paper-with-me

ReLUN

Rectified Linear Unit N

2000년 도입 · 논문 1편에서 사용

The Rectified Linear Unit N, or ReLUN, is a modification of ReLU6 activation function that has trainable parameter n. $$ReLUN(x) = min(max(0, x), n)$$

출처: Trainable Activations for Image Classification

소개 논문: Trainable Activations for Image Classification

Adaptive Activation Functions · GeneralActivation Functions · General