Run Equi2Cube params: - equi (torch.Tensor): 4 dims (b, c, h, w) - rots (List[dict]): dict of ('yaw', 'pitch', 'roll') - w_face (int)

Call Equi2Cube params: - equi (np.ndarray) - rots (List[Dict[str, float]]) - w_face (int) - cube_format (str): ('horizon', 'list'

Run Equi2Pers params: - equi (torch.Tensor): 4 dims (b, c, h, w) - rots (List[dict]): dict of ('yaw', 'pitch', 'roll') - height, widt

Run Equi2Pers params: - equi (np.ndarray): 4 dims (b, c, h, w) - rot (List[dict]): dict of ('yaw', 'pitch', 'roll') - height, width (

Run Equi2Equi params: - src (torch.Tensor): 4 dims (b, c, h, w) - rot (List[dict]): dict of ('yaw', 'pitch', 'roll') - z_down (bool)

Run Equi2Equi params: - src (np.ndarray): 4 dims (b, c, h, w) - rot (List[dict]): dict of ('yaw', 'pitch', 'roll') - z_down (bool)

Get the size of a tensor

Args: cube: torch.Tensor, [b*m, c, h, w] or [b, m, c, h, w] padding: int

uv: ndarray in shape of [..., 2] h: int, height of the equirectangular image w: int, width of the equirectangular image

Visualizes surface normals, such as predictions of the `MarigoldNormalsPipeline`. Args: normals (`Union[np.ndarray, torc

Visualizes dense uncertainties, such as produced by `MarigoldDepthPipeline` or `MarigoldNormalsPipeline`. Args: uncertai

xyz: ndarray in shape of [..., 3]

Return the xyz cordinates of the unit cube in [F R B L U D] format.