问题
In Keras, we can define the network as follows. Are there any way to output the shape after each layer. For instance, I want to print out the shape of inputs after the line defining inputs, then print out the shape of conv1 after the line defining conv1, etc.
inputs = Input((1, img_rows, img_cols))
conv1 = Convolution2D(64, 3, 3, activation='relu', init='lecun_uniform', W_constraint=maxnorm(3), border_mode='same')(inputs)
conv1 = Convolution2D(64, 3, 3, activation='relu', init='lecun_uniform', W_constraint=maxnorm(3), border_mode='same')(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Convolution2D(128, 3, 3, activation='relu', init='lecun_uniform', W_constraint=maxnorm(3), border_mode='same')(pool1)
conv2 = Convolution2D(128, 3, 3, activation='relu', init='lecun_uniform', W_constraint=maxnorm(3), border_mode='same')(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
回答1:
If a layer has a single node (i.e. if it isn't a shared layer), you can get its input tensor, output tensor, input shape and output shape via: layer.input_shape
from keras.utils.layer_utils import layer_from_config
config = layer.get_config()
layer = layer_from_config(config)
Source: https://keras.io/layers/about-keras-layers/
May be this the easiest way to do:
model.layers[layer_of_interest_index].output_shape
回答2:
Just using model.summary(), which gives you pretty print.
回答3:
To print the complete model and all of its dependencies you can also look here: https://keras.io/visualization/
I used this command to save my model visualization as a png:
from keras.utils.visualize_util import plot
plot(model, to_file='model.png')
If you only want to print the layer shape you can do something like this:
layer = model.layers[-1]
print(layer.output._keras_shape)
Prints: (None, 1, 224, 224) # Nr. Filters, Channels, x_dim, y_dim
来源：https://stackoverflow.com/questions/40733984/print-out-the-shape-of-each-layer-in-the-net-architecture