spyrit.misc.statistics

Functions

Cov2Var(Cov[, out_shape])

Extracts Variance Matrix from Covariance Matrix.

cov_1(dataloader, mean, device[, n_loop])

The covariance is computed across batches, channels, and image rows.

cov_2(dataloader, mean, device[, n_loop])

Computes 2D covariance matrix computed across batches and channels.

cov_walsh(dataloader, mean, device[, n_loop])

nloop > 1 is relevant for dataloaders with random crops such as that provided by data_loaders_ImageNet

data_loaders_ImageNet(train_root[, ...])

Args:

data_loaders_imagenet(train_root[, ...])

Args:

data_loaders_stl10(data_root[, img_size, ...])

Args:

img2mask(Ord, M)

Returns subsampling mask from order matrix

mea_abs_model(dataloader, device, model, root)

mean_1(dataloader, device[, n_loop])

The mean is computed across batches, channels, and image rows.

mean_2(dataloader, device[, n_loop])

Computes 2D mean image computed across batches and channels

mean_walsh(dataloader, device[, n_loop])

nloop > 1 is relevant for dataloaders with random crops such as that provided by data_loaders_ImageNet

optim_had(dataloader, root)

Computes image that ranks the hadamard coefficients

stat_1(dataloader, device, root[, n_loop])

1D mean and covariance matrix of an image database.

stat_2(dataloader, device, root[, n_loop])

Computes and saves 2D mean image and covariance matrix of an image database

stat_fwalsh_S(dataloader, device, root)

stat_fwalsh_S_stl10([stat_root, data_root, ...])

Fast Walsh S-transform of X in "2D"

stat_imagenet([stat_root, data_root, ...])

Args:

stat_mean_coef_from_model(dataloader, ...)

stat_model(dataloader, device, model, root)

stat_psnr(model, dataloader, device[, ...])

nloop > 1 is relevant for dataloaders with random crops such as that provided by data_loaders_ImageNet

stat_ssim(model, dataloader, device[, ...])

nloop > 1 is relevant for dataloaders with random crops such as that provided by data_loaders_ImageNet

stat_walsh(dataloader, device, root[, n_loop])

nloop > 1 is relevant for dataloaders with random crops such as that provided by data_loaders_ImageNet

stat_walsh_ImageNet([stat_root, data_root, ...])

Args:

stat_walsh_stl10([stat_root, data_root, ...])

Args:

transform_gray_norm(img_size[, normalize])

Args:

Classes

CenterCrop(img_size)

Args: