Scatter the minimum value along the given dimension of `input` into `src` at the indices specified in `index`.

Apply scaled dot product attention to a sparse tensor. Args: qkv (SparseTensor): A [N, *, 3, H, C] sparse tensor containing Qs, Ks,

Expects a torch tensor with length 2 in the last dimension. The coordinates can be in absolute image or normalized coordinates, If th

Zero out the parameters of a module and return it.

Projector MLP Args: in_features (int): input channels out_features (int): output channels mlp_dim (int, o

Apply non-overlapping constraints to the object scores in pred_masks. Here we keep only the highest scoring object at each spatial lo

Consolidate the per-object temporary outputs in `temp_output_dict_per_obj` on a frame into a single output for all objects, including

Implementation of the forward pass through the DPT head. This method processes a specific chunk of frames from the sequence.

Implementation of the forward pass through the DPT head. This method processes a specific chunk of frames from the sequence.

Get the total number of unique object ids received so far in this session.

Resize the object scores to the original video resolution (video_res_masks) and apply non-overlapping constraints for final output.

Inference interface for the model with padding augmentation Padding augmentation fixes the boundary artifacts in the output depth map

Computes the location of zero-crossings on 'edges_x' using linear interpolation with 'edges_weight'.

Computes the location of zero-crossings on 'edges_x' using linear interpolation with 'edges_weight'.

Prepare and flatten visual features.

Run tracking on a single frame based on current inputs and previous memory.

Add new points to a frame.

Align `points_src` to `points_tgt` with respect to a shared xyz scale and z shift. It is similar to `align_affine` but scale and shift are a

Apply basic activation function to pose parameters. Args: pose_enc: Tensor containing encoded pose parameters act_type: Acti

Apply basic activation function to pose parameters. Args: pose_enc: Tensor containing encoded pose parameters act_type: Acti

r"""Sample a tensor using bilinear interpolation `bilinear_sampler(input, coords)` samples a tensor :attr:`input` at coordinates :attr:`coord

Convert the torso of the model to float16.

Convert video to MP4 format using ffmpeg.

Create and load a backbone model given a config. Args: ---- preset: A backbone preset to load pre-defind configs. Returns: -

Decode the structured latent. Args: slat (sp.SparseTensor): The structured latent. formats (List[str]): The

Revised from 6pack dataset/inference_dataset_nocs.py::projection @bbox: (2,3) min/max @line_color: RGB

@color: BGR

Get the image feature on the input batch.

Load a pretrained model. Args: path (str): The path to the model. Can be either local path or a Hugging Face repository.

Load a SAM model from Hugging Face Hub. Args: model_name: Model name from HF_MODELS or direct HF model path

Get the conditioning information for the model. Args: image (Union[torch.Tensor, list[Image.Image]]): The image prompts.

Get the connected components (8-connectivity) of binary masks of shape (N, 1, H, W). Inputs: - mask: A binary mask tensor of shape (N, 1

@poses_normalized: np array, cam_in_ob (opengl convention), normalized to [-1,1] and centered @pose_array: PoseArray, delta poses Return:

Apply inverse log transform: sign(y) * (exp(|y|) - 1) Args: y: Input tensor Returns: Transformed tensor

Apply inverse log transform: sign(y) * (exp(|y|) - 1) Args: y: Input tensor Returns: Transformed tensor

Inverts a depth map with numerical stability. Args: ---- depth (np.ndarray): Depth map to be inverted. eps (float): Minimum v

Resize 2D map by nearest interpolation. Return the nearest neighbor index and mask of the resized map. ### Parameters - `inputs`: a sing

Resize 2D map by nearest interpolation. Return the nearest neighbor index and mask of the resized map. ### Parameters - `inputs`: a sing

Convert rotations given as rotation matrices to quaternions. Args: matrix: Rotation matrices as tensor of shape (..., 3, 3). Re

Apply Non-Maximum Suppression (NMS) horizontally on the given ratio matrix. Args: ---- ratio (np.ndarray): Input ratio matrix.

UV with left-top corner as (-width / diagonal, -height / diagonal) and right-bottom corner as (width / diagonal, height / diagonal)

Just plain rendering, not support any gradient @K: (3,3) np array @ob_in_cams: (N,4,4) torch tensor, openCV camera @projection_mat: np array (4,

Patchify a tensor. Args: x (torch.Tensor): (N, C, *spatial) tensor patch_size (int): Patch size

Convert a pose encoding back to camera extrinsics and intrinsics. This function performs the inverse operation of extri_intri_to_pose_encoding,

@rgb: np array (H,W,3)

Propagate the input points across frames to track in the entire video.