Create gt, dt which are list of anns/dets. If use_cats is true only anns/dets corresponding to tuple (img_id, cat_id) will be used. El

Given a vector of widths (ws) and heights (hs) around a center (x_ctr, y_ctr), output a set of anchors (windows).

Given a vector of widths (ws) and heights (hs) around a center (x_ctr, y_ctr), output a set of anchors (windows).

Clip boxes so that they lie inside an image of size `size`. Clip's min max are traced as constants. Use torch.min/max to WAR this issue A

Parse response from the remote service.

Return width, height, x center, and y center for an anchor (window).

Return width, height, x center, and y center for an anchor (window).

Add a model to the world tag in xml. Args: xml_root: The root of the xml tree model_id: The model id model_name: The

Log a negative hit @k for given category

Log a positive hit @k for given category

Run all_gather on arbitrary picklable data (not necessarily tensors) Args: data: any picklable object Returns: list[data]

Convert annotation which can be polygons, uncompressed RLE to RLE. Args: ann (dict) : annotation object Returns:

Computes the area of a set of bounding boxes, which are specified by its (x1, y1, x2, y2) coordinates. Args: boxes (Tensor[N, 4]

This gives the gist of it. Not super important because it doesn't change as much

Returns list of utf-8 byte and a corresponding list of unicode strings. The reversible bpe codes work on unicode strings. This means you

Yield successive n-sized chunks from lst.

Check if the bounding box collides with other objects in the world.

Check if a given pose is reachable for the robot. Return True if it is, False otherwise.

Simple function that adds fixed colors depending on the class

Simple function that adds fixed colors depending on the class

Simple function that adds fixed colors depending on the class

Arguments: original_image (np.ndarray): an image as returned by OpenCV Returns: prediction (BoxList): the de

ind_to_class: {0: "__background__", 1 : "person" ...} target: restricted_negative_list : for datasets with restricted negatives, sample

Create grasp detection model from config.

construct a map such that positive_map[i,j] = True iff box i is associated to token j

Map EfficientNet model name to parameter coefficients.

Initialize from a 4x4 matrix.

query: (n_context, queryL, d) context: (n_context, sourceL, d)

Generates a matrix of anchor boxes in (x1, y1, x2, y2) format. Anchors are centered on stride / 2, have (approximate) sqrt areas of the specified

Generates a matrix of anchor boxes in (x1, y1, x2, y2) format. Anchors are centered on stride / 2, have (approximate) sqrt areas of the specified

Use scenario : we want to sample 2500 examples from 10000 examples, while not sampling overlapping examples with other three process.

v3: with simple prompt such as "there are", "are there?" v4: try to merge some are there / there are together, to avoid sequence bein

Get ann ids that satisfy given filter conditions. Args: img_ids (int array): get anns for given imgs cat_ids (int arra

Get latest synchronized rgb and depth image from all cameras

Get 2D mask image of same size as depth image from 2D bounding box @param bbox: 2D bounding box in the form of [min_x, min_y, max_x, max_y]

Get 2D mask image of same size as depth image by projecting 3D bounding box onto the depth image

Return set of symbol pairs in a word. Word is represented as tuple of symbols (symbols being variable-length strings).