Question about Intra-Normalization

[borongyuan]

Hi,
I noticed that for intra-normalization, it's performing L2 norm along axis 1. And axis 1 is n_clusters after reduce_sum. Based on my understanding of intra-normalization, L2 norm should be performed along feature_dim, which is axis 2. Did I misunderstand?

hfnet/hfnet/models/utils/layers.py

Lines 79 to 92 in 0157784

	clusters = slim.model_variable(
	'clusters', shape=[1, 1, 1, config['n_clusters'], feature_dim],
	initializer=slim.initializers.xavier_initializer(), trainable=training)
	residuals = clusters - tf.expand_dims(feature_map, axis=3)
	residuals *= tf.expand_dims(memberships, axis=-1)
	if mask is not None:
	residuals *= tf.to_float(mask)[..., tf.newaxis, tf.newaxis]
	descriptor = tf.reduce_sum(residuals, axis=[1, 2])
	descriptor = tf.nn.l2_normalize(descriptor, axis=1) # intra-normalization
	descriptor = tf.reshape(descriptor,
	[batch_size, feature_dim*config['n_clusters']])
	descriptor = tf.nn.l2_normalize(descriptor, axis=1)
	return descriptor