Args: tp_b: tp of boxes. tp_m: tp of masks. other arguments see `func: ap_per_class`.

Assigns detections to tracked object (both represented as bounding boxes) Returns 3 lists of 1. matches, 2. unmatched_detections

Compute the average precision, given the recall and precision curves # Arguments recall: The recall curve (list) precision:

Load in a clearml dataset and fill the internal data_dict with its contents.

download the model checkpoint artifact if the resume path starts with WANDB_ARTIFACT_PREFIX arguments: opt (namespace) -

commit the log_dict, model artifacts and Tables to W&B and flush the log_dict. arguments: best_result (boolean): Boolean

x: Features to be visualized module_type: Module type stage: Module stage within model n: M

Log metrics if any and finish the current W&B run

Returns the current bounding box estimate # test arr1 = np.array([[1,2,3,4]]) arr2 = np.array([0]) arr3

From SORT: Computes IOU between two boxes in the form [x1,y1,x2,y2]

Creates kmeans-evolved anchors from training dataset Arguments: dataset: path to data.yaml, or a loaded dataset

Log the dataset as W&B artifact and return the new data file with W&B links arguments: data_file (str) -- the .yaml file

Draw the bounding boxes on a single image and report the result as a ClearML debug sample. arguments: image_path (PosixP

Build evaluation Table. Uses reference from validation dataset table. arguments: predn (list): list of predictions in th

Mean AP@0.5:0.95 of all classes. Return: float.

mask1: [N, n] m1 means number of predicted objects mask2: [M, n] m2 means number of gt objects Note: n means image_w x image_h

Parse input arguments.

Args: img (tensor): img is in cuda, shape: [3, h, w], range: [0, 1] masks (tensor): predicted masks on cuda, shape: [n, h, w]

Args: img_size (tuple): The image size. polygons (list[np.ndarray]): each polygon is [N, M], N is the number of p

Return a (640, 640) overlap mask.

Advances the state vector and returns the predicted bounding box estimate