reference/generated/tensordict.TensorDictBase.to_pytree

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

TensorDictBase.to_pytree

TensorDictBase.to_pytree()

Rebuilds the pytree that from_pytree() converted to this tensordict.

If the tensordict has no "_pytree_type" entry, that is, if it was not created by from_pytree(), this method returns self. A nested tensordict without that entry is kept as a tensordict in the pytree.

See also

from_pytree() for the supported types.

Examples

>>> import torch
>>> from tensordict import TensorDict
>>> td = TensorDict.from_pytree({"a": [torch.zeros(2), torch.ones(2)]})
>>> td.to_pytree()
{'a': [tensor([0., 0.]), tensor([1., 1.])]}

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