Rescale image pixels to span range [0, 1]

Total training steps inferred from datamodule and devices.

Patchify an image into patches of a given size and stride. Args: x: Image tensor of shape (B, C, H, W). patch_size: Size of

Plots less dense quiver field. Args: ax: Matplotlib axis flow: motion vectors spacing: space (px) between each arrow in g

Plots less dense quiver field. Args: ax: Matplotlib axis flow: motion vectors spacing: space (px) between each arrow in g

Apply the rotation given by a quaternion to a 3D point. Usual torch rules for broadcasting apply. Args: quaternion: Tensor of qu

Perform per-sample random masking by per-sample shuffling. Per-sample shuffling is done by argsort random noise. x: [N, L, D]

Whether to currently save a checkpoint.

Whether to currently run validation.

Warning: does not synchronize the deque!

Restores the signals that were previously configured before :class:`_SignalConnector` replaced them.

r""" Returns:

r""" Returns:

r""" Returns:

r""" Returns:

x: (N, L, patch_size**2 *3) imgs: (N, 3, H, W)

x: (N, L, patch_size**2 *3) imgs: (N, 3, H, W)

x: (N, L, patch_size**2 *3) patch_size: int imgs: (N, 3, H, W)