code
stringlengths
66
870k
docstring
stringlengths
19
26.7k
func_name
stringlengths
1
138
language
stringclasses
1 value
repo
stringlengths
7
68
path
stringlengths
5
324
url
stringlengths
46
389
license
stringclasses
7 values
def __init__( self, input_tensor_spec, preprocessing_layers=None, preprocessing_combiner=None, conv_layer_params=None, fc_layer_params=(75, 40), dropout_layer_params=None, activation_fn=tf.keras.activations.relu, kernel_initializer=None, batch_squash=True, ...
Creates an instance of `ValueNetwork`. Network supports calls with shape outer_rank + observation_spec.shape. Note outer_rank must be at least 1. Args: input_tensor_spec: A `tensor_spec.TensorSpec` or a tuple of specs representing the input observations. preprocessing_layers: (Optional...
__init__
python
tensorflow/agents
tf_agents/networks/value_network.py
https://github.com/tensorflow/agents/blob/master/tf_agents/networks/value_network.py
Apache-2.0
def __init__( self, input_tensor_spec, preprocessing_layers=None, preprocessing_combiner=None, conv_layer_params=None, input_fc_layer_params=(75, 40), input_dropout_layer_params=None, lstm_size=(40,), output_fc_layer_params=(75, 40), activation_fn=tf.keras.act...
Creates an instance of `ValueRnnNetwork`. Network supports calls with shape outer_rank + input_tensor_shape.shape. Note outer_rank must be at least 1. Args: input_tensor_spec: A nest of `tensor_spec.TensorSpec` representing the input observations. preprocessing_layers: (Optional.) A ne...
__init__
python
tensorflow/agents
tf_agents/networks/value_rnn_network.py
https://github.com/tensorflow/agents/blob/master/tf_agents/networks/value_rnn_network.py
Apache-2.0
def __init__( self, time_step_spec: ts.TimeStep, action_spec: types.NestedTensorSpec, actor_network: network.Network, policy_state_spec: types.NestedTensorSpec = (), info_spec: types.NestedTensorSpec = (), observation_normalizer: Optional[ tensor_normalizer.TensorNorm...
Builds an Actor Policy given an actor network. Args: time_step_spec: A `TimeStep` spec of the expected time_steps. action_spec: A nest of `BoundedTensorSpec` representing the actions. actor_network: An instance of a `tf_agents.networks.network.Network` to be used by the policy. The networ...
__init__
python
tensorflow/agents
tf_agents/policies/actor_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/actor_policy.py
Apache-2.0
def __init__(self, policy_saver: policy_saver_module.PolicySaver): """Initialize an AsyncPolicySaver. Args: policy_saver: An instance of a `policy_saver.PolicySaver`. """ self._policy_saver = policy_saver self._save_condition_variable = threading.Condition() # These vars should only be a...
Initialize an AsyncPolicySaver. Args: policy_saver: An instance of a `policy_saver.PolicySaver`.
__init__
python
tensorflow/agents
tf_agents/policies/async_policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/async_policy_saver.py
Apache-2.0
def _save_loop(self): """Helper method for the saving thread to wait and execute save requests.""" while True: with self._save_condition_variable: while not self._export_dir: self._save_condition_variable.wait() if self._join_save_thread: return if self._sav...
Helper method for the saving thread to wait and execute save requests.
_save_loop
python
tensorflow/agents
tf_agents/policies/async_policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/async_policy_saver.py
Apache-2.0
def _save(self, export_dir, saving_checkpoint, blocking): """Helper save method, generalizes over save and save_checkpoint.""" self._assert_save_thread_is_alive() if blocking: with self._save_condition_variable: while self._export_dir: logging.info("Waiting for AsyncPolicySaver to f...
Helper save method, generalizes over save and save_checkpoint.
_save
python
tensorflow/agents
tf_agents/policies/async_policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/async_policy_saver.py
Apache-2.0
def flush(self): """Blocks until there is no saving happening.""" with self._save_condition_variable: while self._export_dir: logging.info("Waiting for AsyncPolicySaver to finish.") self._save_condition_variable.wait()
Blocks until there is no saving happening.
flush
python
tensorflow/agents
tf_agents/policies/async_policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/async_policy_saver.py
Apache-2.0
def close(self): """Blocks until there is no saving happening and kills the save_thread.""" with self._save_condition_variable: while self._export_dir: logging.info("Waiting for AsyncPolicySaver to finish.") self._save_condition_variable.wait() self._join_save_thread = True sel...
Blocks until there is no saving happening and kills the save_thread.
close
python
tensorflow/agents
tf_agents/policies/async_policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/async_policy_saver.py
Apache-2.0
def __init__( self, policies: Sequence[py_policy.PyPolicy], multithreading: bool = True ): """Batch together multiple (non-batched) py policies. The policies can be different but must use the same action and observation specs. Args: policies: List python policies (must be non-batched). ...
Batch together multiple (non-batched) py policies. The policies can be different but must use the same action and observation specs. Args: policies: List python policies (must be non-batched). multithreading: Python bool describing whether interactions with the given policies should ha...
__init__
python
tensorflow/agents
tf_agents/policies/batched_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/batched_py_policy.py
Apache-2.0
def _action( self, time_step: ts.TimeStep, policy_state: types.NestedArray, seed: Optional[types.Seed] = None, ) -> ps.PolicyStep: """Forward a batch of time_step and policy_states to the wrapped policies. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. ...
Forward a batch of time_step and policy_states to the wrapped policies. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: An Array, or a nested dict, list or tuple of Arrays representing the previous policy_state. seed: Seed value used to initialize a ...
_action
python
tensorflow/agents
tf_agents/policies/batched_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/batched_py_policy.py
Apache-2.0
def _execute_policy(zip_results_element) -> ps.PolicyStep: """Called on each element of zip return value, in _action method.""" (policy, time_step, policy_state) = zip_results_element return policy.action(time_step, policy_state)
Called on each element of zip return value, in _action method.
_execute_policy
python
tensorflow/agents
tf_agents/policies/batched_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/batched_py_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, temperature: types.FloatOrReturningFloat = 1.0, name: Optional[Text] = None, ): """Builds a BoltzmannPolicy wrapping the given policy. Args: policy: A policy implementing the tf_policy.TFPolicy interface, using a distributi...
Builds a BoltzmannPolicy wrapping the given policy. Args: policy: A policy implementing the tf_policy.TFPolicy interface, using a distribution parameterized by logits. temperature: Tensor or function that returns the temperature for sampling when `action` is called. This parameter appli...
__init__
python
tensorflow/agents
tf_agents/policies/boltzmann_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/boltzmann_policy.py
Apache-2.0
def __init__( self, time_step_spec: ts.TimeStep, action_spec: types.NestedTensorSpec, q_network: network.Network, min_q_value: float, max_q_value: float, observation_and_action_constraint_splitter: Optional[ types.Splitter ] = None, temperature: types.Floa...
Builds a categorical Q-policy given a categorical Q-network. Args: time_step_spec: A `TimeStep` spec of the expected time_steps. action_spec: A `BoundedTensorSpec` representing the actions. q_network: A network.Network to use for our policy. min_q_value: A float specifying the minimum Q-val...
__init__
python
tensorflow/agents
tf_agents/policies/categorical_q_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/categorical_q_policy.py
Apache-2.0
def _distribution(self, time_step, policy_state): """Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy...
Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. Returns: A tfp.distributions.Categoric...
_distribution
python
tensorflow/agents
tf_agents/policies/categorical_q_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/categorical_q_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, epsilon: types.FloatOrReturningFloat, exploration_mask: Optional[Sequence[int]] = None, info_fields_to_inherit_from_greedy: Sequence[Text] = (), name: Optional[Text] = None, ): """Builds an epsilon-greedy MixturePolicy wrapping th...
Builds an epsilon-greedy MixturePolicy wrapping the given policy. Args: policy: A policy implementing the tf_policy.TFPolicy interface. epsilon: The probability of taking the random action represented as a float scalar, a scalar Tensor of shape=(), or a callable that returns a float sca...
__init__
python
tensorflow/agents
tf_agents/policies/epsilon_greedy_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/epsilon_greedy_policy.py
Apache-2.0
def __init__( self, wrapped_policy: tf_policy.TFPolicy, scale: types.Float = 1.0, clip: bool = True, name: Optional[Text] = None, ): """Builds an GaussianPolicy wrapping wrapped_policy. Args: wrapped_policy: A policy to wrap and add OU noise to. scale: Stddev of the ...
Builds an GaussianPolicy wrapping wrapped_policy. Args: wrapped_policy: A policy to wrap and add OU noise to. scale: Stddev of the Gaussian distribution from which noise is drawn. clip: Whether to clip actions to spec. Default True. name: The name of this policy.
__init__
python
tensorflow/agents
tf_agents/policies/gaussian_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/gaussian_policy.py
Apache-2.0
def __init__(self, policy: tf_policy.TFPolicy, name: Optional[Text] = None): """Builds a greedy TFPolicy wrapping the given policy. Args: policy: A policy implementing the tf_policy.TFPolicy interface. name: The name of this policy. All variables in this module will fall under that name. De...
Builds a greedy TFPolicy wrapping the given policy. Args: policy: A policy implementing the tf_policy.TFPolicy interface. name: The name of this policy. All variables in this module will fall under that name. Defaults to the class name.
__init__
python
tensorflow/agents
tf_agents/policies/greedy_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/greedy_policy.py
Apache-2.0
def __init__( self, wrapped_policy: tf_policy.TFPolicy, ou_stddev: types.Float = 1.0, ou_damping: types.Float = 1.0, clip: bool = True, name: Optional[Text] = None, ): """Builds an OUNoisePolicy wrapping wrapped_policy. Args: wrapped_policy: A policy to wrap and add ...
Builds an OUNoisePolicy wrapping wrapped_policy. Args: wrapped_policy: A policy to wrap and add OU noise to. ou_stddev: stddev for the Ornstein-Uhlenbeck noise. ou_damping: damping factor for the Ornstein-Uhlenbeck noise. clip: Whether to clip actions to spec. Default True. name: The...
__init__
python
tensorflow/agents
tf_agents/policies/ou_noise_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/ou_noise_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, info_spec: types.NestedTensorSpec, updater_fn: UpdaterFnType, name: Optional[Text] = None, ): """Builds a TFPolicy wrapping the given policy. PolicyInfoUpdaterWrapper class updates `policy_info` using a user-defined updater fun...
Builds a TFPolicy wrapping the given policy. PolicyInfoUpdaterWrapper class updates `policy_info` using a user-defined updater function. The main use case of this policy wrapper is to annotate `policy_info` with some auxiliary information. For example, appending an identifier to specify which model is ...
__init__
python
tensorflow/agents
tf_agents/policies/policy_info_updater_wrapper.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_info_updater_wrapper.py
Apache-2.0
def load(saved_model_path: Text, checkpoint_path: Optional[Text] = None): """Loads a policy. The argument `saved_model_path` is the path of a directory containing a full saved model for the policy. The path typically looks like '/root_dir/policies/policy', it may contain trailing numbers for the train_step. ...
Loads a policy. The argument `saved_model_path` is the path of a directory containing a full saved model for the policy. The path typically looks like '/root_dir/policies/policy', it may contain trailing numbers for the train_step. `saved_model_path` is expected to contain the following files. (There can be...
load
python
tensorflow/agents
tf_agents/policies/policy_loader.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_loader.py
Apache-2.0
def materialize_saved_model( saved_model_path: Text, checkpoint_path: Text, output_path: Text ): """Materializes a full saved model for a policy. Some training processes generate a full saved model only at step 0, and then generate checkpoints for the model variables at different train steps. In this case ...
Materializes a full saved model for a policy. Some training processes generate a full saved model only at step 0, and then generate checkpoints for the model variables at different train steps. In this case there are no full saved models available for these train steps and you must pass both the path to initia...
materialize_saved_model
python
tensorflow/agents
tf_agents/policies/policy_loader.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_loader.py
Apache-2.0
def _check_spec(spec): """Checks for missing or colliding names in specs.""" spec_names = set() checked = [ _true_if_missing_or_collision(s, spec_names) for s in tf.nest.flatten(spec) ] if any(checked): raise ValueError( 'Specs contain either a missing name or a name collision.\n ' ...
Checks for missing or colliding names in specs.
_check_spec
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def _check_compatible(spec, tensor, ignore_outer_dims=True): """Checks if `spec` is compatible with `tensor`, maybe ignoring outer dims.""" if ignore_outer_dims: tensor = tensor_spec.remove_outer_dims_nest( tensor, tensor.shape.ndims - spec.shape.ndims ) if not spec.is_compatible_with(tensor): ...
Checks if `spec` is compatible with `tensor`, maybe ignoring outer dims.
_check_compatible
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, batch_size: Optional[int] = None, use_nest_path_signatures: bool = True, seed: Optional[types.Seed] = None, train_step: Optional[tf.Variable] = None, input_fn_and_spec: Optional[InputFnAndSpecType] = None, metadata: Optional...
Initialize PolicySaver for TF policy `policy`. Args: policy: A TF Policy. batch_size: The number of batch entries the policy will process at a time. This must be either `None` (unknown batch size) or a python integer. use_nest_path_signatures: SavedModel spec signatures will be created b...
__init__
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def distribution_fn(time_step, policy_state): """Wrapper for policy.distribution() in the SavedModel.""" try: time_step = cast(ts.TimeStep, time_step) outs = policy.distribution( time_step=time_step, policy_state=policy_state ) return tf.nest.map_structure(_check_...
Wrapper for policy.distribution() in the SavedModel.
distribution_fn
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def get_train_step(self) -> types.Int: """Returns the train step of the policy. Returns: An integer. """ if tf.executing_eagerly(): return self._train_step.numpy() else: return tf.identity(self._train_step)
Returns the train step of the policy. Returns: An integer.
get_train_step
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def get_metadata(self) -> Dict[Text, tf.Variable]: """Returns the metadata of the policy. Returns: An a dictionary of tf.Variable. """ if tf.executing_eagerly(): return {k: self._metadata[k].numpy() for k in self._metadata} else: return self._metadata
Returns the metadata of the policy. Returns: An a dictionary of tf.Variable.
get_metadata
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def register_function( self, name: str, fn: InputFnType, input_spec: types.NestedTensorSpec, outer_dims: Sequence[Optional[int]] = (None,), ) -> None: """Registers a function into the saved model. Note: There is no easy way to generate polymorphic functions. This pattern can...
Registers a function into the saved model. Note: There is no easy way to generate polymorphic functions. This pattern can be followed and the `get_concerete_function` can be called with named parameters to register more complex signatures. Those functions can then be passed to the `register_concrete_fu...
register_function
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def register_concrete_function( self, name: str, fn: def_function.Function, assets: Optional[Any] = None ) -> None: """Registers a function into the saved model. This gives you the flexibility to register any kind of polymorphic function by creating the concrete function that you wish to register. ...
Registers a function into the saved model. This gives you the flexibility to register any kind of polymorphic function by creating the concrete function that you wish to register. Args: name: Name of the attribute to use for the saved fn. fn: Function to register. Must be a callable following ...
register_concrete_function
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def save( self, export_dir: Text, options: Optional[tf.saved_model.SaveOptions] = None, ): """Save the policy to the given `export_dir`. Args: export_dir: Directory to save the policy to. options: Optional `tf.saved_model.SaveOptions` object. """ tf.compat.v2.saved_model...
Save the policy to the given `export_dir`. Args: export_dir: Directory to save the policy to. options: Optional `tf.saved_model.SaveOptions` object.
save
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def save_checkpoint( self, export_dir: Text, options: Optional[tf.train.CheckpointOptions] = None, ): """Saves the policy as a checkpoint to the given `export_dir`. This will only work with checkpoints generated in TF2.x. For the checkpoint to be useful users should first call `save` t...
Saves the policy as a checkpoint to the given `export_dir`. This will only work with checkpoints generated in TF2.x. For the checkpoint to be useful users should first call `save` to generate a saved_model of the policy. Checkpoints can then be used to update the policy without having to reload the sa...
save_checkpoint
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def _function_with_flat_signature( function, input_specs, output_spec, include_batch_dimension, batch_size=None ): """Create a tf.function with a given signature for export. Args: function: A callable that can be wrapped in tf.function. input_specs: A tuple nested specs declaring ordered arguments to f...
Create a tf.function with a given signature for export. Args: function: A callable that can be wrapped in tf.function. input_specs: A tuple nested specs declaring ordered arguments to function. output_spec: The nested spec describing the output of the function. include_batch_dimension: Python bool, w...
_function_with_flat_signature
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def specs_from_collect_data_spec( loaded_policy_specs: types.NestedTensorSpec, ) -> Dict[types.NestedSpec, types.NestedSpec]: """Creates policy specs from specs loaded from disk. The PolicySaver saves policy specs next to the saved model as a `struct.StructuredValue` proto. This recreates the original spec...
Creates policy specs from specs loaded from disk. The PolicySaver saves policy specs next to the saved model as a `struct.StructuredValue` proto. This recreates the original specs from the proto. Pass the proto loaded from the file with `tensor_spec.from_pbtxt_file()` to this function. Args: loaded_...
specs_from_collect_data_spec
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def _check_composite_distribution(d): """Checks that the tfp Distributions is a CompositeTensor.""" if isinstance(d, tfp.distributions.Distribution): if not hasattr(d, '_type_spec'): raise TypeError(f'{d} is not a composite tensor.') return d
Checks that the tfp Distributions is a CompositeTensor.
_check_composite_distribution
python
tensorflow/agents
tf_agents/policies/policy_saver.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/policy_saver.py
Apache-2.0
def __init__( self, greedy_policy: py_policy.PyPolicy, epsilon: types.Float, random_policy: Optional[random_py_policy.RandomPyPolicy] = None, epsilon_decay_end_count: Optional[types.Float] = None, epsilon_decay_end_value: Optional[types.Float] = None, random_seed: Optional[type...
Initializes the epsilon-greedy policy. Args: greedy_policy: An instance of py_policy.PyPolicy to use as the greedy policy. epsilon: The probability 0.0 <= epsilon <= 1.0 with which an action will be selected at random. random_policy: An instance of random_py_policy.RandomPyPolicy ...
__init__
python
tensorflow/agents
tf_agents/policies/py_epsilon_greedy_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_epsilon_greedy_policy.py
Apache-2.0
def __init__( self, time_step_spec: ts.TimeStep, action_spec: types.NestedArraySpec, policy_state_spec: types.NestedArraySpec = (), info_spec: types.NestedArraySpec = (), observation_and_action_constraint_splitter: Optional[ types.Splitter ] = None, ): """Initia...
Initialization of PyPolicy class. Args: time_step_spec: A `TimeStep` ArraySpec of the expected time_steps. Usually provided by the user to the subclass. action_spec: A nest of BoundedArraySpec representing the actions. Usually provided by the user to the subclass. policy_state_spe...
__init__
python
tensorflow/agents
tf_agents/policies/py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_policy.py
Apache-2.0
def get_initial_state( self, batch_size: Optional[int] = None ) -> types.NestedArray: """Returns an initial state usable by the policy. Args: batch_size: An optional batch size. Returns: An initial policy state. """ return self._get_initial_state(batch_size)
Returns an initial state usable by the policy. Args: batch_size: An optional batch size. Returns: An initial policy state.
get_initial_state
python
tensorflow/agents
tf_agents/policies/py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_policy.py
Apache-2.0
def action( self, time_step: ts.TimeStep, policy_state: types.NestedArray = (), seed: Optional[types.Seed] = None, ) -> policy_step.PolicyStep: """Generates next action given the time_step and policy_state. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`....
Generates next action given the time_step and policy_state. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: An optional previous policy_state. seed: Seed to use if action uses sampling (optional). Returns: A PolicyStep named tuple containing: ...
action
python
tensorflow/agents
tf_agents/policies/py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_policy.py
Apache-2.0
def _action( self, time_step: ts.TimeStep, policy_state: types.NestedArray, seed: Optional[types.Seed] = None, ) -> policy_step.PolicyStep: """Implementation of `action`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: An Array, or a ...
Implementation of `action`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: An Array, or a nested dict, list or tuple of Arrays representing the previous policy_state. seed: Seed to use if action uses sampling (optional). Returns: A `Polic...
_action
python
tensorflow/agents
tf_agents/policies/py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_policy.py
Apache-2.0
def _get_initial_state(self, batch_size: int) -> types.NestedArray: """Default implementation of `get_initial_state`. This implementation returns arrays of all zeros matching `batch_size` and spec `self.policy_state_spec`. Args: batch_size: The batch shape. Returns: A nested object of...
Default implementation of `get_initial_state`. This implementation returns arrays of all zeros matching `batch_size` and spec `self.policy_state_spec`. Args: batch_size: The batch shape. Returns: A nested object of type `policy_state` containing properly initialized Arrays.
_get_initial_state
python
tensorflow/agents
tf_agents/policies/py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, time_step_spec: ts.TimeStep, action_spec: types.NestedArraySpec, policy_state_spec: types.NestedArraySpec, info_spec: types.NestedArraySpec, use_tf_function: bool = False, batch_time_steps=True, ): """Creates a new ins...
Creates a new instance of the policy. Args: policy: `tf_policy.TFPolicy` instance to wrap and expose as a py_policy. time_step_spec: A `TimeStep` ArraySpec of the expected time_steps. Usually provided by the user to the subclass. action_spec: A nest of BoundedArraySpec representing the ac...
__init__
python
tensorflow/agents
tf_agents/policies/py_tf_eager_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_eager_policy.py
Apache-2.0
def __init__( self, model_path: Text, time_step_spec: Optional[ts.TimeStep] = None, action_spec: Optional[types.NestedTensorSpec] = None, policy_state_spec: types.NestedTensorSpec = (), info_spec: types.NestedTensorSpec = (), load_specs_from_pbtxt: bool = False, use_tf_fu...
Initializes a PyPolicy from a saved_model. *Note* (b/151318119): BoundedSpecs are converted to regular specs when saved into a proto as the `nested_structure_coder` from TF currently doesn't handle BoundedSpecs. Shape and dtypes will still match the original specs. Args: model_path: Path to a sa...
__init__
python
tensorflow/agents
tf_agents/policies/py_tf_eager_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_eager_policy.py
Apache-2.0
def update_from_checkpoint(self, checkpoint_path: Text): """Allows users to update saved_model variables directly from a checkpoint. `checkpoint_path` is a path that was passed to either `PolicySaver.save()` or `PolicySaver.save_checkpoint()`. The policy looks for set of checkpoint files with the file ...
Allows users to update saved_model variables directly from a checkpoint. `checkpoint_path` is a path that was passed to either `PolicySaver.save()` or `PolicySaver.save_checkpoint()`. The policy looks for set of checkpoint files with the file prefix `<checkpoint_path>/variables/variables' Args: ...
update_from_checkpoint
python
tensorflow/agents
tf_agents/policies/py_tf_eager_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_eager_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, batch_size: Optional[int] = None, seed: Optional[types.Seed] = None, ): """Initializes a new `PyTFPolicy`. Args: policy: A TF Policy implementing `tf_policy.TFPolicy`. batch_size: (deprecated) seed: Seed to use if polic...
Initializes a new `PyTFPolicy`. Args: policy: A TF Policy implementing `tf_policy.TFPolicy`. batch_size: (deprecated) seed: Seed to use if policy performs random actions (optional).
__init__
python
tensorflow/agents
tf_agents/policies/py_tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_policy.py
Apache-2.0
def _construct(self, batch_size, graph): """Construct the agent graph through placeholders.""" self._batch_size = batch_size self._batched = batch_size is not None outer_dims = [self._batch_size] if self._batched else [1] with graph.as_default(): self._time_step = tensor_spec.to_nest_placeho...
Construct the agent graph through placeholders.
_construct
python
tensorflow/agents
tf_agents/policies/py_tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_policy.py
Apache-2.0
def restore( self, policy_dir: Text, graph: Optional[tf.Graph] = None, assert_consumed: bool = True, ): """Restores the policy from the checkpoint. Args: policy_dir: Directory with the checkpoint. graph: A graph, inside which policy the is restored (optional). assert...
Restores the policy from the checkpoint. Args: policy_dir: Directory with the checkpoint. graph: A graph, inside which policy the is restored (optional). assert_consumed: If true, contents of the checkpoint will be checked for a match against graph variables. Returns: step: Glo...
restore
python
tensorflow/agents
tf_agents/policies/py_tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/py_tf_policy.py
Apache-2.0
def convert_nest_lists_to_np_array( nested_list: types.NestedTensorOrArray, ) -> Union[Tuple[Any], np.ndarray]: """Convert nest lists to numpy array. Args: nested_list: A nested strucutre of lists. Raises: ValueError: If the input is not collections_abc.Mapping/tuple/list/np.ndarray. Return...
Convert nest lists to numpy array. Args: nested_list: A nested strucutre of lists. Raises: ValueError: If the input is not collections_abc.Mapping/tuple/list/np.ndarray. Returns: A nested structure of numpy arrays.
convert_nest_lists_to_np_array
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def _initial_params( self, batch_size: tf.Tensor ) -> Tuple[tf.Tensor, tf.Tensor]: """Returns initial mean and variance tensors. Broadcasts the initial mean and variance to the requested batch_size. Args: batch_size: The requested batch_size. Returns: mean: A [B, A] sized tensors ...
Returns initial mean and variance tensors. Broadcasts the initial mean and variance to the requested batch_size. Args: batch_size: The requested batch_size. Returns: mean: A [B, A] sized tensors where each row is the initial_mean. var: A [B, A] sized tensors where each row is the initia...
_initial_params
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def actor_func( self, observation: types.NestedTensorOrArray, step_type: Optional[tf.Tensor], policy_state: Sequence[tf.Tensor], ) -> Tuple[types.NestedTensor, types.NestedTensor, Sequence[tf.Tensor]]: """Returns an action to perform using CEM given the q network. Args: observat...
Returns an action to perform using CEM given the q network. Args: observation: Observation for which we need to find a CEM based action. step_type: A `StepType` enum value. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. Retu...
actor_func
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def body( mean, var, i, iters, best_actions, best_scores, best_next_policy_state ): """Defines the body of the while loop in graph. Args: mean: A [B, A] sized tensor, where, a is the size of the action space, indicating the mean value of the sample distribution var : [B, A] ...
Defines the body of the while loop in graph. Args: mean: A [B, A] sized tensor, where, a is the size of the action space, indicating the mean value of the sample distribution var : [B, A] sized tensor, where, a is the size of the action space, indicating the variance value o...
body
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def _score( self, observation, sample_actions, step_type=None, policy_state=() ) -> Tuple[tf.Tensor, Sequence[tf.Tensor]]: """Scores the sample actions internally as part of CEM. Args: observation: A batch of observation tensors or NamedTuples, whatever the q_func will handle. CEM is agno...
Scores the sample actions internally as part of CEM. Args: observation: A batch of observation tensors or NamedTuples, whatever the q_func will handle. CEM is agnostic to it. sample_actions: A [B, N, A] sized tensor, where batch is the batch size, N is the sample size for the CEM, a is ...
_score
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def _score_with_time( self, observation: types.NestedTensorOrArray, sample_actions: types.NestedTensorOrArray, step_type: Optional[tf.Tensor], policy_state: Sequence[tf.Tensor], seq_size: tf.Tensor, ) -> Tuple[tf.Tensor, Sequence[tf.Tensor]]: """Scores the sample actions intern...
Scores the sample actions internally as part of CEM. Args: observation: A batch of state tensors or NamedTuples, whatever the q_func will handle. CEM is agnostic to it. sample_actions: A [BxT, N, A] sized tensor, where batch is the batch size, N is the sample size for the CEM, a is the ...
_score_with_time
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def _distribution( self, time_step: ts.TimeStep, policy_state: Sequence[tf.Tensor] ) -> policy_step.PolicyStep: """Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested d...
Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. Returns: A `PolicyStep` named tuple co...
_distribution
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy.py
Apache-2.0
def __init__( self, input_tensor_spec, sampler_type=None, anchor_point=0.0, dist=DIST_BIMODAL, categorical_action_returns=None, ): """Defines a DummyNet class as a simple continuous dist function. It has a clear maximum at the specified anchor_point. By default, all ac...
Defines a DummyNet class as a simple continuous dist function. It has a clear maximum at the specified anchor_point. By default, all action dimensions for all batches must be near the anchor point. Args: input_tensor_spec: Input Tensor Spec. sampler_type: One of 'continuous', 'continuous_and_o...
__init__
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy_test.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy_test.py
Apache-2.0
def test_actor_func( self, sampler_type=None, anchor_point=0.0, dist=DIST_BIMODAL, categorical_action_returns=None, num_iters=5, ): # pylint: disable=g-doc-args """Helper function to run the tests. Creates the right q_func and tests for correctness. See the _create_q_func...
Helper function to run the tests. Creates the right q_func and tests for correctness. See the _create_q_func documentation to understand the various arguments.
test_actor_func
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy_test.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy_test.py
Apache-2.0
def _score_with_map_fn(self, observation, sample_actions, q_network): """Scores the sample actions using map_fn as part of CEM while loop. Args: observation: A batch of observation tensors or NamedTuples, whatever the q_func will handle. CEM is agnostic to it. sample_actions: A [N, B, A] si...
Scores the sample actions using map_fn as part of CEM while loop. Args: observation: A batch of observation tensors or NamedTuples, whatever the q_func will handle. CEM is agnostic to it. sample_actions: A [N, B, A] sized tensor, where batch is the batch size, N is the sample size for t...
_score_with_map_fn
python
tensorflow/agents
tf_agents/policies/qtopt_cem_policy_test.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/qtopt_cem_policy_test.py
Apache-2.0
def __init__( self, time_step_spec: ts.TimeStep, action_spec: types.NestedArraySpec, policy_state_spec: types.NestedArraySpec = (), info_spec: types.NestedArraySpec = (), seed: Optional[types.Seed] = None, outer_dims: Optional[Sequence[int]] = None, observation_and_action...
Initializes the RandomPyPolicy. Args: time_step_spec: Reference `time_step_spec`. If not None and outer_dims is not provided this is used to infer the outer_dims required for the given time_step when action is called. action_spec: A nest of BoundedArraySpec representing the actions to s...
__init__
python
tensorflow/agents
tf_agents/policies/random_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/random_py_policy.py
Apache-2.0
def _calculate_log_probability(outer_dims, action_spec): """Helper function for calculating log prob of a uniform distribution. Each item in the returned tensor will be equal to: |action_spec.shape| * log_prob_of_each_component_of_action_spec. Note that this method expects the same value for all outer_dims be...
Helper function for calculating log prob of a uniform distribution. Each item in the returned tensor will be equal to: |action_spec.shape| * log_prob_of_each_component_of_action_spec. Note that this method expects the same value for all outer_dims because we're sampling uniformly from the same distribution fo...
_calculate_log_probability
python
tensorflow/agents
tf_agents/policies/random_tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/random_tf_policy.py
Apache-2.0
def __init__( self, time_step_spec: ts.TimeStep, action_spec: types.NestedArraySpec, action_script: Sequence[Tuple[int, types.NestedArray]], ): """Instantiates the scripted policy. The Action script can be configured through gin. e.g: ScriptedPyPolicy.action_script = [ (...
Instantiates the scripted policy. The Action script can be configured through gin. e.g: ScriptedPyPolicy.action_script = [ (1, { "action1": [[5, 2], [1, 3]], "action2": [[4, 6]]},), (0, { "action1": [[8, 1], [9, 2]], "action2": [[1, 2]]},), (2, { "acti...
__init__
python
tensorflow/agents
tf_agents/policies/scripted_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/scripted_py_policy.py
Apache-2.0
def __init__( self, policy: tf_policy.TFPolicy, smoothing_coefficient: float, name: Optional[Text] = None, ): """Adds TemporalActionSmoothing to the given policy. smoothed_action = previous_action * smoothing_coefficient + action * (1.0 - smoothing_coefficient)) ...
Adds TemporalActionSmoothing to the given policy. smoothed_action = previous_action * smoothing_coefficient + action * (1.0 - smoothing_coefficient)) Args: policy: A policy implementing the tf_policy.TFPolicy interface. smoothing_coefficient: Coefficient used for smoothing ac...
__init__
python
tensorflow/agents
tf_agents/policies/temporal_action_smoothing.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/temporal_action_smoothing.py
Apache-2.0
def _get_initial_state(self, batch_size): """Creates zero state tuple with wrapped initial state and smoothing vars. Args: batch_size: The batch shape. Returns: A tuple of (wrapped_policy_initial_state, initial_smoothing_state) """ wrapped_initial_state = self._wrapped_policy.get_initi...
Creates zero state tuple with wrapped initial state and smoothing vars. Args: batch_size: The batch shape. Returns: A tuple of (wrapped_policy_initial_state, initial_smoothing_state)
_get_initial_state
python
tensorflow/agents
tf_agents/policies/temporal_action_smoothing.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/temporal_action_smoothing.py
Apache-2.0
def variables(self) -> Sequence[tf.Variable]: """Returns the list of Variables that belong to the policy.""" # Ignore self._variables() in favor of using tf.Module's tracking. return super(TFPolicy, self).variables
Returns the list of Variables that belong to the policy.
variables
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def get_initial_state( self, batch_size: Optional[types.Int] ) -> types.NestedTensor: """Returns an initial state usable by the policy. Args: batch_size: Tensor or constant: size of the batch dimension. Can be None in which case no dimensions gets added. Returns: A nested objec...
Returns an initial state usable by the policy. Args: batch_size: Tensor or constant: size of the batch dimension. Can be None in which case no dimensions gets added. Returns: A nested object of type `policy_state` containing properly initialized Tensors.
get_initial_state
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def action( self, time_step: ts.TimeStep, policy_state: types.NestedTensor = (), seed: Optional[types.Seed] = None, ) -> policy_step.PolicyStep: """Generates next action given the time_step and policy_state. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`...
Generates next action given the time_step and policy_state. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. seed: Seed to use if action performs sampling (op...
action
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def distribution( self, time_step: ts.TimeStep, policy_state: types.NestedTensor = () ) -> policy_step.PolicyStep: """Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a neste...
Generates the distribution over next actions given the time_step. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. Returns: A `PolicyStep` named tuple co...
distribution
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def update( self, policy, tau: float = 1.0, tau_non_trainable: Optional[float] = None, sort_variables_by_name: bool = False, ) -> tf.Operation: """Update the current policy with another policy. This would include copying the variables from the other policy. Args: poli...
Update the current policy with another policy. This would include copying the variables from the other policy. Args: policy: Another policy it can update from. tau: A float scalar in [0, 1]. When tau is 1.0 (the default), we do a hard update. This is used for trainable variables. tau...
update
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def _action( self, time_step: ts.TimeStep, policy_state: types.NestedTensor, seed: Optional[types.Seed] = None, ) -> policy_step.PolicyStep: """Implementation of `action`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a...
Implementation of `action`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. seed: Seed to use if action performs sampling (optional). Returns: A `...
_action
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def _distribution( self, time_step: ts.TimeStep, policy_state: types.NestedTensorSpec ) -> policy_step.PolicyStep: """Implementation of `distribution`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors...
Implementation of `distribution`. Args: time_step: A `TimeStep` tuple corresponding to `time_step_spec()`. policy_state: A Tensor, or a nested dict, list or tuple of Tensors representing the previous policy_state. Returns: A `PolicyStep` named tuple containing: `action`: A (o...
_distribution
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def _get_initial_state(self, batch_size: int) -> types.NestedTensor: """Returns the initial state of the policy network. Args: batch_size: A constant or Tensor holding the batch size. Can be None, in which case the state will not have a batch dimension added. Returns: A nest of zero te...
Returns the initial state of the policy network. Args: batch_size: A constant or Tensor holding the batch size. Can be None, in which case the state will not have a batch dimension added. Returns: A nest of zero tensors matching the spec of the policy network state.
_get_initial_state
python
tensorflow/agents
tf_agents/policies/tf_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy.py
Apache-2.0
def __init__(self, init_var_value, var_scope, name=None): """Initializes policy containing variables with specified value. Args: init_var_value: A scalar specifies the initial value of all variables. var_scope: A String defines variable scope. name: The name of this policy. All variables in t...
Initializes policy containing variables with specified value. Args: init_var_value: A scalar specifies the initial value of all variables. var_scope: A String defines variable scope. name: The name of this policy. All variables in this module will fall under that name. Defaults to the cla...
__init__
python
tensorflow/agents
tf_agents/policies/tf_policy_test.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_policy_test.py
Apache-2.0
def __init__( self, policy: py_policy.PyPolicy, py_policy_is_batched: bool = False, name: Optional[Text] = None, ): """Initializes a new `TFPyPolicy` instance with an Pyton policy . Args: policy: Python policy implementing `py_policy.PyPolicy`. py_policy_is_batched: If Fal...
Initializes a new `TFPyPolicy` instance with an Pyton policy . Args: policy: Python policy implementing `py_policy.PyPolicy`. py_policy_is_batched: If False, time_steps will be unbatched before passing to py_policy.action(), and a batch dimension will be added to the returned action. Th...
__init__
python
tensorflow/agents
tf_agents/policies/tf_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_py_policy.py
Apache-2.0
def _get_initial_state(self, batch_size): """Invokes python policy reset through numpy_function. Args: batch_size: Batch size for the get_initial_state tensor(s). Returns: A tuple of (policy_state, reset_op). policy_state: Tensor, or a nested dict, list or tuple of Tensors, repres...
Invokes python policy reset through numpy_function. Args: batch_size: Batch size for the get_initial_state tensor(s). Returns: A tuple of (policy_state, reset_op). policy_state: Tensor, or a nested dict, list or tuple of Tensors, representing the new policy state. reset_op: a li...
_get_initial_state
python
tensorflow/agents
tf_agents/policies/tf_py_policy.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/tf_py_policy.py
Apache-2.0
def get_num_actions_from_tensor_spec( action_spec: types.BoundedTensorSpec, ) -> int: """Validates `action_spec` and returns number of actions. `action_spec` must specify a scalar int32 or int64 with minimum zero. Args: action_spec: a `BoundedTensorSpec`. Returns: The number of actions described ...
Validates `action_spec` and returns number of actions. `action_spec` must specify a scalar int32 or int64 with minimum zero. Args: action_spec: a `BoundedTensorSpec`. Returns: The number of actions described by `action_spec`. Raises: ValueError: if `action_spec` is not an bounded scalar int32 or...
get_num_actions_from_tensor_spec
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def create_bandit_policy_type_tensor_spec( shape: types.Shape, ) -> types.BoundedTensorSpec: """Create tensor spec for bandit policy type.""" return tensor_spec.BoundedTensorSpec( shape=shape, dtype=tf.int32, minimum=BanditPolicyType.UNKNOWN, maximum=BanditPolicyType.FALCON, )
Create tensor spec for bandit policy type.
create_bandit_policy_type_tensor_spec
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def masked_argmax( input_tensor: types.Tensor, mask: types.Tensor, output_type: tf.DType = tf.int32, ) -> types.Tensor: """Computes the argmax where the allowed elements are given by a mask. If a row of `mask` contains all zeros, then this method will return -1 for the corresponding row of `input_ten...
Computes the argmax where the allowed elements are given by a mask. If a row of `mask` contains all zeros, then this method will return -1 for the corresponding row of `input_tensor`. Args: input_tensor: Rank-2 Tensor of floats. mask: 0-1 valued Tensor of the same shape as input. output_type: Intege...
masked_argmax
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def has_bandit_policy_type( info: Optional[Union[PolicyInfo, PerArmPolicyInfo]], check_for_tensor: bool = False, ) -> bool: """Check if policy info has `bandit_policy_type` field/tensor.""" if not info: return False fields = getattr(info, '_fields', None) has_field = fields is not None and InfoField...
Check if policy info has `bandit_policy_type` field/tensor.
has_bandit_policy_type
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def has_chosen_arm_features( info: Optional[Union[PolicyInfo, PerArmPolicyInfo]], check_for_tensor: bool = False, ) -> bool: """Check if policy info has `chosen_arm_features` field/tensor.""" if not info: return False fields = getattr(info, '_fields', None) has_field = fields is not None and InfoFie...
Check if policy info has `chosen_arm_features` field/tensor.
has_chosen_arm_features
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def set_bandit_policy_type( info: Optional[Union[PolicyInfo, PerArmPolicyInfo]], bandit_policy_type: types.Tensor, ) -> Union[PolicyInfo, PerArmPolicyInfo]: """Sets the InfoFields.BANDIT_POLICY_TYPE on info to bandit_policy_type. If policy `info` does not support InfoFields.BANDIT_POLICY_TYPE, this method ...
Sets the InfoFields.BANDIT_POLICY_TYPE on info to bandit_policy_type. If policy `info` does not support InfoFields.BANDIT_POLICY_TYPE, this method returns `info` as-is (without any modification). Args: info: Policy info on which to set bandit policy type. bandit_policy_type: Tensor containing BanditPoli...
set_bandit_policy_type
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def bandit_policy_uniform_mask( values: types.Tensor, mask: types.Tensor ) -> types.Tensor: """Set bandit policy type tensor to BanditPolicyType.UNIFORM based on mask. Set bandit policy type `values` to BanditPolicyType.UNIFORM; returns tensor where output[i] is BanditPolicyType.UNIFORM if mask[i] is True, o...
Set bandit policy type tensor to BanditPolicyType.UNIFORM based on mask. Set bandit policy type `values` to BanditPolicyType.UNIFORM; returns tensor where output[i] is BanditPolicyType.UNIFORM if mask[i] is True, otherwise it is left as values[i]. Args: values: Tensor containing `BanditPolicyType` enumera...
bandit_policy_uniform_mask
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def create_chosen_arm_features_info_spec( observation_spec: types.NestedTensorSpec, ) -> types.NestedTensorSpec: """Creates the chosen arm features info spec from the arm observation spec.""" arm_spec = observation_spec[bandit_spec_utils.PER_ARM_FEATURE_KEY] return tensor_spec.remove_outer_dims_nest(arm_spec,...
Creates the chosen arm features info spec from the arm observation spec.
create_chosen_arm_features_info_spec
python
tensorflow/agents
tf_agents/policies/utils.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/utils.py
Apache-2.0
def __init__(self, action_spec, sample_clippers=None, sample_rejecters=None): """Creates an ActionsSampler. Args: action_spec: A nest of BoundedTensorSpec representing the actions. sample_clippers: A list of callables that are applied to the generated samples. These callables take in a nest...
Creates an ActionsSampler. Args: action_spec: A nest of BoundedTensorSpec representing the actions. sample_clippers: A list of callables that are applied to the generated samples. These callables take in a nested structure matching the action_spec and must return a matching structure. ...
__init__
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler.py
Apache-2.0
def refit_distribution_to(self, target_sample_indices, samples): """Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A nested structure corresponding to action_spec. Each action is a [B, N, A] sized ...
Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A nested structure corresponding to action_spec. Each action is a [B, N, A] sized tensor. Returns: mean: A nested structure containing [B, A] s...
refit_distribution_to
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous.py
Apache-2.0
def sample_batch_and_clip(self, num_samples, mean, var, state=None): """Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A nested structure containing [B, A] shaped tensor representing the mean of the actions ...
Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A nested structure containing [B, A] shaped tensor representing the mean of the actions to be sampled. var: A nested structure containing [B, A] shaped tensor...
sample_batch_and_clip
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous.py
Apache-2.0
def __init__( self, action_spec, sample_clippers=None, sub_actions_fields=None, sample_rejecters=None, max_rejection_iterations=10, ): """Builds a GaussianActionsSampler. Args: action_spec: A dict of BoundedTensorSpec representing the actions. sample_clippers: ...
Builds a GaussianActionsSampler. Args: action_spec: A dict of BoundedTensorSpec representing the actions. sample_clippers: A list of list of sample clipper functions. The function takes a dict of Tensors of actions and a dict of Tensors of the state, output a dict of Tensors of clipped ...
__init__
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
Apache-2.0
def refit_distribution_to(self, target_sample_indices, samples): """Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A dict corresponding to action_spec. Each action is a [B, N, A] sized tensor. ...
Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A dict corresponding to action_spec. Each action is a [B, N, A] sized tensor. Returns: mean: A dict containing [B, A] sized tensors where each ...
refit_distribution_to
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
Apache-2.0
def sample_batch_and_clip(self, num_samples, mean, var, state=None): """Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A dict containing [B, A] shaped tensor representing the mean of the actions to be sample...
Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A dict containing [B, A] shaped tensor representing the mean of the actions to be sampled. var: A dict containing [B, A] shaped tensor representing the varian...
sample_batch_and_clip
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_continuous_and_one_hot.py
Apache-2.0
def refit_distribution_to(self, target_sample_indices, samples): """Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A nested structure corresponding to action_spec. Each action is a [B, N, A] sized ...
Refits distribution according to actions with index of ind. Args: target_sample_indices: A [B, M] sized tensor indicating the index samples: A nested structure corresponding to action_spec. Each action is a [B, N, A] sized tensor. Returns: mean: A nested structure containing [B, A] s...
refit_distribution_to
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_hybrid.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_hybrid.py
Apache-2.0
def sample_batch_and_clip(self, num_samples, mean, var, state=None): """Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A nested structure containing [B, A] shaped tensor representing the mean of the actions ...
Samples and clips a batch of actions [B, N, A] with mean and var. Args: num_samples: Number of actions to sample each round. mean: A nested structure containing [B, A] shaped tensor representing the mean of the actions to be sampled. var: A nested structure containing [B, A] shaped tensor...
sample_batch_and_clip
python
tensorflow/agents
tf_agents/policies/samplers/qtopt_cem_actions_sampler_hybrid.py
https://github.com/tensorflow/agents/blob/master/tf_agents/policies/samplers/qtopt_cem_actions_sampler_hybrid.py
Apache-2.0
def __init__( self, data_spec, capacity=1000, completed_only=False, buffer_size=8, name_prefix='EpisodicReplayBuffer', device='cpu:*', seed=None, begin_episode_fn=None, end_episode_fn=None, dataset_drop_remainder=False, dataset_window_shift=None, ...
Creates an EpisodicReplayBuffer. This class receives a dataspec and capacity and creates a replay buffer supporting read/write operations, organized into episodes. This uses an underlying EpisodicTable with capacity equal to capacity. Each row in the table can have an episode of unbounded length. ...
__init__
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def create_episode_ids(self, num_episodes=None): """Returns a new tensor containing initial invalid episode ID(s). This tensor is meant to be passed to methods like `add_batch` and `extend_episodes`; those methods will return an updated set of episode id values in their output. To keep track of update...
Returns a new tensor containing initial invalid episode ID(s). This tensor is meant to be passed to methods like `add_batch` and `extend_episodes`; those methods will return an updated set of episode id values in their output. To keep track of updated episode IDs across multiple TF1 session run calls,...
create_episode_ids
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def add_sequence(self, items, episode_id): """Adds a sequence of items to the replay buffer for the selected episode. Args: items: A sequence of items to be added to the buffer. Items will have the same structure as the data_spec of this class, but the tensors in items will have an outer ...
Adds a sequence of items to the replay buffer for the selected episode. Args: items: A sequence of items to be added to the buffer. Items will have the same structure as the data_spec of this class, but the tensors in items will have an outer sequence dimension in addition to the correspondin...
add_sequence
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _add_steps(): """Add sequence of items to the buffer.""" inc_episode_length = self._increment_episode_length_locked( episode_location, num_steps ) write_data_op = self._data_table.append(episode_location, items) with tf.control_dependencies([inc_episod...
Add sequence of items to the buffer.
_add_steps
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def add_batch(self, items, episode_ids): """Adds a batch of single steps for the corresponding episodes IDs. Args: items: A batch of items to be added to the buffer. Items will have the same structure as the data_spec of this class, but the tensors in items will have an extra outer dimens...
Adds a batch of single steps for the corresponding episodes IDs. Args: items: A batch of items to be added to the buffer. Items will have the same structure as the data_spec of this class, but the tensors in items will have an extra outer dimension `(num_episodes, ...)` in addition to ...
add_batch
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _get_next( self, sample_batch_size=None, num_steps=None, time_stacked=None ): """Returns an episode sampled uniformly from the buffer. Args: sample_batch_size: Not used num_steps: Not used time_stacked: Not used Returns: A 2-tuple containing: - An episode sampl...
Returns an episode sampled uniformly from the buffer. Args: sample_batch_size: Not used num_steps: Not used time_stacked: Not used Returns: A 2-tuple containing: - An episode sampled uniformly from the buffer. - BufferInfo NamedTuple, containing the episode id.
_get_next
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _as_dataset( self, sample_batch_size=None, num_steps=None, sequence_preprocess_fn=None, num_parallel_calls=tf.data.experimental.AUTOTUNE, ): """Creates a dataset that returns episodes entries from the buffer. The dataset behaves differently depending on if `num_steps` is pro...
Creates a dataset that returns episodes entries from the buffer. The dataset behaves differently depending on if `num_steps` is provided or not. If `num_steps = None`, then entire episodes are sampled uniformly at random from the buffer. If `num_steps != None`, then we attempt to sample uniformly acr...
_as_dataset
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _get_episode_locations(_): """Sample episode ids according to value of num_steps.""" if num_steps is None: # Just want to get a uniform sampling of episodes. episode_ids = self._sample_episode_ids( shape=[episode_id_buffer_size], seed=self._seed ) else: ...
Sample episode ids according to value of num_steps.
_get_episode_locations
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _read_tensor_list_and_id(row): """Read the TensorLists out of the table row, get id and num_frames.""" # Return a flattened tensor list flat_tensor_lists = tuple( tf.nest.flatten(self._data_table.get_episode_lists(row)) ) # Due to race conditions, not all entries ...
Read the TensorLists out of the table row, get id and num_frames.
_read_tensor_list_and_id
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _random_slice(flat_tensor_lists, id_, num_frames): """Take a random slice from the episode, of length num_steps.""" # Sample uniformly between [0, num_frames - num_steps] start_slice = tf.random.uniform( (), minval=0, maxval=num_frames - num_steps + 1, ...
Take a random slice from the episode, of length num_steps.
_random_slice
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _single_deterministic_pass_dataset( self, sample_batch_size=None, num_steps=None, sequence_preprocess_fn=None, num_parallel_calls=tf.data.experimental.AUTOTUNE, ): """Creates a dataset that returns entries from the buffer in fixed order. Args: sample_batch_size: (Optio...
Creates a dataset that returns entries from the buffer in fixed order. Args: sample_batch_size: (Optional.) An optional batch_size to specify the number of items to return. See as_dataset() documentation. **NOTE** This argument may only be provided when `num_steps is not None`. Otherwise ...
_single_deterministic_pass_dataset
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0
def _gather_all(self): """Returns all the items currently in the buffer. Returns: A tuple containing two entries: - All the items currently in the buffer (nested). - The items ids. Raises: ValueError: If the data spec contains lists that must be converted to tuples. ...
Returns all the items currently in the buffer. Returns: A tuple containing two entries: - All the items currently in the buffer (nested). - The items ids. Raises: ValueError: If the data spec contains lists that must be converted to tuples.
_gather_all
python
tensorflow/agents
tf_agents/replay_buffers/episodic_replay_buffer.py
https://github.com/tensorflow/agents/blob/master/tf_agents/replay_buffers/episodic_replay_buffer.py
Apache-2.0