:param episodes: how much traces to collect. :param timesteps: the size in timesteps (number of states/transitions) of a single trace. :p

Compute relative angle and distance between a target_location and a current_location :param target_location: location of the target object

Converts the given carla.Image into grayscale in which each pixel contains the depth distance.

Source: https://stackoverflow.com/a/31039095

-1: opposite vectors (pointing in the opposite direction), 0: orthogonal, 1: exactly the same (pointing in the same direction)

Returns the number of traces available at the given folder.

Converts a floating or LearningRateSchedule `value` into a DynamicParameter object

Convert a RGB or BGR image to grayscale using OpenCV (cv2). :param image: input image, a numpy.ndarray. :param is_bgr: tells whether t

Transform some tensors data into a dataset of mini-batches

Decomposes a given number [n] in a scientific-like notation: - n = fractional_part * 10^exponent - e.g. 2.34 could be represented as (0.

Displays the given image on a pygame window :param blend: whether to blend or not the given image. :param window_size: the size of the pygame'

Shows how to draw traffic light actor bounding boxes from a worlds snapshot.

Draws the closest route's waypoint to the given location

Code adapted from carla/PythonAPI/examples/manual_control.py: - White: means static points. - Red: indicates points moving towards the

Draws the path of an actor through the worlds, printing information at each waypoint. @:arg vehicle: see recipe 'actor_spectator'. @:a

dSiLu activation function (i.e. the derivative of SiLU/Swish). Paper: Sigmoid-Weighted Linear Units for Neural Network Function Approximation i

Returns a list of blueprints for vehicles and pedestrians

Connects to the simulator. @:returns a carla.Client instance if the CARLA simulator accepts the connection.

Returns a display used to render images and text. :param window_size: a tuple (width: int, height: int) :param mode: pygame rendering

Returns a dict with additional information either for debugging or learning

Get method for protected member local planner

Learn from experience traces collected by 'collect

Performs imitation learning on the already recorded experience traces (given by `name` argument). - refer to CARLAgent.imitation_learning(...)

Checks whether some input [x] has a shape of the form (H, W, C)

Checks whether some input [x] has a shape of the form (N, D) or (D,)

Kullback-Leibler divergence - Source: https://www.tensorflow.org/api_docs/python/tf/keras/losses/KLD

Computes the Euclidean distance between two carla.Location objects.

Non-Parameteric Linearly Scaled Hyperbolic Tangent Activation Function Sources: - https://www.tensorflow.org/addons/api_docs/python/tfa

Returns the magnitude (norm) of the given 3D vector (tuple/list or carla.Vector3D).

Creates a directory

A Self Regularized Non-Monotonic Neural Activation Function Source: - https://www.tensorflow.org/addons/api_docs/python/tfa/activations