Initialize the VGGT model.

* Return the array of supernode-to-column mapping */

This function will generate tracks by camera transform Args: queries: B T N 4 c2w_traj: B T 4 4

\returns an expression of P H P^-1 where H is the matrix represented by \c *this */

Jacobian of the up-projection offset.

\brief Adjoint (conjugate transpose) of the Householder sequence. */

If trunc is None, solve `min sum_i w_i * |a * x_i - y_i|`, otherwise solve `min sum_i min(trunc, w_i * |a * x_i - y_i|)`. w_i must be >=

If trunc is None, solve `min sum_i w_i * |a * x_i - y_i|`, otherwise solve `min sum_i min(trunc, w_i * |a * x_i - y_i|)`. w_i must be >=

This method forwards all its arguments to Qwen2TokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring o

This function will generate tracks by camera transform Args: queries: B T N 3 scale_est: 1 1 shi

Link the output of each node to the input of the next node.

\sa MatrixBase::coeff() * \warning the coordinates must fit into the referenced triangular part */

Args: @ file_path: string representing path to config file. Returns: # config: dict containing all configs.

\returns the size of the main diagonal, which is min(rows(),cols()). * \sa rows(), cols(), SizeAtCompileTime. */

Distort normalized 2D coordinates.

Extract camera trajectory from extrinsic matrices. Args: extrinsic_matrices: torch.Tensor or numpy array of shape (batch, num_ca

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

Args: @ configs: configures for the network. @ load_imagenet_model: whether to initialize from ImageNet-pretrained model. @ imagenet_

\returns a const pointer to the array of the number of non zeros of the inner vectors. * This function is aimed at interoperability with other l

Returns the inverse of the similarity transform.

Merge the patched input into a image with sliding window.

\returns the number of non zero coefficients */

Pixel coordinates in self frame. Returns: torch.Tensor: Pixel coordinates as a tensor of shape (B, h * w, 2).

Predict masks for the given input prompts, using the currently set image. Arguments: point_coords (np.ndarray or None): A

Process all video segments using VIPE SLAM to extract trajectories. Args: segments_dir: Path to segments directory containing vi

\returns the rank of the matrix of which *this is the complete orthogonal * decomposition. * * \note This method has to determine which pivot