credit.ensemble.gaussian_noise

credit.ensemble.gaussian_noise#

Functions#

add_gaussian_noise(x[, std, mean, N])

Adds Gaussian noise to the first portion of channels in a tensor.

Module Contents#

credit.ensemble.gaussian_noise.add_gaussian_noise(x, std=0.1, mean=0.0, N=1)#

Adds Gaussian noise to the first portion of channels in a tensor.

Parameters:

  • x (torch.Tensor) – Input tensor of shape (B, C, T, lat, lon).

  • std (float) – Standard deviation of the Gaussian noise.

  • mean (float) – Mean of the Gaussian noise.

  • N (int) – Number of different noisy tensors to return.

Returns:

Single noisy tensor if N=1, else a list of N noisy tensors.

Return type:

A list of torch.Tensor

Contents