def dropout(x):
if dropout_rate:
if is_sparse_tensor(x):
x = SparseDropout(dropout_rate)(x)
else:
x = tf.keras.layers.Dropout(dropout_rate)(x)
return x
def propagate(x, adj, filters, activation=None, kernel_regularizer=None):
if dropout_rate:
if is_sparse_tensor(x):
x = SparseDropout(dropout_rate)(x)
else:
x = tf.keras.layers.Dropout(dropout_rate)(x)
kwargs = dict(kernel_regularizer=kernel_regularizer, use_bias=False)
if linear_skip_connections:
skip = tf.keras.layers.Dense(filters, **kwargs)
x = graph_conv_factory(filters, **kwargs)([x, adj])
x = x + skip
if activation is not None:
x = activation(x)
else:
x = graph_conv_factory(filters, activation=activation,
**kwargs)([x, adj])
return x
def __init__(
self,
channels,
dropout_rate=0.5,
activation=None,
use_bias=True,
kernel_initializer="glorot_uniform",
bias_initializer="zeros",
attn_kernel_initializer="glorot_uniform",
kernel_regularizer=None,
bias_regularizer=None,
attn_kernel_regularizer=None,
attn_bias_regularizer=None,
activity_regularizer=None,
kernel_constraint=None,
bias_constraint=None,
attn_kernel_constraint=None,
**kwargs,
):
super().__init__(activity_regularizer=activity_regularizer, **kwargs)
self.channels = channels
self.dropout_rate = dropout_rate
self.attn_kernel_initializer = initializers.get(
attn_kernel_initializer)
self.attn_kernel_regularizer = regularizers.get(
attn_kernel_regularizer)
self.attn_kernel_constraint = constraints.get(attn_kernel_constraint)
self.attn_bias_regularizer = regularizers.get(attn_bias_regularizer)
self.activation = activations.get(activation)
kwargs = dict(
kernel_regularizer=attn_kernel_regularizer,
kernel_constraint=attn_kernel_constraint,
kernel_initializer=attn_kernel_initializer,
bias_regularizer=attn_bias_regularizer,
)
self.key_dense = tf.keras.layers.Dense(1,
use_bias=False,
name="key-dense",
**kwargs)
self.query_dense = tf.keras.layers.Dense(1,
use_bias=False,
name="query-dense",
**kwargs)
# self.attn_dense = tf.keras.layers.Dense(
# 2, use_bias=False, name="key-dense", **kwargs
# )
self.values_dense = tf.keras.layers.Dense(
channels,
kernel_regularizer=kernel_regularizer,
kernel_initializer=kernel_initializer,
kernel_constraint=kernel_constraint,
use_bias=False,
name="values-dense",
)
self.input_dropout = tf.keras.layers.Dropout(dropout_rate)
self.attn_dropout = SparseDropout(dropout_rate)
self.values_dropout = tf.keras.layers.Dropout(dropout_rate)
self.use_bias = use_bias
self.bias_constraint = constraints.get(bias_constraint)
self.bias_initializer = initializers.get(bias_initializer)
self.bias_regularizer = regularizers.get(bias_regularizer)
self.bias = None