tf_agents.utils.nest_utils.flatten_multi_batched_nested_tensors

Reshape tensors to contain only one batch dimension.

tf_agents.utils.nest_utils.flatten_multi_batched_nested_tensors(
    tensors, specs
)

For each tensor, it checks the number of extra dimensions beyond those in the spec, and reshapes tensor to have only one batch dimension. NOTE: Each tensor's batch dimensions must be the same.

Args
`tensors`	Nested list/tuple or dict of batched Tensors or SparseTensors.
`specs`	Nested list/tuple or dict of TensorSpecs, describing the shape of the non-batched Tensors.

Returns
A nested version of each tensor with a single batch dimension. A list of the batch dimensions which were flattened.

Raises
`ValueError`	if the tensors and specs have incompatible dimensions or shapes.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2024-04-26 UTC.

tf_agents.utils.nest_utils.flatten_multi_batched_nested_tensors

Args

Returns

Raises