Nilang.jl – A Reversible Julia DSL

Incredible. Given a (one-to-one) Julia function `f`, this package generates `~f`, the inverse of f. Clever use of automatic differentiation.<p>Paper: <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2003.04617" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2003.04617</a>

,,The performance of reversible programming automatic differentiation is much better than most traditional frameworks.&#x27;&#x27;<p>It would be good to see a ResNet training benchmark comparison with PyTorch as an example if this is really true.

This is the latest NiLang tutorial notebook: <a href="https:&#x2F;&#x2F;github.com&#x2F;JuliaReverse&#x2F;NiLangTutorial" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;JuliaReverse&#x2F;NiLangTutorial</a>

Wait a second, does it yield an <i>exact</i> inverse or a numerical approximation?

How does this compare to zygote ?