# Why isn't leaky ReLU always preferable to ReLU given the zero gradient for x<0?

It looks to me like the leaky ReLU should have much better performance since the standard ReLU can’t use half of its space (x < 0 where the gradient is zero). But this doesn't happen and in practice most people use standard ReLU.

• Your question confuses me, what exactly do you want to know? Also can you trim down the heading and add more in the content. What do you mean by 'all' are using? Do you mean generic entreprises? You mention that leaky ReLU should perform better in the question content but contradict yourself in the heading. – Hima Varsha Jan 31 '17 at 12:07

Also, leaky RelU (if parametric) introduces another parameter (the slope for $x<0$) that needs to be learned during training and therefore adds more complexity/training time.