def test_positive_parameterizer(self):
shape = (1, 2, 3, 4)
var = self._test_parameterizer(
parameterizers.NonnegativeParameterizer(minimum=.1),
tf.initializers.random_uniform(), shape)
self.assertEqual(var.shape, shape)
self.assertTrue(np.all(var >= .1))
from __future__ import print_function
# Dependency imports
from tensorflow.python.eager import context
from tensorflow.python.framework import ops
from tensorflow.python.framework import tensor_shape
from tensorflow.python.layers import base
from tensorflow.python.ops import array_ops
from tensorflow.python.ops import init_ops
from tensorflow.python.ops import math_ops
from tensorflow.python.ops import nn
from tensorflow_compression.python.layers import parameterizers
_default_beta_param = parameterizers.NonnegativeParameterizer(minimum=1e-6)
_default_gamma_param = parameterizers.NonnegativeParameterizer()
class GDN(base.Layer):
"""Generalized divisive normalization layer.
Based on the papers:
> "Density modeling of images using a generalized normalization
> transformation"<br />
> J. Ballé, V. Laparra, E.P. Simoncelli<br />
> https://arxiv.org/abs/1511.06281
> "End-to-end optimized image compression"<br />
> J. Ballé, V. Laparra, E.P. Simoncelli<br />