768/768 [==============================] - 0s - loss: 0.3862 - acc: 0.8190
Epoch 110/150
768/768 [==============================] - 0s - loss: 0.3872 - acc: 0.8177
Epoch 111/150
768/768 [==============================] - 0s - loss: 0.3866 - acc: 0.8216
Epoch 112/150
768/768 [==============================] - 0s - loss: 0.3845 - acc: 0.8190
Epoch 113/150
768/768 [==============================] - 0s - loss: 0.3831 - acc: 0.8216
Epoch 114/150
768/768 [==============================] - 0s - loss: 0.3848 - acc: 0.8216
Epoch 115/150
768/768 [==============================] - 0s - loss: 0.3839 - acc: 0.8255
Epoch 116/150
768/768 [==============================] - 0s - loss: 0.3838 - acc: 0.8216
Epoch 117/150
768/768 [==============================] - 0s - loss: 0.3834 - acc: 0.8203
Epoch 118/150
768/768 [==============================] - 0s - loss: 0.3814 - acc: 0.8216
Epoch 119/150
768/768 [==============================] - 0s - loss: 0.3825 - acc: 0.8216
Epoch 120/150
768/768 [==============================] - 0s - loss: 0.3825 - acc: 0.8216
Epoch 121/150
768/768 [==============================] - 0s - loss: 0.3818 - acc: 0.8216
Epoch 122/150
768/768 [==============================] - 0s - loss: 0.3814 - acc: 0.8203
Epoch 123/150
768/768 [==============================] - 0s - loss: 0.3804 - acc: 0.8216
Epoch 124/150
768/768 [==============================] - 0s - loss: 0.3798 - acc: 0.8203
Epoch 125/150
768/768 [==============================] - 0s - loss: 0.3799 - acc: 0.8164
Epoch 126/150
768/768 [==============================] - 0s - loss: 0.3799 - acc: 0.8190
Epoch 127/150
768/768 [==============================] - 0s - loss: 0.3798 - acc: 0.8242
Epoch 128/150
768/768 [==============================] - 0s - loss: 0.3789 - acc: 0.8268
Epoch 129/150
768/768 [==============================] - 0s - loss: 0.3789 - acc: 0.8268
Epoch 130/150
768/768 [==============================] - 0s - loss: 0.3778 - acc: 0.8255
Epoch 131/150
768/768 [==============================] - 0s - loss: 0.3770 - acc: 0.8216
Epoch 132/150
768/768 [==============================] - 0s - loss: 0.3778 - acc: 0.8242
Epoch 133/150
768/768 [==============================] - 0s - loss: 0.3767 - acc: 0.8268
Epoch 134/150
768/768 [==============================] - 0s - loss: 0.3781 - acc: 0.8294
Epoch 135/150
768/768 [==============================] - 0s - loss: 0.3770 - acc: 0.8255
Epoch 136/150
768/768 [==============================] - 0s - loss: 0.3758 - acc: 0.8242
Epoch 137/150
768/768 [==============================] - 0s - loss: 0.3755 - acc: 0.8268
Epoch 138/150
768/768 [==============================] - 0s - loss: 0.3748 - acc: 0.8281
Epoch 139/150
768/768 [==============================] - 0s - loss: 0.3742 - acc: 0.8255
Epoch 140/150
768/768 [==============================] - 0s - loss: 0.3740 - acc: 0.8320
Epoch 141/150
768/768 [==============================] - 0s - loss: 0.3738 - acc: 0.8229
Epoch 142/150
768/768 [==============================] - 0s - loss: 0.3746 - acc: 0.8294
Epoch 143/150
768/768 [==============================] - 0s - loss: 0.3725 - acc: 0.8320
Epoch 144/150
768/768 [==============================] - 0s - loss: 0.3724 - acc: 0.8307
Epoch 145/150
768/768 [==============================] - 0s - loss: 0.3717 - acc: 0.8359
Epoch 146/150
768/768 [==============================] - 0s - loss: 0.3728 - acc: 0.8320
Epoch 147/150
768/768 [==============================] - 0s - loss: 0.3716 - acc: 0.8333
Epoch 148/150
768/768 [==============================] - 0s - loss: 0.3719 - acc: 0.8268
Epoch 149/150
768/768 [==============================] - 0s - loss: 0.3715 - acc: 0.8320
Epoch 150/150
768/768 [==============================] - 0s - loss: 0.3712 - acc: 0.8294
In [165]:
model_json = model.to_json()
with open('model.json', 'w') as json_file:
json_file.write(model_json)
model.save_weights('model.h5')
In [167]:
del model
In [170]:
from keras.models import model_from_json
json_file = open('model.json', 'r')
loaded_model_json = json_file.read()
json_file.close()
loaded_model = model_from_json(loaded_model_json)
loaded_model.load_weights('model.h5')
In [172]:
loaded_model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_9 (InputLayer) (None, 8) 0
_________________________________________________________________
dense_85 (Dense) (None, 12) 108
_________________________________________________________________
dense_86 (Dense) (None, 8) 104
_________________________________________________________________
dense_87 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
In [175]:
del loaded_model
In [177]:
model.save('mymodel.h5')
In [178]:
del model
In [180]:
from keras.models import load_model
model = load_model('mymodel.h5')
In [181]:
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_88 (Dense) (None, 12) 108
_________________________________________________________________
dense_89 (Dense) (None, 8) 104
_________________________________________________________________
dense_90 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
In [183]:
model.compile(loss='binary_crossentropy', optimizer='sgd', metrics=['accuracy'])
In [184]:
model.optimizer
Out[184]:
<keras.optimizers.SGD at 0x137fc0198>
Content source: aidiary/notebooks
Similar notebooks: