torch.distributions.transforms.ExpTransform

class ExpTransform extends Transform

new ExpTransform(options?: TransformOptions)

readonlydomain(unknown)
readonlycodomain(unknown)
readonlybijective(unknown)
readonlysign(unknown)

Exponential transform: y = exp(x) (maps ℝ → (0, ∞)).

Transform via the mapping y = exp(x). This transform maps real numbers to positive numbers. It is commonly used to transform Normal distributions to LogNormal.

\begin{aligned} \text{Forward: } y = e^x \\ \text{Inverse: } x = \ln(y) \\ \text{Log-det Jacobian: } \log\left|\frac{dy}{dx}\right| = x \\ \text{Property: } y = e^x \Rightarrow \log|\det J| = \log(y) \\ \text{Application: } \text{Normal} \xrightarrow{\exp} \text{LogNormal} \end{aligned}

torch.distributions.transforms.ExpTransform

readonlydomain(unknown)

readonlycodomain(unknown)

readonlybijective(unknown)

readonlysign(unknown)

Exponential transform: y = exp(x) (maps ℝ → (0, ∞)).

Transform via the mapping y = exp(x). This transform maps real numbers to positive numbers. It is commonly used to transform Normal distributions to LogNormal.

\begin{aligned} \text{Forward: } y = e^x \\ \text{Inverse: } x = \ln(y) \\ \text{Log-det Jacobian: } \log\left|\frac{dy}{dx}\right| = x \\ \text{Property: } y = e^x \Rightarrow \log|\det J| = \log(y) \\ \text{Application: } \text{Normal} \xrightarrow{\exp} \text{LogNormal} \end{aligned}