RReLU — PyTorch 2.7 documentation (original) (raw)

class torch.nn.RReLU(lower=0.125, upper=0.3333333333333333, inplace=False)[source][source]¶

Applies the randomized leaky rectified linear unit function, element-wise.

Method described in the paper:Empirical Evaluation of Rectified Activations in Convolutional Network.

The function is defined as:

RReLU(x)={xif x≥0ax otherwise \text{RReLU}(x) = \begin{cases} x & \text{if } x \geq 0 \\ ax & \text{ otherwise } \end{cases}

where aa is randomly sampled from uniform distributionU(lower,upper)\mathcal{U}(\text{lower}, \text{upper}) during training while during evaluation aa is fixed with a=lower+upper2a = \frac{\text{lower} + \text{upper}}{2}.

Parameters

• lower (float) – lower bound of the uniform distribution. Default: 18\frac{1}{8}
• upper (float) – upper bound of the uniform distribution. Default: 13\frac{1}{3}
• inplace (bool) – can optionally do the operation in-place. Default: False

Shape:

• Input: (∗)(*), where ∗* means any number of dimensions.
• Output: (∗)(*), same shape as the input.

Examples:

m = nn.RReLU(0.1, 0.3) input = torch.randn(2) output = m(input)