computes and stores the average and current value

Convert ndarrays in sample to Tensors.

Crop randomly the image in a sample. Args: output_size (tuple or int): Desired output size. If int, square crop is made.

shape is (h,w,c)

Crop randomly the image in a sample. Args: output_size (tuple or int): Desired output size. If int, square crop is made.

Training loss

BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond Only support x4 upsampling. Args: num_feat (int):

Structure of self_.records: seq: frame: path of images -> {'Blur': <path>, 'Sharp': <path>}

Structure of self_.records: seq: frame: path of images -> {'Blur': <path>, 'Sharp': <path>}

Structure of self_.records: seq: frame: path of images -> {'Blur': <path>, 'Sharp': <path>}

Dense layer for residual dense block

Residual Dense Block

Initialize a basic Conv LSTM cell. Args: shape: int tuple thats the height and width of the hidden states h and c() filter_size: int t

GAN loss

Gradient loss

L1 Charbonnierloss

L1 Charbonnierloss color

Middle network of residual dense blocks

Residual block

shape is (h,w,c)

Pixel Unshuffle