我可以运行此代码，一切正常：
from keras.datasets import mnist
def plot_history(net_history):
history = net_history.history
import matplotlib.pyplot as plt
losses = history['loss']
val_losses = history['val_loss']
accuracies = history['acc']
val_accuracies = history['val_acc']
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.plot(losses)
plt.plot(val_losses)
plt.legend(['loss', 'val_loss'])
plt.figure()
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.plot(accuracies)
plt.plot(val_accuracies)
plt.legend(['acc', 'val_acc'])
# Load data
(train_images, train_labels), (test_images, test_labels) = mnist.load_data()
# Data attributes
X_train = train_images.reshape(60000, 784)
X_test = test_images.reshape(10000, 784)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
X_train /= 255
X_test /= 255
from keras.utils import np_utils
Y_train = np_utils.to_categorical(train_labels)
Y_test = np_utils.to_categorical(test_labels)
#==================================================
# Creating our model
from keras.models import Sequential
from keras.layers import Dense, Dropout
from keras.optimizers import SGD
from keras.losses import categorical_crossentropy
myModel = Sequential()
myModel.add(Dense(500, activation='relu', input_shape=(784,)))
myModel.add(Dropout(20))
myModel.add(Dense(100, activation='relu'))
myModel.add(Dropout(20))
myModel.add(Dense(10, activation='softmax'))
myModel.summary()
myModel.compile(optimizer=SGD(lr=0.001), loss=categorical_crossentropy,
metrics=['accuracy'])
#==================================================
# Train our model
network_history = myModel.fit(X_train, Y_train, batch_size=128, epochs=20,
validation_split=0.2)
plot_history(network_history)
但是当我无法运行code_2时：
#code_2
from keras.datasets import mnist
(train_images, train_labels), (test_images, test_labels) = mnist.load_data()
# Data attributes
X_train = train_images.reshape(60000, 28 , 28 ,1)
X_test = test_images.reshape(10000, 28, 28, 1)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
X_train /= 255
X_test /= 255
from keras.utils import np_utils
Y_train = np_utils.to_categorical(train_labels)
Y_test = np_utils.to_categorical(test_labels)
# Creating our model
from keras.models import Model
from keras.layers import Conv2D, MaxPool2D, Input, Flatten, Dense
import keras
myInput = Input(shape=(28,28,1))
conv1 = Conv2D(16, (3,3), activation = 'relu', padding = 'same')(myInput)
pool1= MaxPool2D(pool_size = 2)(conv1)
conv2 = Conv2D(32, (3,3), activation = 'relu', padding = 'same')(pool1)
pool2= MaxPool2D(pool_size = 2)(conv2)
flat = Flatten()(pool2)
out_layer = Dense(10, activation = 'softmax')(flat)
myModel = Model(myInput, out_layer)
myModel.summary()
myModel.compile(optimizer = keras.optimizers.Adam(),
loss=keras.losses.categorical_crossentropy, metrics=['accuracy'])
#==================================================
# Train our model
network_history = myModel.fit(X_train, Y_train, batch_size=128, epochs=2,
validation_split=0.2)
ERORR
启动内核时出错 2018 15：26：08.943071：I T：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：1344]找到具有属性的设备0： 名称：GeForce GT 740M主要：3次要：5 memoryClockRate（GHz）：1.0325 pciBusID：0000：01：00.0 totalMemory：2.00GiB freeMemory：1.91GiB 2018 15：26：08.943124：I T：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：1423]添加可见的gpu设备：0 2018 15：26：09.399518：I T：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：911]具有强度1边缘矩阵的设备互连StreamExecutor： 2018 15：26：09.399535：我T：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：917] 0 2018 15：26：09.399540：我T：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：930] 0：N 2018 15：26：09.399664：IT：\ src \ github \ tensorflow \ tensorflow \ core \ common_runtime \ gpu \ gpu_device.cc：1041]创建TensorFlow设备（/ job：localhost / replica：0 / task：0 / device：GPU ：0和1700 MB内存） - ＆gt;物理GPU（设备：0，名称：GeForce GT 740M，pci总线ID：0000：01：00.0，计算能力：3.5）
依赖关系
IPython＆gt; = 4.0：6.2.1（好）
cython＆gt; = 0.21：0.27.3（好）
jedi＆gt; = 0.9.0：0.11.1（OK）
nbconvert＆gt; = 4.0：5.3.1（确定）
numpy＆gt; = 1.7：1.14.2（OK）
pandas＆gt; = 0.13.1：0.22.0（OK）
pycodestyle＆gt; = 2.3：2.3.1（好）
pyflakes＆gt; = 0.6.0：1.6.0（OK）
pygments＆gt; = 2.0：2.2.0（OK）
pylint＆gt; = 0.25：1.8.2（OK）
qtconsole＆gt; = 4.2.0：4.3.1（确定）
rope＆gt; = 0.9.4：0.10.7（OK）
sphinx＆gt; = 0.6.6：1.6.6（OK）
sympy＆gt; = 0.7.3：1.1.1（OK）