Stack new image. Args: image (numpy.ndarray): image observation.

Returns PyTorch's state-action enocder module. Args: observation_shape: observation shape. action_size: action size.

Extract features for the decoder network from a prior or posterior.

Returns a scaler type. Returns: scaler type.

Determine whether a periodic event should happen at this step. Args: phase_step: The incrementing step. batch: The number of steps pr

Return the mean estimate and reset the streaming statistics.

Send a close message to the external process and join it.

The number of submitted samples.

Returns dimension of action-space. Returns: int: dimension of action-space.

Handle multiple type options of `cell_size`. In order to keep the old API of following functions, as well as add support for non-square grids

Returns the dimension of the input. Returns: The dimension of the input

Returns the dimension of the input. Returns: The dimension of the input

Registers reward scaler class. Args: cls: scaler class inheriting ``RewardScaler``.

Returns reversely transformed observations. Args: x: observation. Returns: reversely transformed observation

Get a tensor dtype from a OpenAI Gym space. Args: space: Gym space. Raises: NotImplementedError: For spaces other than Box and D

Signal activity by writing the current timestamp to the PING file. Args: overwrite: Write even if the PING file lists another worker.

Adds a new layer to the network. Arguments: layer: (layer) The new layer to be added to the network. If this i

Clear stacked observation by filling 0.

Instantiate implementation objects with the dataset shapes. This method will be used internally when `fit` method is called. Args:

See predict() above for documentation.

Extract the latent distribution from a prior or posterior state.

Generate greedy sample that maximizes the probability. Borrowed from https://github.com/HorizonRobotics/alf/blob/0f8d0ec5d60ef6f30307c6a66ba38

Returns scorer function of evaluation on environment. This function returns scorer function, which is suitable to the standard scikit-learn s

Returns algorithm configured with json file. The Json file should be the one saved during fitting. .. code-block:: python

When input to a model, each agent is represented by a set of one-hot binary vectors {i, t, l, h, c} encoding its team ID, unique ID,

Returns GPU id. Returns: GPU id.

Generate a new observation with the observation at the index position and num_stacked_observations past observations and actions stac

Registers encoder factory class. Args: cls: encoder factory class inheriting ``EncoderFactory``.

Registers Q function factory class. Args: cls: Q function factory class inheriting ``QFunctionFactory``.

Display the game observation.

samples action for each agent from uniform distribution