acidflask

A port of Andrej Karpathy's llama2.c to pure Julia. Works on Llama2 and Andrej's tinyllamas models. REPL mode allows interaction from the Julia console

How long until we can have Julia wrappers?

This question keeps me up at night, as I presume it does for my colleagues at Julia Computing. Not that switching to a fully commercial model is necessarily a bad thing, but Julia Labs, like any academic group, always has to worry about where funding will come from.

See Slide 4 for why it's important to be open source.

TL;DR: Researcher A finds things he wants improved in Magma (closed source) but can't. Researcher B tries to write improved FOSS implementation, but lost his job, likely in part by spending too much time writing said code and not doing other things like writing papers. Researcher A moves on and has a successful academic career. Moral: writing FOSS can cost you your academic job; it's safer to find something else to do.

I'll be there for moral support, if I'm allowed to crash!

And the problem with academic funding is that it's really hard to pay people to improve the documentation with those funds.

As Stein explains in the slides: because he couldn't sustainably pay people to work with him on SAGE, and there is only so much you can do alone.