The last axis should have dimension 2, for x and y values.

Same as tf.foldl but with support for a possibly nested sequence of tensors.

Functional interface for the instance normalization layer. Reference: https://arxiv.org/abs/1607.08022. "Instance Normalization: The Missing I

Performs a backward warp of an image using the predicted flow. Args: im: Batch of images. [num_batch, height, width, channels] fl

2-D locally connected operation with filters represented as a kronecker product of smaller filters Args: inputs: A 4-D tensor of sha

2-D locally connected operation. Works similarly to 2-D convolution except that the weights are unshared, that is, a different set of filter

Configures custom yaml loading behavior and parses config file

Configures custom yaml loading behavior and parses config file

Image quality metric based on maximal signal power vs. power of the noise. Args: true: the ground truth image. pred: the pre

Replaces read ops of each variable in `vars` with new read ops obtained from `read_value()`, thus forcing to read the most up-to-date values

2-D locally connected operation with separable filters. Note that, unlike tf.nn.separable_conv2d, this is spatial separability between dimen

Upsamples the inputs by a factor using bilinear interpolation and the performs conv2d on the upsampled input. This function is more computati

rgb_image: 4-D tensor with pixel intensities between 0 and 1.

rgb_image: 4-D tensor with pixel intensities between 0 and 1.