Transform a numpy array with shape (C, H, W) into torch 4D Tensor (1, C, H, W) or (C, H, W)

Wrap mat data from folder

UpsampleConvLayer Upsamples the input and then does a convolution. This method gives better results compared to ConvTranspose2d. ref: http

add deadline noise to the given numpy array (B,H,W)

add impulse noise to the given numpy array (B,H,W)

add stripe noise to the given numpy array (B,H,W)

add gaussian noise to the given numpy array (B,H,W)

add blind gaussian noise to the given numpy array (B,H,W)

add non-iid gaussian noise to the given numpy array (B,H,W)

Data Parallel with a replication callback. An replication callback `__data_parallel_replicate__` of each module will be invoked after being

A thread-safe future implementation. Used only as one-to-one pipe.

Note: num_memblock denotes the number of MemoryBlock currently

ResidualBlock introduced in: https://arxiv.org/abs/1512.03385 x - Relu - Conv - Relu - Conv - x

Pipe for master-slave communication.

An abstract `SyncMaster` object. - During the replication, as the data parallel will trigger an callback of each module, all slave devices should

add blind gaussian noise to the given numpy array (B,H,W)

add mixed noise to the given numpy array (B,H,W) Args: noise_bank: list of noise maker (e.g. AddNoiseImpulse) num_bands: list of n

Wrap data from image folder

Residual Encoder-Decoder Convolution 3D Args: downsample: downsample times, None denotes no downsample

For HSI (c x h x w)

r"""Applies Synchronized Batch Normalization over a 2d or 3d input that is seen as a mini-batch. .. math:: y = \frac{x - mean[x]}{ \

r"""Applies Batch Normalization over a 4d input that is seen as a mini-batch of 3d inputs .. math:: y = \frac{x - mean[x]}{ \sqrt{Va

r"""Applies Batch Normalization over a 5d input that is seen as a mini-batch of 4d inputs .. math:: y = \frac{x - mean[x]}{ \sqrt{Va