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

Tensor.exponential_(lambd=1, *, generator=None) → Tensor¶

Fills self tensor with elements drawn from the PDF (probability density function):

f(x)=λe−λx,x>0f(x) = \lambda e^{-\lambda x}, x > 0

Note

In probability theory, exponential distribution is supported on interval [0, inf⁡\inf) (i.e., x>=0x >= 0) implying that zero can be sampled from the exponential distribution. However, torch.Tensor.exponential_() does not sample zero, which means that its actual support is the interval (0, inf⁡\inf).

Note that torch.distributions.exponential.Exponential() is supported on the interval [0, inf⁡\inf) and can sample zero.