window_size: list, the downsample window size in pyramidal attention. inner_size: int, the size of neighbour attention

patch_len: int, patch len for patch_embedding stride: int, stride for patch_embedding

Number of images in the object dataset.

Does a single level 2d wavelet decomposition of an input. Does separate row and column filtering by two calls to :py:func:`pytorch_wavelets.d

Does a single level 2d wavelet decomposition of an input. Does separate row and column filtering by two calls to :py:func:`pytorch_wavelets.d

Does a 1 level 2d wavelet decomposition of an input. Doesn't do separate row and column filtering. Inputs: x (torch.Tensor): Input t

Download a file to the given path. :param url: URL to download :param file_path: Where to download the content.

Computes the DTW of two sequences. :param prototype: np array [0..b] :param sample: np array [0..t] :param extended: bool

MAPE loss as defined in: https://en.wikipedia.org/wiki/Mean_absolute_percentage_error :param forecast: Forecast values. Shape: batch

sMAPE loss as defined in https://robjhyndman.com/hyndsight/smape/ (Makridakis 1993) :param forecast: Forecast values. Shape: batch,

MASE loss as defined in "Scaled Errors" https://robjhyndman.com/papers/mase.pdf :param insample: Insample values. Shape: batch, time

hidden_states: (B, L, D) Returns: same shape as hidden_states

Forward pass of the DWT. Args: x (tensor): Input of shape :math:`(N, C_{in}, L_{in})` Returns: (yl, yh)

Args: coeffs (yl, yh): tuple of lowpass and bandpass coefficients, should match the format returned by DWT1DForward