Keras：ValueError：检查目标时出错：预期密集具有形状（10，）但得到形状为（400，）的数组

Question

尝试在 Keras 和 Tensorflow 后端训练我的模型时遇到问题。

import numpy as np
from sklearn.utils import shuffle

# Load Data
df = np.loadtxt("features.txt", delimiter=',') 
print('Features shape:', df.shape)

labels = np.loadtxt("labels.txt", delimiter=',') 
print('Labels shape', labels.shape)

# Replace 10 by 0
labels = np.where(labels == 10, 0, labels) 

# Randomize the data 
data_shuffled, labels_shuffled = shuffle(df, labels, random_state=42)

from keras.models import Sequential
from keras.layers import Dense

# split into input (X) and output (y) variables
X = data_shuffled[:4000]
y = data_shuffled[4000:]

print(X.shape, y.shape)

# define the keras model
model = Sequential()
model.add(Dense(12, input_dim=400, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(10, activation='sigmoid'))
# compile the keras model
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
# fit the keras model on the dataset
model.fit(X, y, epochs=1000, batch_size=100, verbose=0)
# make class predictions with the model
predictions = model.predict_classes(X)
# summarize the first 5 cases
for i in range(5):
    print('%s => %d (expected %d)' % (X[i].tolist(), predictions[i], y[i]))

# evaluate the keras model
accuracy = model.evaluate(X, y)
print('Accuracy: %.2f' % (accuracy*100))

特征形状：(5000, 400) 标签形状 (5000,)

我在调用 model.fit() 的那一行遇到了错误。

我尝试修复此错误的方法： 我尝试使用以下代码重塑 X 和 y numpy 数组： X.reshape(400, -1) y.reshape(400, -1)

但这无济于事。

Answer 1

我认为这行是错误的：y = data_shuffled[4000:]

应该是：y = labels_shuffled[:4000]

那么你需要one-hot-encode：

from sklearn.preprocessing import OneHotEncoder
onehotencoder = OneHotEncoder()
y = onehotencoder.fit_transform(y.reshape(-1,1)).toarray()