Apply Gaussian Blur to the PIL image.

Normlize image by given mean and std.

Prepare sample for usage as network input.

Resize sample to given size (width, height).

A dictionary of running averages.

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

Gradient loss

Convert a ``PIL Image`` or ``numpy.ndarray`` to tensor, or keep as is if already a tensor.

Residual convolution module.

Residual convolution module.

SILog loss (pixel-wise)

Track a series of values and provide access to smoothed values over a window or the global series average.

Cross entropy loss

Feature fusion block.

Feature fusion block.

Kozachenko-Leonenko entropic loss regularizer from Sablayrolles et al. - 2018 - Spreading vectors for similarity search

Linear layer to train on top of frozen features

Resize sample to given size (width, height).

Resize sample to given size (width, height).

Use DINOv2 pre-trained models