Question

无论我改变什么，dense_1_input总是想要＃（无，296）＆＃39;。错误是：

ValueError: Error when checking model input: expected dense_1_input to have shape (None, 296) but got array with shape (296, 1)    `

代码：

from keras.models import Sequential
from keras.layers import Dense
from random import randrange
import numpy
# fix random seed for reproducibility
numpy.random.seed(7)


def myGenerator():
    # load pima indians dataset
    dataset = numpy.genfromtxt("vectorFile.csv", delimiter=",")
    # split into input (X) and output (Y) variables
    global X
    global Y
    X = dataset[:,0:148]
    Y = dataset[:,149]
    size = len(X)

    while 1:
        outputData = []
        outputAnswer = []

        for i in range(1):          
            firstPick = randrange(0,size)
            firstResult = Y[firstPick]
            firstPlayer = X[firstPick][0]

            while True:
                secondPick = randrange(0,size)
                if firstPlayer==X[secondPick][0]:
                    break

            if Y[firstPick]>Y[secondPick]:
                outputAnswer.append([1,0])
            else:
                outputAnswer.append([0,1])

            result = numpy.concatenate((X[firstPick], X[secondPick]))
            result.reshape(1, 296)
            outputData.append(result)
        yield outputData,outputAnswer

# create model
model = Sequential()

model.add(Dense(12, input_shape=(296, ), activation='relu'))
#model.add(Dense(12, input_dim=296, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(2, activation='sigmoid'))
# Compile model
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=
['accuracy'])
# Fit the model
#model.fit(X, Y, epochs=150, batch_size=10)

#samples_per_epoch = batch_size * number_of_batches
#samples_per_epoch = 100 * 1000

model.fit_generator(myGenerator(), steps_per_epoch=5)
# evaluate the model
scores = model.evaluate(X, Y)
print("\n%s: %.2f%%" % (model.metrics_names[1], scores[1]*100))

Answer 1

以这种方式更改myGenerator()功能：

result = result.reshape(1, 296)

因此，重塑运算符的结果将被保存。

Numpy和keras张量混乱

1 个答案: