Creates the replay buffer used by the agent.

Creates the replay buffer used by the agent.

Creates the replay buffer used by the agent.

Builds a training op. Returns: train_op: An op performing one step of training from replay data.

Builds a training op. Returns: train_op: An op performing one step of training.

Builds a training op. Returns: train_op: An op performing one step of training from replay data.

The least generic type all column values are convertible to. >>> _column_type(["1", "2"]) is _int_type True >>> _column_type(["1", "2.3"]

Builds a multi-head Q-network that outputs Q-values for multiple heads. Args: name: str, this name is passed to the tf.keras.Model and used

Builds a Quantile ConvNet. Equivalent to Rainbow ConvNet, only now the output logits are interpreted as quantiles. Args: name: str

Builds a multi-network Q-network that outputs Q-values for each network. Args: name: str, this name is passed to the tf.keras.Model and use

Load multiple checkpoints into a list of replay buffers.

Return a horizontal line with optional colons to indicate column's alignment (as in `pipe` output format).

Runs one iteration of agent/environment interaction.

Visible width of a printed string. ANSI color codes are removed. >>> _visible_width('\x1b[31mhello\x1b[0m'), _visible_width("world") (5, 5)

Split an interable into batches of size `n`. If `n` does not evenly divide `iterable`, the last slice will be smaller. https://stackover

Returns the agent's first action for this episode. Args: observation: numpy array, the environment's initial observation. Returns:

Calculates the distribution of Q-values using the input state tensor.

Creates the output tensor/op given the input state tensor. See https://www.tensorflow.org/api_docs/python/tf/keras/Model for more information

Creates the output tensor/op given the state tensor as input. See https://www.tensorflow.org/api_docs/python/tf/keras/Model for more informat

Creates the output tensor/op given the input state tensor. See https://www.tensorflow.org/api_docs/python/tf/keras/Model for more information

Clip the magnitude of value to be within some value. :param value: :param magnitude: :return:

Creates a DQN agent. Args: sess: A `tf.Session`object for running associated ops. environment: An Atari 2600 environment. replay_log_d

Creates a DQN agent. Args: sess: A `tf.Session`object for running associated ops. environment: An Atari 2600 environment. replay_data_

Creates an online agent. Args: sess: A `tf.Session`object for running associated ops. environment: An Atari 2600 environment. summary_

Given an image $$x$$, samples a bunch of latents from the prior $$z_i$$ and decode them $$\hat x_i$$. Compare this to $$\hat

Only denormalize the scale. Do not add the mean.

``` x = {'foo': [3, 4, 5], 'bar': [1, 2, 3]} ppp.dict_of_list__to__list_of_dicts(x, 3) # Output: # [ # {'foo': 3, 'bar':

Convert each value in the dictionary into a JSON'able primitive. :param d: :return:

Convert something like ``` { 'one.two.three.four': 4, 'one.six.seven.eight': None, 'five.nine.ten': 10, '