Module: tf_agents.utils

Utils module.


common module: Common utilities for TF-Agents.

composite module: Utilities for dealing with CompositeTensors.

eager_utils module: Common utilities for TF-Agents.

example_encoding module: Utilities for easily encoding nests of numpy arrays into example protos.

lazy_loader module: Lazy loader class.

nest_utils module: Utilities for handling nested tensors.

numpy_storage module: NumpyStorage stores nested objects across multiple numpy arrays.

object_identity module: Utilities for collecting objects based on "is" comparison.

session_utils module: A class to create objects which needs a session to be functional.

tensor_normalizer module: Tensor statistics and normalization.

test_utils module: Common utility functions for testing.

timer module: Timing utility for TF-Agents.

value_ops module: Methods for computing advantages and target values.