Keras入门——（4）长短期记忆网络LSTM（一）

参考：

https://blog.csdn.net/zwqjoy/article/details/80493341

https://blog.csdn.net/u012735708/article/details/82769711

执行代码：

# Naive LSTM to learn three-char window to one-char mapping
import numpy
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
from keras.utils import np_utils
# fix random seed for reproducibility
numpy.random.seed(7)
# define the raw dataset
alphabet = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
# create mapping of characters to integers (0-25) and the reverse
char_to_int = dict((c, i) for i, c in enumerate(alphabet))
int_to_char = dict((i, c) for i, c in enumerate(alphabet))
# prepare the dataset of input to output pairs encoded as integers
seq_length = 3
dataX = []
dataY = []
for i in range(0, len(alphabet) - seq_length, 1):
    seq_in = alphabet[i:i + seq_length]
    seq_out = alphabet[i + seq_length]
    dataX.append([char_to_int[char] for char in seq_in])
    dataY.append(char_to_int[seq_out])
    print(seq_in, '->', seq_out)
# reshape X to be [samples, time steps, features]
X = numpy.reshape(dataX, (len(dataX), 1, seq_length))
# normalize
X = X / float(len(alphabet))
# one hot encode the output variable
y = np_utils.to_categorical(dataY)
# create and fit the model
model = Sequential()
model.add(LSTM(32, input_shape=(X.shape[1], X.shape[2])))
model.add(Dense(y.shape[1], activation='softmax'))
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
model.fit(X, y, epochs=500, batch_size=1, verbose=2)
# summarize performance of the model
scores = model.evaluate(X, y, verbose=0)
print("Model Accuracy: %.2f%%" % (scores[1]*100))
# demonstrate some model predictions
for pattern in dataX:
    x = numpy.reshape(pattern, (1, 1, len(pattern)))
    x = x / float(len(alphabet))
    prediction = model.predict(x, verbose=0)
    index = numpy.argmax(prediction)
    result = int_to_char[index]
    seq_in = [int_to_char[value] for value in pattern]
    print(seq_in, "->", result)

返回信息：

Using TensorFlow backend.
ABC -> D
BCD -> E
CDE -> F
DEF -> G
EFG -> H
FGH -> I
GHI -> J
HIJ -> K
IJK -> L
JKL -> M
KLM -> N
LMN -> O
MNO -> P
NOP -> Q
OPQ -> R
PQR -> S
QRS -> T
RST -> U
STU -> V
TUV -> W
UVW -> X
VWX -> Y
WXY -> Z
WARNING:tensorflow:From D:\ProgramData\Anaconda2\lib\site-packages\tensorflow\python\framework\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.
WARNING:tensorflow:From D:\ProgramData\Anaconda2\lib\site-packages\tensorflow\python\ops\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.
Epoch 1/500
 - 8s - loss: 3.2651 - acc: 0.0000e+00
Epoch 2/500
 - 0s - loss: 3.2527 - acc: 0.0435
Epoch 3/500
 - 0s - loss: 3.2462 - acc: 0.0435
Epoch 4/500
 - 0s - loss: 3.2402 - acc: 0.0000e+00
Epoch 5/500
 - 0s - loss: 3.2339 - acc: 0.0435
Epoch 6/500
 - 0s - loss: 3.2274 - acc: 0.0435
Epoch 7/500
 - 0s - loss: 3.2209 - acc: 0.0435
Epoch 8/500
 - 0s - loss: 3.2142 - acc: 0.0000e+00
Epoch 9/500
 - 0s - loss: 3.2067 - acc: 0.0435
Epoch 10/500
 - 0s - loss: 3.1993 - acc: 0.0435
Epoch 11/500
 - 0s - loss: 3.1918 - acc: 0.0435
Epoch 12/500
 - 0s - loss: 3.1839 - acc: 0.0000e+00
Epoch 13/500
 - 0s - loss: 3.1756 - acc: 0.0435
Epoch 14/500
 - 0s - loss: 3.1674 - acc: 0.0435
Epoch 15/500
 - 0s - loss: 3.1586 - acc: 0.0000e+00
Epoch 16/500
 - 0s - loss: 3.1498 - acc: 0.0435
Epoch 17/500
 - 0s - loss: 3.1418 - acc: 0.0000e+00
Epoch 18/500
 - 0s - loss: 3.1340 - acc: 0.0000e+00
Epoch 19/500
 - 0s - loss: 3.1245 - acc: 0.0435
Epoch 20/500
 - 0s - loss: 3.1167 - acc: 0.0435
Epoch 21/500
 - 0s - loss: 3.1096 - acc: 0.0435
Epoch 22/500
 - 0s - loss: 3.1018 - acc: 0.0435
Epoch 23/500
 - 0s - loss: 3.0935 - acc: 0.0435
Epoch 24/500
 - 0s - loss: 3.0857 - acc: 0.0435
Epoch 25/500
 - 0s - loss: 3.0788 - acc: 0.0435
Epoch 26/500
 - 0s - loss: 3.0721 - acc: 0.0435
Epoch 27/500
 - 0s - loss: 3.0647 - acc: 0.0435
Epoch 28/500
 - 0s - loss: 3.0584 - acc: 0.0435
Epoch 29/500
 - 0s - loss: 3.0530 - acc: 0.0435
Epoch 30/500
 - 0s - loss: 3.0449 - acc: 0.0435
Epoch 31/500
 - 0s - loss: 3.0398 - acc: 0.0435
Epoch 32/500
 - 0s - loss: 3.0328 - acc: 0.0870
Epoch 33/500
 - 0s - loss: 3.0257 - acc: 0.0870
Epoch 34/500
 - 0s - loss: 3.0200 - acc: 0.0870
Epoch 35/500
 - 0s - loss: 3.0132 - acc: 0.0870
Epoch 36/500
 - 0s - loss: 3.0077 - acc: 0.0870
Epoch 37/500
 - 0s - loss: 2.9992 - acc: 0.0870
Epoch 38/500
 - 0s - loss: 2.9946 - acc: 0.0870
Epoch 39/500
 - 0s - loss: 2.9855 - acc: 0.0870
Epoch 40/500
 - 0s - loss: 2.9790 - acc: 0.0870
Epoch 41/500
 - 0s - loss: 2.9725 - acc: 0.0870
Epoch 42/500
 - 0s - loss: 2.9655 - acc: 0.0870
Epoch 43/500
 - 0s - loss: 2.9576 - acc: 0.0870
Epoch 44/500
 - 0s - loss: 2.9501 - acc: 0.0870
Epoch 45/500
 - 0s - loss: 2.9420 - acc: 0.0870
Epoch 46/500
 - 0s - loss: 2.9353 - acc: 0.0870
Epoch 47/500
 - 0s - loss: 2.9271 - acc: 0.0870
Epoch 48/500
 - 0s - loss: 2.9193 - acc: 0.0870
Epoch 49/500
 - 0s - loss: 2.9104 - acc: 0.0870
Epoch 50/500
 - 0s - loss: 2.9012 - acc: 0.0870
Epoch 51/500
 - 0s - loss: 2.8931 - acc: 0.0870
Epoch 52/500
 - 0s - loss: 2.8841 - acc: 0.0870
Epoch 53/500
 - 0s - loss: 2.8759 - acc: 0.0870
Epoch 54/500
 - 0s - loss: 2.8653 - acc: 0.0870
Epoch 55/500
 - 0s - loss: 2.8574 - acc: 0.0870
Epoch 56/500
 - 0s - loss: 2.8467 - acc: 0.0870
Epoch 57/500
 - 0s - loss: 2.8372 - acc: 0.0870
Epoch 58/500
 - 0s - loss: 2.8272 - acc: 0.0870
Epoch 59/500
 - 0s - loss: 2.8180 - acc: 0.0870
Epoch 60/500
 - 0s - loss: 2.8074 - acc: 0.0870
Epoch 61/500
 - 0s - loss: 2.7979 - acc: 0.0870
Epoch 62/500
 - 0s - loss: 2.7865 - acc: 0.1304
Epoch 63/500
 - 0s - loss: 2.7778 - acc: 0.1304
Epoch 64/500
 - 0s - loss: 2.7675 - acc: 0.1304
Epoch 65/500
 - 0s - loss: 2.7577 - acc: 0.0870
Epoch 66/500
 - 0s - loss: 2.7471 - acc: 0.0870
Epoch 67/500
 - 0s - loss: 2.7384 - acc: 0.0870
Epoch 68/500
 - 0s - loss: 2.7288 - acc: 0.0870
Epoch 69/500
 - 0s - loss: 2.7165 - acc: 0.0870
Epoch 70/500
 - 0s - loss: 2.7084 - acc: 0.0870
Epoch 71/500
 - 0s - loss: 2.6975 - acc: 0.0870
Epoch 72/500
 - 0s - loss: 2.6891 - acc: 0.0870
Epoch 73/500
 - 0s - loss: 2.6801 - acc: 0.0870
Epoch 74/500
 - 0s - loss: 2.6708 - acc: 0.0870
Epoch 75/500
 - 0s - loss: 2.6624 - acc: 0.0870
Epoch 76/500
 - 0s - loss: 2.6537 - acc: 0.0870
Epoch 77/500
 - 0s - loss: 2.6471 - acc: 0.0870
Epoch 78/500
 - 0s - loss: 2.6378 - acc: 0.1304
Epoch 79/500
 - 0s - loss: 2.6304 - acc: 0.1304
Epoch 80/500
 - 0s - loss: 2.6220 - acc: 0.1304
Epoch 81/500
 - 0s - loss: 2.6150 - acc: 0.1304
Epoch 82/500
 - 0s - loss: 2.6070 - acc: 0.1304
Epoch 83/500
 - 0s - loss: 2.6006 - acc: 0.1304
Epoch 84/500
 - 0s - loss: 2.5950 - acc: 0.1304
Epoch 85/500
 - 0s - loss: 2.5855 - acc: 0.0870
Epoch 86/500
 - 0s - loss: 2.5784 - acc: 0.0870
Epoch 87/500
 - 0s - loss: 2.5741 - acc: 0.0870
Epoch 88/500
 - 0s - loss: 2.5655 - acc: 0.1304
Epoch 89/500
 - 0s - loss: 2.5596 - acc: 0.0870
Epoch 90/500
 - 0s - loss: 2.5528 - acc: 0.0870
Epoch 91/500
 - 0s - loss: 2.5470 - acc: 0.1304
Epoch 92/500
 - 0s - loss: 2.5402 - acc: 0.1304
Epoch 93/500
 - 0s - loss: 2.5350 - acc: 0.1304
Epoch 94/500
 - 0s - loss: 2.5291 - acc: 0.1304
Epoch 95/500
 - 0s - loss: 2.5234 - acc: 0.1304
Epoch 96/500
 - 0s - loss: 2.5174 - acc: 0.1304
Epoch 97/500
 - 0s - loss: 2.5107 - acc: 0.1304
Epoch 98/500
 - 0s - loss: 2.5043 - acc: 0.1304
Epoch 99/500
 - 0s - loss: 2.4984 - acc: 0.1304
Epoch 100/500
 - 0s - loss: 2.4939 - acc: 0.1304
Epoch 101/500
 - 0s - loss: 2.4886 - acc: 0.1304
Epoch 102/500
 - 0s - loss: 2.4820 - acc: 0.1304
Epoch 103/500
 - 0s - loss: 2.4761 - acc: 0.1739
Epoch 104/500
 - 0s - loss: 2.4696 - acc: 0.1739
Epoch 105/500
 - 0s - loss: 2.4660 - acc: 0.1304
Epoch 106/500
 - 0s - loss: 2.4610 - acc: 0.1304
Epoch 107/500
 - 0s - loss: 2.4551 - acc: 0.1304
Epoch 108/500
 - 0s - loss: 2.4498 - acc: 0.1304
Epoch 109/500
 - 0s - loss: 2.4431 - acc: 0.1304
Epoch 110/500
 - 0s - loss: 2.4387 - acc: 0.1739
Epoch 111/500
 - 0s - loss: 2.4333 - acc: 0.1304
Epoch 112/500
 - 0s - loss: 2.4270 - acc: 0.1304
Epoch 113/500
 - 0s - loss: 2.4243 - acc: 0.1739
Epoch 114/500
 - 0s - loss: 2.4161 - acc: 0.1304
Epoch 115/500
 - 0s - loss: 2.4130 - acc: 0.1304
Epoch 116/500
 - 0s - loss: 2.4078 - acc: 0.1739
Epoch 117/500
 - 0s - loss: 2.4023 - acc: 0.1739
Epoch 118/500
 - 0s - loss: 2.3974 - acc: 0.1304
Epoch 119/500
 - 0s - loss: 2.3921 - acc: 0.2174
Epoch 120/500
 - 0s - loss: 2.3869 - acc: 0.1304
Epoch 121/500
 - 0s - loss: 2.3831 - acc: 0.1304
Epoch 122/500
 - 0s - loss: 2.3777 - acc: 0.1739
Epoch 123/500
 - 0s - loss: 2.3728 - acc: 0.2174
Epoch 124/500
 - 0s - loss: 2.3682 - acc: 0.1739
Epoch 125/500
 - 0s - loss: 2.3634 - acc: 0.1739
Epoch 126/500
 - 0s - loss: 2.3586 - acc: 0.1739
Epoch 127/500
 - 0s - loss: 2.3532 - acc: 0.1739
Epoch 128/500
 - 0s - loss: 2.3482 - acc: 0.1739
Epoch 129/500
 - 0s - loss: 2.3463 - acc: 0.2174
Epoch 130/500
 - 0s - loss: 2.3414 - acc: 0.2174
Epoch 131/500
 - 0s - loss: 2.3363 - acc: 0.2174
Epoch 132/500
 - 0s - loss: 2.3322 - acc: 0.1739
Epoch 133/500
 - 0s - loss: 2.3270 - acc: 0.2174
Epoch 134/500
 - 0s - loss: 2.3238 - acc: 0.2174
Epoch 135/500
 - 0s - loss: 2.3194 - acc: 0.2174
Epoch 136/500
 - 0s - loss: 2.3152 - acc: 0.2174
Epoch 137/500
 - 0s - loss: 2.3090 - acc: 0.2174
Epoch 138/500
 - 0s - loss: 2.3051 - acc: 0.2174
Epoch 139/500
 - 0s - loss: 2.3028 - acc: 0.2174
Epoch 140/500
 - 0s - loss: 2.2952 - acc: 0.2174
Epoch 141/500
 - 0s - loss: 2.2936 - acc: 0.2174
Epoch 142/500
 - 0s - loss: 2.2890 - acc: 0.1739
Epoch 143/500
 - 0s - loss: 2.2830 - acc: 0.1739
Epoch 144/500
 - 0s - loss: 2.2797 - acc: 0.2174
Epoch 145/500
 - 0s - loss: 2.2757 - acc: 0.2174
Epoch 146/500
 - 0s - loss: 2.2710 - acc: 0.2174
Epoch 147/500
 - 0s - loss: 2.2676 - acc: 0.2174
Epoch 148/500
 - 0s - loss: 2.2635 - acc: 0.1739
Epoch 149/500
 - 0s - loss: 2.2603 - acc: 0.2174
Epoch 150/500
 - 0s - loss: 2.2570 - acc: 0.2174
Epoch 151/500
 - 0s - loss: 2.2524 - acc: 0.2174
Epoch 152/500
 - 0s - loss: 2.2483 - acc: 0.1739
Epoch 153/500
 - 0s - loss: 2.2437 - acc: 0.2174
Epoch 154/500
 - 0s - loss: 2.2409 - acc: 0.2174
Epoch 155/500
 - 0s - loss: 2.2361 - acc: 0.1739
Epoch 156/500
 - 0s - loss: 2.2345 - acc: 0.2174
Epoch 157/500
 - 0s - loss: 2.2296 - acc: 0.2174
Epoch 158/500
 - 0s - loss: 2.2252 - acc: 0.2174
Epoch 159/500
 - 0s - loss: 2.2219 - acc: 0.2174
Epoch 160/500
 - 0s - loss: 2.2190 - acc: 0.2174
Epoch 161/500
 - 0s - loss: 2.2161 - acc: 0.2609
Epoch 162/500
 - 0s - loss: 2.2119 - acc: 0.2609
Epoch 163/500
 - 0s - loss: 2.2065 - acc: 0.2609
Epoch 164/500
 - 0s - loss: 2.2046 - acc: 0.2609
Epoch 165/500
 - 0s - loss: 2.2011 - acc: 0.2609
Epoch 166/500
 - 0s - loss: 2.1987 - acc: 0.3043
Epoch 167/500
 - 0s - loss: 2.1948 - acc: 0.2174
Epoch 168/500
 - 0s - loss: 2.1914 - acc: 0.3043
Epoch 169/500
 - 0s - loss: 2.1882 - acc: 0.2609
Epoch 170/500
 - 0s - loss: 2.1863 - acc: 0.2609
Epoch 171/500
 - 0s - loss: 2.1808 - acc: 0.2174
Epoch 172/500
 - 0s - loss: 2.1779 - acc: 0.3478
Epoch 173/500
 - 0s - loss: 2.1744 - acc: 0.3478
Epoch 174/500
 - 0s - loss: 2.1736 - acc: 0.3478
Epoch 175/500
 - 0s - loss: 2.1686 - acc: 0.3478
Epoch 176/500
 - 0s - loss: 2.1652 - acc: 0.3043
Epoch 177/500
 - 0s - loss: 2.1617 - acc: 0.2609
Epoch 178/500
 - 0s - loss: 2.1613 - acc: 0.2609
Epoch 179/500
 - 0s - loss: 2.1553 - acc: 0.3478
Epoch 180/500
 - 0s - loss: 2.1534 - acc: 0.2609
Epoch 181/500
 - 0s - loss: 2.1511 - acc: 0.2609
Epoch 182/500
 - 0s - loss: 2.1477 - acc: 0.3043
Epoch 183/500
 - 0s - loss: 2.1445 - acc: 0.2609
Epoch 184/500
 - 0s - loss: 2.1416 - acc: 0.3913
Epoch 185/500
 - 0s - loss: 2.1383 - acc: 0.3478
Epoch 186/500
 - 0s - loss: 2.1366 - acc: 0.3478
Epoch 187/500
 - 0s - loss: 2.1328 - acc: 0.3043
Epoch 188/500
 - 0s - loss: 2.1317 - acc: 0.3043
Epoch 189/500
 - 0s - loss: 2.1284 - acc: 0.3478
Epoch 190/500
 - 0s - loss: 2.1242 - acc: 0.3478
Epoch 191/500
 - 0s - loss: 2.1225 - acc: 0.3043
Epoch 192/500
 - 0s - loss: 2.1178 - acc: 0.3043
Epoch 193/500
 - 0s - loss: 2.1171 - acc: 0.2609
Epoch 194/500
 - 0s - loss: 2.1141 - acc: 0.2609
Epoch 195/500
 - 0s - loss: 2.1108 - acc: 0.3043
Epoch 196/500
 - 0s - loss: 2.1100 - acc: 0.3478
Epoch 197/500
 - 0s - loss: 2.1051 - acc: 0.3043
Epoch 198/500
 - 0s - loss: 2.1025 - acc: 0.3478
Epoch 199/500
 - 0s - loss: 2.1005 - acc: 0.3478
Epoch 200/500
 - 0s - loss: 2.0982 - acc: 0.3478
Epoch 201/500
 - 0s - loss: 2.0951 - acc: 0.3478
Epoch 202/500
 - 0s - loss: 2.0926 - acc: 0.3043
Epoch 203/500
 - 0s - loss: 2.0919 - acc: 0.3043
Epoch 204/500
 - 0s - loss: 2.0876 - acc: 0.3478
Epoch 205/500
 - 0s - loss: 2.0844 - acc: 0.3043
Epoch 206/500
 - 0s - loss: 2.0838 - acc: 0.3043
Epoch 207/500
 - 0s - loss: 2.0798 - acc: 0.3043
Epoch 208/500
 - 0s - loss: 2.0777 - acc: 0.3478
Epoch 209/500
 - 0s - loss: 2.0767 - acc: 0.3043
Epoch 210/500
 - 0s - loss: 2.0723 - acc: 0.2609
Epoch 211/500
 - 0s - loss: 2.0716 - acc: 0.3043
Epoch 212/500
 - 0s - loss: 2.0690 - acc: 0.3043
Epoch 213/500
 - 0s - loss: 2.0663 - acc: 0.3478
Epoch 214/500
 - 0s - loss: 2.0632 - acc: 0.3913
Epoch 215/500
 - 0s - loss: 2.0628 - acc: 0.3478
Epoch 216/500
 - 0s - loss: 2.0603 - acc: 0.3478
Epoch 217/500
 - 0s - loss: 2.0567 - acc: 0.3913
Epoch 218/500
 - 0s - loss: 2.0559 - acc: 0.3913
Epoch 219/500
 - 0s - loss: 2.0509 - acc: 0.3913
Epoch 220/500
 - 0s - loss: 2.0499 - acc: 0.3043
Epoch 221/500
 - 0s - loss: 2.0482 - acc: 0.3478
Epoch 222/500
 - 0s - loss: 2.0439 - acc: 0.3478
Epoch 223/500
 - 0s - loss: 2.0427 - acc: 0.3913
Epoch 224/500
 - 0s - loss: 2.0404 - acc: 0.4348
Epoch 225/500
 - 0s - loss: 2.0393 - acc: 0.3913
Epoch 226/500
 - 0s - loss: 2.0379 - acc: 0.4348
Epoch 227/500
 - 0s - loss: 2.0360 - acc: 0.4348
Epoch 228/500
 - 0s - loss: 2.0330 - acc: 0.4348
Epoch 229/500
 - 0s - loss: 2.0307 - acc: 0.4348
Epoch 230/500
 - 0s - loss: 2.0269 - acc: 0.4783
Epoch 231/500
 - 0s - loss: 2.0251 - acc: 0.3913
Epoch 232/500
 - 0s - loss: 2.0234 - acc: 0.4783
Epoch 233/500
 - 0s - loss: 2.0222 - acc: 0.4348
Epoch 234/500
 - 0s - loss: 2.0190 - acc: 0.4783
Epoch 235/500
 - 0s - loss: 2.0175 - acc: 0.5652
Epoch 236/500
 - 0s - loss: 2.0161 - acc: 0.4783
Epoch 237/500
 - 0s - loss: 2.0133 - acc: 0.4348
Epoch 238/500
 - 0s - loss: 2.0097 - acc: 0.4348
Epoch 239/500
 - 0s - loss: 2.0094 - acc: 0.3913
Epoch 240/500
 - 0s - loss: 2.0077 - acc: 0.4783
Epoch 241/500
 - 0s - loss: 2.0048 - acc: 0.4348
Epoch 242/500
 - 0s - loss: 2.0028 - acc: 0.4348
Epoch 243/500
 - 0s - loss: 2.0002 - acc: 0.4348
Epoch 244/500
 - 0s - loss: 1.9974 - acc: 0.4348
Epoch 245/500
 - 0s - loss: 1.9958 - acc: 0.4783
Epoch 246/500
 - 0s - loss: 1.9956 - acc: 0.4348
Epoch 247/500
 - 0s - loss: 1.9929 - acc: 0.4783
Epoch 248/500
 - 0s - loss: 1.9916 - acc: 0.4783
Epoch 249/500
 - 0s - loss: 1.9888 - acc: 0.5652
Epoch 250/500
 - 0s - loss: 1.9895 - acc: 0.5217
Epoch 251/500
 - 0s - loss: 1.9838 - acc: 0.4348
Epoch 252/500
 - 0s - loss: 1.9840 - acc: 0.4348
Epoch 253/500
 - 0s - loss: 1.9814 - acc: 0.5652
Epoch 254/500
 - 0s - loss: 1.9812 - acc: 0.4783
Epoch 255/500
 - 0s - loss: 1.9768 - acc: 0.5217
Epoch 256/500
 - 0s - loss: 1.9759 - acc: 0.4348
Epoch 257/500
 - 0s - loss: 1.9741 - acc: 0.4783
Epoch 258/500
 - 0s - loss: 1.9703 - acc: 0.5652
Epoch 259/500
 - 0s - loss: 1.9713 - acc: 0.4348
Epoch 260/500
 - 0s - loss: 1.9653 - acc: 0.5217
Epoch 261/500
 - 0s - loss: 1.9658 - acc: 0.5217
Epoch 262/500
 - 0s - loss: 1.9624 - acc: 0.5652
Epoch 263/500
 - 0s - loss: 1.9614 - acc: 0.5217
Epoch 264/500
 - 0s - loss: 1.9632 - acc: 0.5217
Epoch 265/500
 - 0s - loss: 1.9588 - acc: 0.5217
Epoch 266/500
 - 0s - loss: 1.9556 - acc: 0.5217
Epoch 267/500
 - 0s - loss: 1.9556 - acc: 0.5217
Epoch 268/500
 - 0s - loss: 1.9511 - acc: 0.5217
Epoch 269/500
 - 0s - loss: 1.9522 - acc: 0.5652
Epoch 270/500
 - 0s - loss: 1.9502 - acc: 0.5652
Epoch 271/500
 - 0s - loss: 1.9494 - acc: 0.5652
Epoch 272/500
 - 0s - loss: 1.9450 - acc: 0.5652
Epoch 273/500
 - 0s - loss: 1.9455 - acc: 0.5217
Epoch 274/500
 - 0s - loss: 1.9446 - acc: 0.3913
Epoch 275/500
 - 0s - loss: 1.9406 - acc: 0.4783
Epoch 276/500
 - 0s - loss: 1.9392 - acc: 0.4783
Epoch 277/500
 - 0s - loss: 1.9353 - acc: 0.5652
Epoch 278/500
 - 0s - loss: 1.9356 - acc: 0.4348
Epoch 279/500
 - 0s - loss: 1.9355 - acc: 0.6087
Epoch 280/500
 - 0s - loss: 1.9345 - acc: 0.5652
Epoch 281/500
 - 0s - loss: 1.9291 - acc: 0.6087
Epoch 282/500
 - 0s - loss: 1.9311 - acc: 0.6087
Epoch 283/500
 - 0s - loss: 1.9298 - acc: 0.4783
Epoch 284/500
 - 0s - loss: 1.9264 - acc: 0.5217
Epoch 285/500
 - 0s - loss: 1.9245 - acc: 0.6087
Epoch 286/500
 - 0s - loss: 1.9233 - acc: 0.5652
Epoch 287/500
 - 0s - loss: 1.9217 - acc: 0.4783
Epoch 288/500
 - 0s - loss: 1.9193 - acc: 0.5217
Epoch 289/500
 - 0s - loss: 1.9149 - acc: 0.5217
Epoch 290/500
 - 0s - loss: 1.9153 - acc: 0.5217
Epoch 291/500
 - 0s - loss: 1.9128 - acc: 0.6087
Epoch 292/500
 - 0s - loss: 1.9112 - acc: 0.6957
Epoch 293/500
 - 0s - loss: 1.9112 - acc: 0.6087
Epoch 294/500
 - 0s - loss: 1.9095 - acc: 0.6087
Epoch 295/500
 - 0s - loss: 1.9077 - acc: 0.5652
Epoch 296/500
 - 0s - loss: 1.9059 - acc: 0.6087
Epoch 297/500
 - 0s - loss: 1.9054 - acc: 0.6522
Epoch 298/500
 - 0s - loss: 1.9045 - acc: 0.6087
Epoch 299/500
 - 0s - loss: 1.9010 - acc: 0.6522
Epoch 300/500
 - 0s - loss: 1.8994 - acc: 0.5217
Epoch 301/500
 - 0s - loss: 1.8975 - acc: 0.4348
Epoch 302/500
 - 0s - loss: 1.8957 - acc: 0.5652
Epoch 303/500
 - 0s - loss: 1.8956 - acc: 0.6087
Epoch 304/500
 - 0s - loss: 1.8962 - acc: 0.4783
Epoch 305/500
 - 0s - loss: 1.8935 - acc: 0.5217
Epoch 306/500
 - 0s - loss: 1.8892 - acc: 0.5652
Epoch 307/500
 - 0s - loss: 1.8881 - acc: 0.6087
Epoch 308/500
 - 0s - loss: 1.8867 - acc: 0.5652
Epoch 309/500
 - 0s - loss: 1.8869 - acc: 0.5652
Epoch 310/500
 - 0s - loss: 1.8837 - acc: 0.6087
Epoch 311/500
 - 0s - loss: 1.8825 - acc: 0.6522
Epoch 312/500
 - 0s - loss: 1.8791 - acc: 0.5217
Epoch 313/500
 - 0s - loss: 1.8790 - acc: 0.6087
Epoch 314/500
 - 0s - loss: 1.8771 - acc: 0.6087
Epoch 315/500
 - 0s - loss: 1.8766 - acc: 0.6087
Epoch 316/500
 - 0s - loss: 1.8746 - acc: 0.5652
Epoch 317/500
 - 0s - loss: 1.8720 - acc: 0.5652
Epoch 318/500
 - 0s - loss: 1.8711 - acc: 0.6087
Epoch 319/500
 - 0s - loss: 1.8699 - acc: 0.5652
Epoch 320/500
 - 0s - loss: 1.8688 - acc: 0.4783
Epoch 321/500
 - 0s - loss: 1.8674 - acc: 0.5652
Epoch 322/500
 - 0s - loss: 1.8677 - acc: 0.5652
Epoch 323/500
 - 0s - loss: 1.8627 - acc: 0.5217
Epoch 324/500
 - 0s - loss: 1.8636 - acc: 0.6087
Epoch 325/500
 - 0s - loss: 1.8623 - acc: 0.6522
Epoch 326/500
 - 0s - loss: 1.8608 - acc: 0.5217
Epoch 327/500
 - 0s - loss: 1.8619 - acc: 0.6522
Epoch 328/500
 - 0s - loss: 1.8582 - acc: 0.6087
Epoch 329/500
 - 0s - loss: 1.8554 - acc: 0.5652
Epoch 330/500
 - 0s - loss: 1.8540 - acc: 0.6522
Epoch 331/500
 - 0s - loss: 1.8567 - acc: 0.5652
Epoch 332/500
 - 0s - loss: 1.8520 - acc: 0.5652
Epoch 333/500
 - 0s - loss: 1.8515 - acc: 0.6522
Epoch 334/500
 - 0s - loss: 1.8484 - acc: 0.6087
Epoch 335/500
 - 0s - loss: 1.8498 - acc: 0.6087
Epoch 336/500
 - 0s - loss: 1.8451 - acc: 0.6522
Epoch 337/500
 - 0s - loss: 1.8434 - acc: 0.6522
Epoch 338/500
 - 0s - loss: 1.8431 - acc: 0.5217
Epoch 339/500
 - 0s - loss: 1.8418 - acc: 0.6087
Epoch 340/500
 - 0s - loss: 1.8410 - acc: 0.5217
Epoch 341/500
 - 0s - loss: 1.8395 - acc: 0.6522
Epoch 342/500
 - 0s - loss: 1.8392 - acc: 0.6087
Epoch 343/500
 - 0s - loss: 1.8362 - acc: 0.5652
Epoch 344/500
 - 0s - loss: 1.8336 - acc: 0.6087
Epoch 345/500
 - 0s - loss: 1.8320 - acc: 0.6087
Epoch 346/500
 - 0s - loss: 1.8316 - acc: 0.6522
Epoch 347/500
 - 0s - loss: 1.8325 - acc: 0.5652
Epoch 348/500
 - 0s - loss: 1.8284 - acc: 0.5652
Epoch 349/500
 - 0s - loss: 1.8278 - acc: 0.6087
Epoch 350/500
 - 0s - loss: 1.8263 - acc: 0.6087
Epoch 351/500
 - 0s - loss: 1.8234 - acc: 0.5217
Epoch 352/500
 - 0s - loss: 1.8244 - acc: 0.6087
Epoch 353/500
 - 0s - loss: 1.8224 - acc: 0.6522
Epoch 354/500
 - 0s - loss: 1.8208 - acc: 0.6522
Epoch 355/500
 - 0s - loss: 1.8225 - acc: 0.6522
Epoch 356/500
 - 0s - loss: 1.8181 - acc: 0.6522
Epoch 357/500
 - 0s - loss: 1.8170 - acc: 0.5217
Epoch 358/500
 - 0s - loss: 1.8182 - acc: 0.6522
Epoch 359/500
 - 0s - loss: 1.8146 - acc: 0.5652
Epoch 360/500
 - 0s - loss: 1.8114 - acc: 0.6957
Epoch 361/500
 - 0s - loss: 1.8111 - acc: 0.7391
Epoch 362/500
 - 0s - loss: 1.8091 - acc: 0.6522
Epoch 363/500
 - 0s - loss: 1.8096 - acc: 0.5652
Epoch 364/500
 - 0s - loss: 1.8078 - acc: 0.6087
Epoch 365/500
 - 0s - loss: 1.8069 - acc: 0.5652
Epoch 366/500
 - 0s - loss: 1.8060 - acc: 0.6522
Epoch 367/500
 - 0s - loss: 1.8041 - acc: 0.6087
Epoch 368/500
 - 0s - loss: 1.8021 - acc: 0.6957
Epoch 369/500
 - 0s - loss: 1.8003 - acc: 0.6957
Epoch 370/500
 - 0s - loss: 1.8004 - acc: 0.6957
Epoch 371/500
 - 0s - loss: 1.7980 - acc: 0.5652
Epoch 372/500
 - 0s - loss: 1.7977 - acc: 0.6522
Epoch 373/500
 - 0s - loss: 1.7946 - acc: 0.6957
Epoch 374/500
 - 0s - loss: 1.7930 - acc: 0.6957
Epoch 375/500
 - 0s - loss: 1.7939 - acc: 0.6957
Epoch 376/500
 - 0s - loss: 1.7907 - acc: 0.6087
Epoch 377/500
 - 0s - loss: 1.7892 - acc: 0.6522
Epoch 378/500
 - 0s - loss: 1.7899 - acc: 0.6087
Epoch 379/500
 - 0s - loss: 1.7861 - acc: 0.6522
Epoch 380/500
 - 0s - loss: 1.7871 - acc: 0.6522
Epoch 381/500
 - 0s - loss: 1.7870 - acc: 0.6087
Epoch 382/500
 - 0s - loss: 1.7850 - acc: 0.7391
Epoch 383/500
 - 0s - loss: 1.7811 - acc: 0.6957
Epoch 384/500
 - 0s - loss: 1.7812 - acc: 0.6522
Epoch 385/500
 - 0s - loss: 1.7824 - acc: 0.7391
Epoch 386/500
 - 0s - loss: 1.7790 - acc: 0.6522
Epoch 387/500
 - 0s - loss: 1.7762 - acc: 0.6957
Epoch 388/500
 - 0s - loss: 1.7761 - acc: 0.7826
Epoch 389/500
 - 0s - loss: 1.7763 - acc: 0.6957
Epoch 390/500
 - 0s - loss: 1.7740 - acc: 0.6957
Epoch 391/500
 - 0s - loss: 1.7719 - acc: 0.6957
Epoch 392/500
 - 0s - loss: 1.7698 - acc: 0.6957
Epoch 393/500
 - 0s - loss: 1.7712 - acc: 0.6522
Epoch 394/500
 - 0s - loss: 1.7673 - acc: 0.6522
Epoch 395/500
 - 0s - loss: 1.7690 - acc: 0.6957
Epoch 396/500
 - 0s - loss: 1.7659 - acc: 0.6522
Epoch 397/500
 - 0s - loss: 1.7666 - acc: 0.6087
Epoch 398/500
 - 0s - loss: 1.7657 - acc: 0.6087
Epoch 399/500
 - 0s - loss: 1.7630 - acc: 0.6957
Epoch 400/500
 - 0s - loss: 1.7623 - acc: 0.6522
Epoch 401/500
 - 0s - loss: 1.7604 - acc: 0.6957
Epoch 402/500
 - 0s - loss: 1.7576 - acc: 0.7391
Epoch 403/500
 - 0s - loss: 1.7580 - acc: 0.6522
Epoch 404/500
 - 0s - loss: 1.7584 - acc: 0.6957
Epoch 405/500
 - 0s - loss: 1.7561 - acc: 0.6522
Epoch 406/500
 - 0s - loss: 1.7555 - acc: 0.6522
Epoch 407/500
 - 0s - loss: 1.7526 - acc: 0.8261
Epoch 408/500
 - 0s - loss: 1.7531 - acc: 0.6957
Epoch 409/500
 - 0s - loss: 1.7507 - acc: 0.6957
Epoch 410/500
 - 0s - loss: 1.7508 - acc: 0.7391
Epoch 411/500
 - 0s - loss: 1.7495 - acc: 0.6957
Epoch 412/500
 - 0s - loss: 1.7495 - acc: 0.7391
Epoch 413/500
 - 0s - loss: 1.7469 - acc: 0.6957
Epoch 414/500
 - 0s - loss: 1.7459 - acc: 0.6522
Epoch 415/500
 - 0s - loss: 1.7434 - acc: 0.6957
Epoch 416/500
 - 0s - loss: 1.7414 - acc: 0.6522
Epoch 417/500
 - 0s - loss: 1.7393 - acc: 0.6957
Epoch 418/500
 - 0s - loss: 1.7383 - acc: 0.6522
Epoch 419/500
 - 0s - loss: 1.7388 - acc: 0.6957
Epoch 420/500
 - 0s - loss: 1.7389 - acc: 0.6087
Epoch 421/500
 - 0s - loss: 1.7379 - acc: 0.6957
Epoch 422/500
 - 0s - loss: 1.7335 - acc: 0.6957
Epoch 423/500
 - 0s - loss: 1.7331 - acc: 0.7391
Epoch 424/500
 - 0s - loss: 1.7339 - acc: 0.6957
Epoch 425/500
 - 0s - loss: 1.7338 - acc: 0.7391
Epoch 426/500
 - 0s - loss: 1.7303 - acc: 0.6957
Epoch 427/500
 - 0s - loss: 1.7278 - acc: 0.7826
Epoch 428/500
 - 0s - loss: 1.7274 - acc: 0.6522
Epoch 429/500
 - 0s - loss: 1.7277 - acc: 0.7391
Epoch 430/500
 - 0s - loss: 1.7264 - acc: 0.6957
Epoch 431/500
 - 0s - loss: 1.7249 - acc: 0.6522
Epoch 432/500
 - 0s - loss: 1.7245 - acc: 0.6522
Epoch 433/500
 - 0s - loss: 1.7202 - acc: 0.7391
Epoch 434/500
 - 0s - loss: 1.7201 - acc: 0.6522
Epoch 435/500
 - 0s - loss: 1.7186 - acc: 0.7391
Epoch 436/500
 - 0s - loss: 1.7177 - acc: 0.8261
Epoch 437/500
 - 0s - loss: 1.7187 - acc: 0.7391
Epoch 438/500
 - 0s - loss: 1.7170 - acc: 0.7391
Epoch 439/500
 - 0s - loss: 1.7148 - acc: 0.7391
Epoch 440/500
 - 0s - loss: 1.7130 - acc: 0.6957
Epoch 441/500
 - 0s - loss: 1.7140 - acc: 0.8261
Epoch 442/500
 - 0s - loss: 1.7124 - acc: 0.7826
Epoch 443/500
 - 0s - loss: 1.7077 - acc: 0.7826
Epoch 444/500
 - 0s - loss: 1.7108 - acc: 0.6957
Epoch 445/500
 - 0s - loss: 1.7080 - acc: 0.7391
Epoch 446/500
 - 0s - loss: 1.7068 - acc: 0.7391
Epoch 447/500
 - 0s - loss: 1.7061 - acc: 0.6522
Epoch 448/500
 - 0s - loss: 1.7056 - acc: 0.6957
Epoch 449/500
 - 0s - loss: 1.7052 - acc: 0.6957
Epoch 450/500
 - 0s - loss: 1.7015 - acc: 0.7391
Epoch 451/500
 - 0s - loss: 1.7008 - acc: 0.7391
Epoch 452/500
 - 0s - loss: 1.6998 - acc: 0.6957
Epoch 453/500
 - 0s - loss: 1.7005 - acc: 0.7391
Epoch 454/500
 - 0s - loss: 1.6990 - acc: 0.7826
Epoch 455/500
 - 0s - loss: 1.6948 - acc: 0.6957
Epoch 456/500
 - 0s - loss: 1.6984 - acc: 0.8261
Epoch 457/500
 - 0s - loss: 1.6917 - acc: 0.7826
Epoch 458/500
 - 0s - loss: 1.6947 - acc: 0.6087
Epoch 459/500
 - 0s - loss: 1.6923 - acc: 0.7826
Epoch 460/500
 - 0s - loss: 1.6934 - acc: 0.7391
Epoch 461/500
 - 0s - loss: 1.6918 - acc: 0.7391
Epoch 462/500
 - 0s - loss: 1.6893 - acc: 0.7391
Epoch 463/500
 - 0s - loss: 1.6865 - acc: 0.6957
Epoch 464/500
 - 0s - loss: 1.6843 - acc: 0.6957
Epoch 465/500
 - 0s - loss: 1.6856 - acc: 0.7391
Epoch 466/500
 - 0s - loss: 1.6861 - acc: 0.7391
Epoch 467/500
 - 0s - loss: 1.6828 - acc: 0.7826
Epoch 468/500
 - 0s - loss: 1.6819 - acc: 0.7826
Epoch 469/500
 - 0s - loss: 1.6800 - acc: 0.8261
Epoch 470/500
 - 0s - loss: 1.6785 - acc: 0.7826
Epoch 471/500
 - 0s - loss: 1.6795 - acc: 0.8261
Epoch 472/500
 - 0s - loss: 1.6761 - acc: 0.7391
Epoch 473/500
 - 0s - loss: 1.6770 - acc: 0.8261
Epoch 474/500
 - 0s - loss: 1.6755 - acc: 0.8261
Epoch 475/500
 - 0s - loss: 1.6722 - acc: 0.7826
Epoch 476/500
 - 0s - loss: 1.6703 - acc: 0.7826
Epoch 477/500
 - 0s - loss: 1.6705 - acc: 0.7391
Epoch 478/500
 - 0s - loss: 1.6700 - acc: 0.7826
Epoch 479/500
 - 0s - loss: 1.6676 - acc: 0.8696
Epoch 480/500
 - 0s - loss: 1.6700 - acc: 0.7826
Epoch 481/500
 - 0s - loss: 1.6695 - acc: 0.7826
Epoch 482/500
 - 0s - loss: 1.6668 - acc: 0.6957
Epoch 483/500
 - 0s - loss: 1.6669 - acc: 0.7391
Epoch 484/500
 - 0s - loss: 1.6657 - acc: 0.6957
Epoch 485/500
 - 0s - loss: 1.6640 - acc: 0.7391
Epoch 486/500
 - 0s - loss: 1.6613 - acc: 0.7391
Epoch 487/500
 - 0s - loss: 1.6623 - acc: 0.7826
Epoch 488/500
 - 0s - loss: 1.6612 - acc: 0.6957
Epoch 489/500
 - 0s - loss: 1.6574 - acc: 0.7391
Epoch 490/500
 - 0s - loss: 1.6580 - acc: 0.7826
Epoch 491/500
 - 0s - loss: 1.6575 - acc: 0.7826
Epoch 492/500
 - 0s - loss: 1.6556 - acc: 0.8261
Epoch 493/500
 - 0s - loss: 1.6568 - acc: 0.7391
Epoch 494/500
 - 0s - loss: 1.6551 - acc: 0.7391
Epoch 495/500
 - 0s - loss: 1.6500 - acc: 0.8261
Epoch 496/500
 - 0s - loss: 1.6521 - acc: 0.7391
Epoch 497/500
 - 0s - loss: 1.6502 - acc: 0.7391
Epoch 498/500
 - 0s - loss: 1.6516 - acc: 0.8261
Epoch 499/500
 - 0s - loss: 1.6491 - acc: 0.7826
Epoch 500/500
 - 0s - loss: 1.6453 - acc: 0.7826
Model Accuracy: 86.96%
['A', 'B', 'C'] -> D
['B', 'C', 'D'] -> E
['C', 'D', 'E'] -> F
['D', 'E', 'F'] -> G
['E', 'F', 'G'] -> H
['F', 'G', 'H'] -> I
['G', 'H', 'I'] -> J
['H', 'I', 'J'] -> K
['I', 'J', 'K'] -> L
['J', 'K', 'L'] -> M
['K', 'L', 'M'] -> N
['L', 'M', 'N'] -> O
['M', 'N', 'O'] -> P
['N', 'O', 'P'] -> Q
['O', 'P', 'Q'] -> R
['P', 'Q', 'R'] -> S
['Q', 'R', 'S'] -> T
['R', 'S', 'T'] -> U
['S', 'T', 'U'] -> V
['T', 'U', 'V'] -> X
['U', 'V', 'W'] -> Z
['V', 'W', 'X'] -> Z
['W', 'X', 'Y'] -> Z