Since Keras 2.0 removed the Highway Network layer, here's my attempt at implementing something equivalent using the functional API

keras2-highway-network.py

import keras.backend as K
from keras.layers import Dense, Activation, Multiply, Add, Lambda
import keras.initializers

def highway_layers(value, n_layers, activation="tanh", gate_bias=-3):
    dim = K.int_shape(value)[-1]
    gate_bias_initializer = keras.initializers.Constant(gate_bias)
    for i in range(n_layers):     
        gate = Dense(units=dim, bias_initializer=gate_bias_initializer)(value)
        gate = Activation("sigmoid")(gate)
        negated_gate = Lambda(
            lambda x: 1.0 - x,
            output_shape=(dim,))(gate)
        transformed = Dense(units=dim)(value)
        transformed = Activation(activation)(transformed)
        transformed_gated = Multiply()([gate, transformed])
        identity_gated = Multiply()([negated_gate, value])
        value = Add()([transformed_gated, identity_gated])
    return value


def conv_highway_layer(input, activation="tanh",kernel_size=3, gate_bias=-3):
    # y = H(x, W_h)*T(x, W_t) + x*(1-T(x, W_t))
    # T(x, W_t, b_t) = sigmoid(W_t'*x+b_t)
    # H(x, W_h) = tanh(W_h'*x)
    # T = gate, H = regular layer, x = input, y = output
    dim = K.int_shape(input)[-1]
    gate = Convolution2D(dim, (kernel_size, kernel_size), padding='same', activation='sigmoid',
                         bias_initializer=keras.initializers.Constant(gate_bias))(input)  #T(x, W_t)
    H = Convolution2D(dim, (kernel_size, kernel_size), padding='same',
                                activation=activation)(input) #tanh(W_h'*x)
    negated_gate = Lambda(lambda x: 1.0 - x)(gate)
    transformed_gated = Multiply()([H, gate])  # H(x, W_h)*T(x, W_t)
    identity_gated = Multiply()([input, negated_gate])  #x*(1-T(x, W_t))
    output = Add()([transformed_gated, identity_gated])  # H(x, W_h)*T(x, W_t) + x*(1-T(x, W_t))
    return output

https://arxiv.org/pdf/1505.00387.pdf