Keras: How can i implement inception style network with keras

[erogol]

As inception style networks uses I like to branch a layer's output and merge at the end of the next layer. I implemented that by the following snippet but I got DisconnectedInputError

    # model is defined previously as the last layers output
    # conv 3x3
    left = Sequential()
    left = ConvFactory(left, kernel=(3, 3), stride=(2, 2), num_filter=ch_3x3,
                      input_shape = model.output_shape[1:])
    # pool
    right = Sequential()
    right.add(layers.MaxPooling2D(pool_size=(3, 3), strides=(2, 2),
             input_shape = model.output_shape[1:]))

    model.add(layers.Merge([left, right], mode='concat',concat_axis=1))

[huster-wgm]

`from keras.layers import merge, Convolution2D, MaxPooling2D, Input

input_img = Input(shape=(3, 256, 256))

tower_1 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(input_img)
tower_1 = Convolution2D(64, 3, 3, border_mode='same', activation='relu')(tower_1)

tower_2 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(input_img)
tower_2 = Convolution2D(64, 5, 5, border_mode='same', activation='relu')(tower_2)

tower_3 = MaxPooling2D((3, 3), strides=(1, 1), border_mode='same')(input_img)
tower_3 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(tower_3)

output = merge([tower_1, tower_2, tower_3], mode='concat', concat_axis=1)`

These codes come directly from official document. Hope it will help.
http://keras.io/getting-started/functional-api-guide/#getting-started-with-the-keras-functional-api

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs, but feel free to re-open it if needed.