Dive into deep learning compilers

Re:Dive into deep learning compilers by 37ef_ced3 on Thursday February 24, 2022 @12:16PM (#30457405)

Related: https://NN-512.com is a compiler that generates C99 code for neural net inference. It takes as input a simple text description of a convolutional neural net inference graph. It produces as output a human-readable, stand-alone (no dependencies) C99 implementation of that graph. The generated C99 code uses AVX-512 vector instructions to perform inference.

[ Parent | on HN ]

Re:Dive into deep learning compilers by jhgb on Saturday March 05, 2022 @09:52PM (#30574574)

That's cool for anyone who has AVX-512 enabled CPUs. Is there anything that works on CPUs supporting only up to AVX2?

[ Parent | on HN ]

Re:Dive into deep learning compilers by theK on Friday February 25, 2022 @12:02AM (#30463882)

Tangential question, why not just host a git repo instead of hosting the files without history?

[ Parent | on HN ]

Re:Dive into deep learning compilers by 37ef_ced3 on Friday February 25, 2022 @12:10AM (#30463911)

The compiler generates and serves its own website with no dependencies.

For personal projects (including NN-512) I use Mercurial for version control, not git. But in any case I don't want the compiler (or the website it generates) to depend on git or any other version control software.

[ Parent | on HN ]

Re:Dive into deep learning compilers by smaudet on Friday February 25, 2022 @03:19AM (#30464853)

What do you do for version control?

I think you have this backwards, anybody could check in the contents of your tarball into a vcs, so not sure the point...

[ Parent | on HN ]