torch.Tensor.bernoulli_ — PyTorch 2.7 documentation (original) (raw)

Tensor.bernoulli_(p=0.5, *, generator=None) → Tensor¶

Fills each location of self with an independent sample fromBernoulli(p)\text{Bernoulli}(\texttt{p}). self can have integraldtype.

p should either be a scalar or tensor containing probabilities to be used for drawing the binary random number.

If it is a tensor, the ith\text{i}^{th} element of self tensor will be set to a value sampled fromBernoulli(p_tensor[i])\text{Bernoulli}(\texttt{p\_tensor[i]}). In this case p must have floating point dtype.

See also bernoulli() and torch.bernoulli()