tensorflow/provider.py

[NayeeC]

In provider.py have a get_model function, it look like this:

def get_model(point_cloud, is_training, bn_decay=None):
""" ConvNet baseline, input is BxNx9 gray image """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)

k = 20

adj = tf_util.pairwise_distance(point_cloud[:, :, 6:])
nn_idx = tf_util.knn(adj, k=k)  # (batch, num_points, k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)

out1 = tf_util.conv2d(edge_feature, 64, [1, 1],
                      padding='VALID', stride=[1,1],
                      bn=True, is_training=is_training, weight_decay=weight_decay,
                      scope='adj_conv1', bn_decay=bn_decay, is_dist=True)

(just a part of this function)
in the out1, you can see there is a parameter named weight_decay, whose value is weight_decay, however I cant find this value anywhere, so I just set it as 0. Would you please help me about this question?

[LeopoldACC]

the default value of weight_decay is set 0 in conv2d as shown below
def conv2d(inputs,
num_output_channels,
kernel_size,
scope,
stride=[1, 1],
padding='SAME',
use_xavier=True,
stddev=1e-3,
weight_decay=0.0,
activation_fn=tf.nn.relu,
bn=False,
bn_decay=None,
is_training=None,
is_dist=False):