reference/generated/tensordict.TensorDictBase.rsub

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

TensorDictBase.rsub

TensorDictBase.rsub(other: TensorDictBase | Tensor | float, *, alpha: float | None = None, default: str | Tensor | TensorCollection | None = None) → Self

Subtracts self, scaled by alpha, from other, leaf by leaf.

\[\text{out}_i = \text{other}_i - \text{alpha} \times \text{self}_i\]

Supports broadcasting, type promotion, and integer, float, and complex inputs.

Parameters:

other (TensorDict, Tensor or Number) – the value from which self is subtracted. The dimensions of a tensor match the leading dimensions of each leaf, and the tensor is expanded over the other dimensions.

Keyword Arguments:

• alpha (Number) – the multiplier for self.

• default (Tensor or str, optional) – the value to use for an entry that only one of the two tensordicts has. If None (default), the two tensordicts must have the same keys, and a missing key raises KeyError. If "intersection", only the keys that both tensordicts have are computed, and the other keys are dropped. Any other value replaces the missing entries on both sides of the operation.

Examples

>>> import torch
>>> from tensordict import TensorDict
>>> td = TensorDict(a=torch.full((2,), 10.0), batch_size=[2])
>>> other = TensorDict(a=torch.ones(2), batch_size=[2])
>>> td.rsub(other)["a"]
tensor([-9., -9.])
>>> td.rsub(other, alpha=2)["a"]  # 1 - 2 * 10
tensor([-19., -19.])
>>> td.rsub(torch.tensor([1.0, 2.0]))["a"]
tensor([-9., -8.])

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