reference/generated/tensordict.TensorDictBase.bytes

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

TensorDictBase.bytes

TensorDictBase.bytes(*, count_duplicates: bool = True) → int

Returns the number of bytes of the tensor leaves.

Each tensor leaf counts numel() * dtype.itemsize bytes, and a leaf with a gradient counts the bytes of its gradient too. Non-tensor data is not counted. A view counts the bytes of its own elements, not those of the storage it shares.

Keyword Arguments:

count_duplicates (bool, optional) – if True, a tensor stored under several keys is counted once per key. If False, it is counted once. Two different views of one tensor are counted twice in both cases. Defaults to True.

Examples

>>> import torch
>>> from tensordict import TensorDict
>>> x = torch.zeros(6)
>>> td = TensorDict(a=x, b=x, c=torch.zeros(2, 3))
>>> td.bytes()
72
>>> td.bytes(count_duplicates=False)
48

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