Keras：如何使用fit_generator和多个输入

Question

是否有可能有两个fit_generator？

我正在创建一个带有两个输入的模型，模型配置如下所示。

标签Y对X1和X2数据使用相同的标签。

将继续发生以下错误。

检查模型输入时出错：传递给模型的Numpy数组列表不是模型预期的大小。预期看到2个阵列，但是得到以下1个阵列的列表：[array（[[[[0.75686276,0.75686276,0.75686276]，[0.75686276,0.75686276,0.75686276]，[0.75686276,0.75686276,0.75686276]，.... ..，[0.65882355,0.65882355,0.65882355 ...

我的代码看起来像这样：

def generator_two_img(X1, X2, Y,batch_size):
    generator = ImageDataGenerator(rotation_range=15,
                                   width_shift_range=0.2,
                                   height_shift_range=0.2,
                                   shear_range=0.2,
                                   zoom_range=0.2,
                                   horizontal_flip=True,
                                   fill_mode='nearest')

    genX1 = generator.flow(X1, Y, batch_size=batch_size)
    genX2 = generator.flow(X2, Y, batch_size=batch_size)

    while True:
        X1 = genX1.__next__()
        X2 = genX2.__next__()
        yield [X1, X2], Y
  """
      .................................
  """
hist = model.fit_generator(generator_two_img(x_train, x_train_landmark, 
                y_train, batch_size),
                steps_per_epoch=len(x_train) // batch_size, epochs=nb_epoch,
                callbacks = callbacks,
                validation_data=(x_validation, y_validation),
                validation_steps=x_validation.shape[0] // batch_size, 
                `enter code here`verbose=1)

Answer 1

试试这个发电机：

def generator_two_img(X1, X2, y, batch_size):
    genX1 = gen.flow(X1, y,  batch_size=batch_size, seed=1)
    genX2 = gen.flow(X2, y, batch_size=batch_size, seed=1)
    while True:
        X1i = genX1.next()
        X2i = genX2.next()
        yield [X1i[0], X2i[0]], X1i[1]

在Thanh Nguyen评论之后编辑

3输入发电机：

def generator_three_img(X1, X2, X3, y, batch_size):
    genX1 = gen.flow(X1, y,  batch_size=batch_size, seed=1)
    genX2 = gen.flow(X2, y, batch_size=batch_size, seed=1)
    genX3 = gen.flow(X3, y, batch_size=batch_size, seed=1)
    while True:
        X1i = genX1.next()
        X2i = genX2.next()
        X3i = genX3.next()
        yield [X1i[0], X2i[0], X3i[0]], X1i[1]