Computes the intersection of 3D boxes_dt and boxes_gt. Inputs boxes_dt, boxes_gt are tensors of shape (B, 8, 3) (where B doesn't have to

Args: img_size (int): Input image size. patch_size (int): Patch size. in_chans (int): Number of input ima

Perform the appropriate action at the given iteration. Args: iteration (int): the current iteration, ranged in [0, max_i

Args: input_ch(int) : the number of input channel out_features (list[str]): name of the layers whose outputs should

Returns the Omni3D categories for dataset Args: dataset: str Returns: cats: set of strings with category names

Accumulate per image evaluation results and store the result in self.eval :param p: input params for evaluation :return: None

Draws a scene from multiple different modes. Args: im (array): the image to draw onto K (array): the 3x3 matrix for projecti

Create configs and perform basic setups.

Compute and display summary metrics for evaluation results. Note this functin can *only* be applied on the default parameter setting

Return an activation function given a string

Return an activation function given a string

Input: bbox_preds shape: (num_query, attr_num), attr order: ('loc', 'dim', 'pose', 'uncern') cls_scores shape: (num_q

1x1 convolution with padding

3x3 convolution with padding

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). This is the same as the DropConnect impl I created for Ef

Runs the evaluation for an individual dataset split, assuming all predictions have been passed in. Args: datas

Run per image evaluation on given images and store results (a list of dict) in self.evalImgs

Computes vertices and faces from a 3D cuboid representation. Args: bbox3d (flexible): [[X Y Z W H L]] R (flexible): [np.array

Computes the amount of intersect between two different sets of boxes. Args: box_a (nparray): Mx4 boxes, defined by [x1, y1, x2, y2]

Compute (fractional) per-image repeat factors based on category frequency. The repeat factor for an image is a function of the freque

Checks that plane vertices are coplanar. Returns a bool tensor of size B, where True indicates a box is coplanar.

Checks that the sides of the box have a non zero area. Returns a bool tensor of size B, where True indicates a box is nonzero.

Derive the desired score numbers from summarized COCOeval. Args: omni_eval (None or Omni3Deval): None represents no predi

Evaluate the coco results using COCOEval API.

Input: - src: [bs, sum(hi*wi), 256] - spatial_shapes: h,w of each level [num_level, 2] - level_start_inde

Randomly modifies given color to produce a slightly different color than the color given. Args: color (tuple[double]): a tuple of 3 e

Prepare ._gts and ._dts for evaluation based on params

Adds predictions to the evaluator for dataset_name. This can be any number of predictions, including all predictions passed in at onc

Reduce the values in the dictionary from all processes so that process with rank 0 has the reduced results. Args: input_dict (dic

Approximates the resolution an image with h x w resolution would run through a model at which constrains the scale to a min and max. Arg

Create a list of default :class:`Augmentation` from config. Now it includes resizing and flipping. Returns: list[Augmentation]

Build a LR scheduler from config.

Build the whole model architecture, defined by ``cfg.MODEL.META_ARCHITECTURE``. Note that it does not load any weights from ``cfg``.

ComputeIoU computes the IoUs by sorting based on "score" for either 2D boxes (in 2D mode) or 3D boxes (in 3D mode)