The library already has tinygrad support it seems, so it's not limited to Apple & MLX

That is true. However (as of two days ago, it may have rapidly changed since then) the python program did not differentiate based on your architecture and would try to import mlx regardless if it's installable on your system or not, causing import errors.

This is fixed now, with these commits: - https://github.com/exo-explore/exo/commit/dbbc7be57fb1871d2b... - https://github.com/exo-explore/exo/commit/ce46f000591d8d59c1...

Please keep the bug reports coming, we're moving fast to get this stable on all platforms.

You can use other models like a vision LLM, or use AI agents as well

The embedding size is only 8k so while the parameters are 70B. So it's a huge difference

This is the most relatable comment in this thread. I had the same experience when I was using ros.org and moved to a new distro

Correct, which one of the main moats Nvidia has when it comes to training

Potato/ Jerusalem artichoke ? (“Erdapfel”) is my fav