In [1]:
# %load /home/sjkim/.jupyter/head.py
# %%writefile /home/sjkim/.jupyter/head.py
# %load /home/sjkim/.jupyter/head.py
%matplotlib inline
%load_ext autoreload
%autoreload 2
from importlib import reload
import matplotlib.pyplot as plt
import numpy as np
In [2]:
from keraspp import ch4_3_cnn_cifar10
reload(ch4_3_cnn_cifar10)
Using TensorFlow backend.
Out[2]:
<module 'keraspp.ch4_3_cnn_cifar10' from '/home/sjkim/Dropbox/Aspuru-Guzik/python_lab/py3/keraspp/keraspp/ch4_3_cnn_cifar10.py'>
In [3]:
m = ch4_3_cnn_cifar10.Machine()
m.run()
(40000, 32, 32, 3) (40000, 1)
X_train shape: (40000, 32, 32, 3)
40000 train samples
10000 test samples
data.input_shape (32, 32, 3)
Train on 40000 samples, validate on 10000 samples
Epoch 1/100
40000/40000 [==============================] - 92s - loss: 1.8967 - acc: 0.3136 - val_loss: 1.5593 - val_acc: 0.4473
Epoch 2/100
40000/40000 [==============================] - 3s - loss: 1.5012 - acc: 0.4640 - val_loss: 1.3153 - val_acc: 0.5383
Epoch 3/100
40000/40000 [==============================] - 3s - loss: 1.3343 - acc: 0.5249 - val_loss: 1.2203 - val_acc: 0.5609
Epoch 4/100
40000/40000 [==============================] - 3s - loss: 1.2235 - acc: 0.5675 - val_loss: 1.1701 - val_acc: 0.5869
Epoch 5/100
40000/40000 [==============================] - 3s - loss: 1.1436 - acc: 0.6000 - val_loss: 1.0937 - val_acc: 0.6165
Epoch 6/100
40000/40000 [==============================] - 3s - loss: 1.0835 - acc: 0.6194 - val_loss: 1.0542 - val_acc: 0.6269
Epoch 7/100
40000/40000 [==============================] - 3s - loss: 1.0253 - acc: 0.6414 - val_loss: 0.9890 - val_acc: 0.6532
Epoch 8/100
40000/40000 [==============================] - 3s - loss: 0.9859 - acc: 0.6557 - val_loss: 1.0044 - val_acc: 0.6494
Epoch 9/100
40000/40000 [==============================] - 3s - loss: 0.9448 - acc: 0.6703 - val_loss: 0.9654 - val_acc: 0.6571
Epoch 10/100
40000/40000 [==============================] - 3s - loss: 0.9041 - acc: 0.6854 - val_loss: 0.9898 - val_acc: 0.6523
Epoch 11/100
40000/40000 [==============================] - 3s - loss: 0.8653 - acc: 0.6989 - val_loss: 0.9197 - val_acc: 0.6796
Epoch 12/100
40000/40000 [==============================] - 3s - loss: 0.8323 - acc: 0.7090 - val_loss: 0.9159 - val_acc: 0.6753
Epoch 13/100
40000/40000 [==============================] - 3s - loss: 0.8043 - acc: 0.7202 - val_loss: 0.8886 - val_acc: 0.6894
Epoch 14/100
40000/40000 [==============================] - 3s - loss: 0.7768 - acc: 0.7291 - val_loss: 0.8856 - val_acc: 0.6936
Epoch 15/100
40000/40000 [==============================] - 3s - loss: 0.7443 - acc: 0.7400 - val_loss: 0.8778 - val_acc: 0.6947
Epoch 16/100
40000/40000 [==============================] - 3s - loss: 0.7250 - acc: 0.7461 - val_loss: 0.8708 - val_acc: 0.6969
Epoch 17/100
40000/40000 [==============================] - 3s - loss: 0.7010 - acc: 0.7540 - val_loss: 0.8652 - val_acc: 0.7005
Epoch 18/100
40000/40000 [==============================] - 3s - loss: 0.6754 - acc: 0.7605 - val_loss: 0.9107 - val_acc: 0.6891
Epoch 19/100
40000/40000 [==============================] - 3s - loss: 0.6550 - acc: 0.7706 - val_loss: 0.8904 - val_acc: 0.6976
Epoch 20/100
40000/40000 [==============================] - 3s - loss: 0.6299 - acc: 0.7787 - val_loss: 0.8515 - val_acc: 0.7084
Epoch 21/100
40000/40000 [==============================] - 3s - loss: 0.6027 - acc: 0.7898 - val_loss: 0.8766 - val_acc: 0.7013
Epoch 22/100
40000/40000 [==============================] - 3s - loss: 0.5925 - acc: 0.7932 - val_loss: 0.8590 - val_acc: 0.7071
Epoch 23/100
40000/40000 [==============================] - 3s - loss: 0.5749 - acc: 0.7988 - val_loss: 0.8924 - val_acc: 0.7009
Epoch 24/100
40000/40000 [==============================] - 3s - loss: 0.5591 - acc: 0.8051 - val_loss: 0.8720 - val_acc: 0.7051
Epoch 25/100
40000/40000 [==============================] - 3s - loss: 0.5386 - acc: 0.8107 - val_loss: 0.9126 - val_acc: 0.7078
Epoch 26/100
40000/40000 [==============================] - 3s - loss: 0.5208 - acc: 0.8185 - val_loss: 0.8702 - val_acc: 0.7080
Epoch 27/100
40000/40000 [==============================] - 3s - loss: 0.5137 - acc: 0.8186 - val_loss: 0.8656 - val_acc: 0.7082
Epoch 28/100
40000/40000 [==============================] - 3s - loss: 0.4997 - acc: 0.8252 - val_loss: 0.8761 - val_acc: 0.7134
Epoch 29/100
40000/40000 [==============================] - 3s - loss: 0.4829 - acc: 0.8307 - val_loss: 0.9071 - val_acc: 0.7068
Epoch 30/100
40000/40000 [==============================] - 3s - loss: 0.4749 - acc: 0.8335 - val_loss: 0.8904 - val_acc: 0.7141
Epoch 31/100
40000/40000 [==============================] - 3s - loss: 0.4646 - acc: 0.8389 - val_loss: 0.9134 - val_acc: 0.7088
Epoch 32/100
40000/40000 [==============================] - 3s - loss: 0.4510 - acc: 0.8407 - val_loss: 0.9080 - val_acc: 0.7101
Epoch 33/100
40000/40000 [==============================] - 3s - loss: 0.4431 - acc: 0.8467 - val_loss: 0.8726 - val_acc: 0.7131
Epoch 34/100
40000/40000 [==============================] - 3s - loss: 0.4317 - acc: 0.8498 - val_loss: 0.9201 - val_acc: 0.7118
Epoch 35/100
40000/40000 [==============================] - 3s - loss: 0.4226 - acc: 0.8511 - val_loss: 0.9330 - val_acc: 0.7151
Epoch 36/100
40000/40000 [==============================] - 3s - loss: 0.4218 - acc: 0.8529 - val_loss: 0.9559 - val_acc: 0.7156
Epoch 37/100
40000/40000 [==============================] - 3s - loss: 0.4082 - acc: 0.8537 - val_loss: 0.9942 - val_acc: 0.7155
Epoch 38/100
40000/40000 [==============================] - 3s - loss: 0.4009 - acc: 0.8599 - val_loss: 0.9542 - val_acc: 0.7056
Epoch 39/100
40000/40000 [==============================] - 3s - loss: 0.3918 - acc: 0.8634 - val_loss: 0.9114 - val_acc: 0.7084
Epoch 40/100
40000/40000 [==============================] - 3s - loss: 0.3915 - acc: 0.8649 - val_loss: 1.0364 - val_acc: 0.7142
Epoch 41/100
40000/40000 [==============================] - 3s - loss: 0.3863 - acc: 0.8674 - val_loss: 0.9200 - val_acc: 0.7158
Epoch 42/100
40000/40000 [==============================] - 3s - loss: 0.3810 - acc: 0.8664 - val_loss: 0.9044 - val_acc: 0.7088
Epoch 43/100
40000/40000 [==============================] - 3s - loss: 0.3751 - acc: 0.8682 - val_loss: 0.9827 - val_acc: 0.7167
Epoch 44/100
40000/40000 [==============================] - 3s - loss: 0.3673 - acc: 0.8730 - val_loss: 0.9563 - val_acc: 0.7126
Epoch 45/100
40000/40000 [==============================] - 3s - loss: 0.3662 - acc: 0.8728 - val_loss: 0.9247 - val_acc: 0.7176
Epoch 46/100
40000/40000 [==============================] - 3s - loss: 0.3606 - acc: 0.8740 - val_loss: 1.0758 - val_acc: 0.7174
Epoch 47/100
40000/40000 [==============================] - 3s - loss: 0.3560 - acc: 0.8756 - val_loss: 1.0478 - val_acc: 0.7110
Epoch 48/100
40000/40000 [==============================] - 3s - loss: 0.3550 - acc: 0.8774 - val_loss: 1.1299 - val_acc: 0.7174
Epoch 49/100
40000/40000 [==============================] - 3s - loss: 0.3533 - acc: 0.8772 - val_loss: 1.1774 - val_acc: 0.7192
Epoch 50/100
40000/40000 [==============================] - 3s - loss: 0.3468 - acc: 0.8808 - val_loss: 0.9025 - val_acc: 0.7034
Epoch 51/100
40000/40000 [==============================] - 3s - loss: 0.3526 - acc: 0.8777 - val_loss: 0.9376 - val_acc: 0.7181
Epoch 52/100
40000/40000 [==============================] - 3s - loss: 0.3388 - acc: 0.8852 - val_loss: 0.9904 - val_acc: 0.7210
Epoch 53/100
40000/40000 [==============================] - 3s - loss: 0.3443 - acc: 0.8816 - val_loss: 1.1140 - val_acc: 0.7137
Epoch 54/100
40000/40000 [==============================] - 3s - loss: 0.3454 - acc: 0.8838 - val_loss: 1.0140 - val_acc: 0.7210
Epoch 55/100
40000/40000 [==============================] - 3s - loss: 0.3415 - acc: 0.8817 - val_loss: 1.0218 - val_acc: 0.7195
Epoch 56/100
40000/40000 [==============================] - 3s - loss: 0.3360 - acc: 0.8845 - val_loss: 1.1122 - val_acc: 0.7136
Epoch 57/100
40000/40000 [==============================] - 3s - loss: 0.3350 - acc: 0.8858 - val_loss: 0.9474 - val_acc: 0.6842
Epoch 58/100
40000/40000 [==============================] - 3s - loss: 0.3364 - acc: 0.8864 - val_loss: 1.1625 - val_acc: 0.7200
Epoch 59/100
40000/40000 [==============================] - 3s - loss: 0.3357 - acc: 0.8859 - val_loss: 1.2038 - val_acc: 0.7205
Epoch 60/100
40000/40000 [==============================] - 3s - loss: 0.3390 - acc: 0.8864 - val_loss: 1.0123 - val_acc: 0.7150
Epoch 61/100
40000/40000 [==============================] - 3s - loss: 0.3396 - acc: 0.8865 - val_loss: 1.0195 - val_acc: 0.7189
Epoch 62/100
40000/40000 [==============================] - 3s - loss: 0.3393 - acc: 0.8881 - val_loss: 1.3219 - val_acc: 0.7190
Epoch 63/100
40000/40000 [==============================] - 3s - loss: 0.3360 - acc: 0.8866 - val_loss: 1.0100 - val_acc: 0.7229
Epoch 64/100
40000/40000 [==============================] - 3s - loss: 0.3474 - acc: 0.8846 - val_loss: 1.1325 - val_acc: 0.7183
Epoch 65/100
40000/40000 [==============================] - 3s - loss: 0.3250 - acc: 0.8917 - val_loss: 0.9521 - val_acc: 0.7147
Epoch 66/100
40000/40000 [==============================] - 3s - loss: 0.3388 - acc: 0.8887 - val_loss: 0.9979 - val_acc: 0.7183
Epoch 67/100
40000/40000 [==============================] - 3s - loss: 0.3429 - acc: 0.8869 - val_loss: 1.0285 - val_acc: 0.7179
Epoch 68/100
40000/40000 [==============================] - 3s - loss: 0.3359 - acc: 0.8896 - val_loss: 1.0319 - val_acc: 0.7184
Epoch 69/100
40000/40000 [==============================] - 3s - loss: 0.3463 - acc: 0.8880 - val_loss: 1.1072 - val_acc: 0.7182
Epoch 70/100
40000/40000 [==============================] - 3s - loss: 0.3421 - acc: 0.8886 - val_loss: 1.1975 - val_acc: 0.7189
Epoch 71/100
40000/40000 [==============================] - 3s - loss: 0.3342 - acc: 0.8914 - val_loss: 1.3263 - val_acc: 0.7214
Epoch 72/100
40000/40000 [==============================] - 3s - loss: 0.3394 - acc: 0.8892 - val_loss: 1.0824 - val_acc: 0.7164
Epoch 73/100
40000/40000 [==============================] - 3s - loss: 0.3395 - acc: 0.8888 - val_loss: 1.1018 - val_acc: 0.7195
Epoch 74/100
40000/40000 [==============================] - 3s - loss: 0.3404 - acc: 0.8886 - val_loss: 0.9730 - val_acc: 0.6867
Epoch 75/100
40000/40000 [==============================] - 3s - loss: 0.3360 - acc: 0.8914 - val_loss: 1.3034 - val_acc: 0.7246
Epoch 76/100
40000/40000 [==============================] - 3s - loss: 0.3444 - acc: 0.8888 - val_loss: 1.1297 - val_acc: 0.7186
Epoch 77/100
40000/40000 [==============================] - 3s - loss: 0.3472 - acc: 0.8885 - val_loss: 1.0597 - val_acc: 0.7165
Epoch 78/100
40000/40000 [==============================] - 3s - loss: 0.3490 - acc: 0.8847 - val_loss: 1.2927 - val_acc: 0.7218
Epoch 79/100
40000/40000 [==============================] - 3s - loss: 0.3503 - acc: 0.8862 - val_loss: 1.1642 - val_acc: 0.7210
Epoch 80/100
40000/40000 [==============================] - 3s - loss: 0.3499 - acc: 0.8878 - val_loss: 1.0784 - val_acc: 0.7181
Epoch 81/100
40000/40000 [==============================] - 3s - loss: 0.3459 - acc: 0.8883 - val_loss: 1.0262 - val_acc: 0.7232
Epoch 82/100
40000/40000 [==============================] - 3s - loss: 0.3534 - acc: 0.8868 - val_loss: 1.0196 - val_acc: 0.7157
Epoch 83/100
40000/40000 [==============================] - 3s - loss: 0.3583 - acc: 0.8867 - val_loss: 1.0146 - val_acc: 0.7145
Epoch 84/100
40000/40000 [==============================] - 3s - loss: 0.3435 - acc: 0.8897 - val_loss: 0.9619 - val_acc: 0.7183
Epoch 85/100
40000/40000 [==============================] - 3s - loss: 0.3522 - acc: 0.8889 - val_loss: 0.9880 - val_acc: 0.7123
Epoch 86/100
40000/40000 [==============================] - 3s - loss: 0.3522 - acc: 0.8857 - val_loss: 1.1175 - val_acc: 0.7219
Epoch 87/100
40000/40000 [==============================] - 3s - loss: 0.3608 - acc: 0.8869 - val_loss: 1.0715 - val_acc: 0.7239
Epoch 88/100
40000/40000 [==============================] - 3s - loss: 0.3597 - acc: 0.8842 - val_loss: 1.2776 - val_acc: 0.7181
Epoch 89/100
40000/40000 [==============================] - 3s - loss: 0.3559 - acc: 0.8872 - val_loss: 0.9562 - val_acc: 0.7105
Epoch 90/100
40000/40000 [==============================] - 3s - loss: 0.3629 - acc: 0.8832 - val_loss: 1.4673 - val_acc: 0.7145
Epoch 91/100
40000/40000 [==============================] - 3s - loss: 0.3582 - acc: 0.8883 - val_loss: 0.9302 - val_acc: 0.7034
Epoch 92/100
40000/40000 [==============================] - 3s - loss: 0.3584 - acc: 0.8877 - val_loss: 1.0776 - val_acc: 0.7191
Epoch 93/100
40000/40000 [==============================] - 3s - loss: 0.3539 - acc: 0.8863 - val_loss: 0.9306 - val_acc: 0.7121
Epoch 94/100
40000/40000 [==============================] - 3s - loss: 0.3515 - acc: 0.8882 - val_loss: 1.0448 - val_acc: 0.7198
Epoch 95/100
40000/40000 [==============================] - 3s - loss: 0.3614 - acc: 0.8871 - val_loss: 1.0237 - val_acc: 0.7196
Epoch 96/100
40000/40000 [==============================] - 3s - loss: 0.3633 - acc: 0.8857 - val_loss: 0.9401 - val_acc: 0.7081
Epoch 97/100
40000/40000 [==============================] - 3s - loss: 0.3525 - acc: 0.8883 - val_loss: 1.2315 - val_acc: 0.7196
Epoch 98/100
40000/40000 [==============================] - 3s - loss: 0.3507 - acc: 0.8884 - val_loss: 1.2278 - val_acc: 0.7210
Epoch 99/100
40000/40000 [==============================] - 3s - loss: 0.3589 - acc: 0.8857 - val_loss: 0.9363 - val_acc: 0.6944
Epoch 100/100
40000/40000 [==============================] - 3s - loss: 0.3570 - acc: 0.8859 - val_loss: 0.9947 - val_acc: 0.7160
Confusion matrix
[[744 27 43 20 17 4 12 4 89 44]
[ 15 810 5 12 4 1 13 3 16 107]
[ 58 6 640 72 65 50 62 39 20 4]
[ 26 10 79 591 47 128 81 28 18 22]
[ 31 2 73 72 605 28 56 81 13 6]
[ 10 4 65 228 43 564 32 46 10 14]
[ 8 6 53 45 30 11 788 2 5 5]
[ 10 3 39 55 72 44 7 746 3 20]
[ 61 45 18 16 3 1 4 4 807 30]
[ 30 75 12 18 4 3 7 10 16 865]]
Test score: 0.99474778223
Test accuracy: 0.716
Output results are saved in output_2c07614e-13c5-4299-9791-b489bcb58e24
Out[3]:
'output_2c07614e-13c5-4299-9791-b489bcb58e24'
In [ ]:
Content source: jskDr/keraspp
Similar notebooks: