reference/generated/tensordict.TensorDictBase.rename

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

TensorDictBase.rename

TensorDictBase.rename(*names, **rename_map)

Returns a shallow copy of the tensordict with new names for its batch dimensions.

Pass either the new names, one per batch dimension, or keyword arguments that map old names to new ones. The leaves are not copied (see copy()), and the names of the tensordict do not change. The nested tensordicts of the copy get the new names for the batch dimensions they share with it.

Parameters:

• *names (str or None) – the new names, one per batch dimension, or a single None to clear every name. Exclusive with **rename_map.

• **rename_map (str or None) – new names by old name, as in rename(old="new"). Exclusive with *names.

Raises:

ValueError – if both or neither of names and rename_map are passed, or if a key of rename_map is not a name of the tensordict.

Examples

>>> from tensordict import TensorDict
>>> td = TensorDict(batch_size=[1, 2, 3, 4], names=list("abcd"))
>>> td.rename(c="g").names
['a', 'b', 'g', 'd']
>>> td.rename("w", "x", "y", "z").names
['w', 'x', 'y', 'z']
>>> td.rename(None).names
[None, None, None, None]
>>> td.names
['a', 'b', 'c', 'd']

Every tensordict class and every tensorclass has this method. For the other methods, see Methods by task.