Allows multi-gpu split along batch

Allows multi-gpu split along batch

Allows multi-gpu split along batch

Allows multi-gpu split along batch

Allows multi-gpu split along batch

Allows multi-gpu split along batch

Allows multi-gpu split along batch

Dilates non-zero sparse depth values to kernel size specified Arg(s): sparse_depth : torch.Tensor[float32] N x 1 x H x W

Sets model to evaluation mode

Sets model to evaluation mode

Sets model to evaluation mode

Sets model to evaluation mode

Exponential map of rotation vector using Rodrigues Arg(s): v : torch.Tensor[float32] 3 element rotation vector Retur

Forward input x through a convolution layer Arg(s): x : torch.Tensor[float32] N x C x H x W input tensor

Get parameters for ``crop`` for center crop. Args: img (numpy.ndarray (C x H x W)): Image to be cropped. output_size

Get parameters for ``crop`` for bottom crop. Args: img (numpy.ndarray (C x H x W)): Image to be cropped. output_size

Get parameters for ``crop`` for bottom crop. Args: img (numpy.ndarray (C x H x W)): Image to be cropped. output_size

Get parameters for ``crop`` for center crop. Args: img (numpy.ndarray (C x H x W)): Image to be cropped. output_size

Convert a translation vector into a 4x4 transformation matrix

Convert a translation vector into a 4x4 transformation matrix

Convert a translation vector into a 4x4 transformation matrix

Samples the image at x, y locations to target x, y locations Arg(s): image : torch.Tensor[float32] N x 3 x H x W RGB ima

Perform horizontal flip on each sample Arg(s): images_arr : list[np.array[float32]] list of N x C x H x W tensors Re

Fill in almost dense depth_map with mask using Scikit Learn Image's inpaint_biharmonic Used for post processing depth maps Arg(s):

Temporarily hardcoding the calibration matrix using calib file from 2011_09_26

Load in the state dict, handle the case if the model was saved with nn.DataParallel Arg(s): model : torch model Model to

Logarithm map of rotation matrix element Arg(s): R : torch.Tensor[float32] 3 x 3 rotation matrix Returns: to

Map camera intrinsics and focal length + baseline to directories Arg(s): paths_only : bool boolean flag if set then crea

Creates N x 2 x H x W meshgrid in x, y directions Arg(s): n_batch : int batch size n_height : int he

Creates N x 2 x H x W meshgrid in x, y directions Arg(s): n_batch : int batch size n_height : int he

Creates N x 2 x H x W meshgrid from -1 to 1 in x, y directions Arg(s): n_batch : int batch size n_height : int

Creates N x 2 x H x W meshgrid in x, y directions Arg(s): n_batch : int batch size n_height : int he

Normalize image to a given range Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H x W

Normalize image to a given range Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H x W

Normalize image to a given range Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H x W

Returns the set of parameters

Returns the list of parameters in the model Returns: list[torch.Tensor[float32]] : list of parameters

Returns the set of parameters

Projects points in camera coordinates to 2D pixel coordinates Arg(s): points : torch.Tensor[float32] N x 4 x (H x W) dep

Apply crop to inputs e.g. images, depth and if available adjust camera intrinsics Arg(s): inputs : list[numpy[float32]]

Apply crop to inputs e.g. images, depth and if available adjust camera intrinsics Arg(s): inputs : list[numpy[float32]]

Randomly selects nonzero elements Arg(s): T : torch.Tensor[float32] N x C x H x W tensor den

Randomly selects nonzero elements Arg(s): T : torch.Tensor[float32] N x C x H x W tensor den

Randomly selects nonzero elements Arg(s): T : torch.Tensor[float32] N x C x H x W tensor den

Remove random nonzero for each sample Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H

Remove random nonzero for each sample Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H

Remove random nonzero for each sample Arg(s): images_arr : list[torch.Tensor[float32]] list of N x C x H

Restore weights of the model Arg(s): checkpoint_path : str path to checkpoint optimizer : to

Convert an axisangle rotation into a 4x4 transformation matrix (adapted from https://github.com/Wallacoloo/printipi) Input 'vec' has to be Bx1

Convert an axisangle rotation into a 4x4 transformation matrix (adapted from https://github.com/Wallacoloo/printipi) Input 'vec' has to be Bx1

Convert an axisangle rotation into a 4x4 transformation matrix (adapted from https://github.com/Wallacoloo/printipi) Input 'vec' has to be Bx1

Runs ScaffNet model if output paths are provided, then will save outputs to a predetermined list of paths Arg(s): model : ScaffN

Fetch image, sparse depth, and ground truth paths for training Arg(s): paths_only : bool if set, then only produces path

Creates teacher_output based on external models Arg(s): external_models : list[str] name of external models to use

Fetch image, sparse depth, and ground truth paths for validation and testing Arg(s): paths_only : bool if set, then only

Creates teacher output based on external models Arg(s): external_models : list[str] name of external models to use