def forward()

in models.py [0:0]

• 28 lines of code
• 8 McCabe index (conditional complexity)

    def forward(self, inp, weights, reuse=False, scope='', stop_grad=False, label=None, **kwargs):
        channels = self.channels
        weights = weights.copy()
        inp = tf.reshape(inp, (tf.shape(inp)[0], 28, 28, 1))

        if FLAGS.swish_act:
            act = swish
        else:
            act = tf.nn.leaky_relu

        if stop_grad:
            for k, v in weights.items():
                if type(v) == dict:
                    v = v.copy()
                    weights[k] = v
                    for k_sub, v_sub in v.items():
                        v[k_sub] = tf.stop_gradient(v_sub)
                else:
                    weights[k] = tf.stop_gradient(v)

        if FLAGS.cclass:
            label_d = tf.reshape(label, shape=(tf.shape(label)[0], 1, 1, self.label_size))
            inp = conv_cond_concat(inp, label_d)

        h1 = smart_conv_block(inp, weights, reuse, 'c1_pre', use_stride=False, activation=act)
        h2 = smart_conv_block(h1, weights, reuse, 'c1', use_stride=True, downsample=True, label=label, extra_bias=False, activation=act)
        h3 = smart_conv_block(h2, weights, reuse, 'c2', use_stride=True, downsample=True, label=label, extra_bias=False, activation=act)
        h4 = smart_conv_block(h3, weights, reuse, 'c3', use_stride=True, downsample=True, label=label, use_scale=False, extra_bias=False, activation=act)

        h5 = tf.reshape(h4, [-1, np.prod([int(dim) for dim in h4.get_shape()[1:]])])
        h6 = act(smart_fc_block(h5, weights, reuse, 'fc_dense'))
        hidden6 = smart_fc_block(h6, weights, reuse, 'fc5')

        return hidden6