tensortools.utils

standardize

standardize(arr)

Normalizes an array by subtracting the mean and dividing by the standard deviation. See https://en.wikipedia.org/wiki/Feature_scaling.

standardize([1, 2, 3]) array([-1.22474487, 0. , 1.22474487])

Arguments:

  • arr: The array to normalize.

Returns:

The normalized array.

iou

iou(box_a, box_b)

Calculates the IOU between two boxes.

For example:

iou([0.5, 0.5], [1, 1]) 0.25

Arguments:

  • box_a:
  • box_b:

Returns:

avg_iou

avg_iou(annotations, anchors)

Calculates the average iou between the given anchors. Only the max IOU between each annotation and anchor is used to calculate the average.

Arguments:

  • annotations: The annotations, shape [n_annotations, 2].
  • anchors: The anchors, shape [n_anchors, 2].

Returns:

The average IOU.

count

count(arr, value)

Counts the occurrences of the value in the given iterable.

Arguments:

  • arr: The iterable.
  • value: The value to count.

Returns:

The number of occurrences.

download_file

download_file(url, dst)

Downloads a file from a url to the given destination with % finished bar.

Arguments:

  • url: The url.
  • dst: The destination.