Загрузка видео...
Не удалось загрузить видео
llama 4 maverick do visually look stunning.
62,407 просмотров • 1 год назад •via X (Twitter)
Комментарии: 10

losh1 год назад
a little explanation of what you are seeing here. The models when stored have a structure (not execution graph, just how parameters are grouped). the color coding represent types of param blocks and the block size is log(param_size). I did similar plots in 2d sometime back:

Viraat1 год назад
This is really cool - in a few months if you are looking for a job and are interested hmu

losh1 год назад
sure, and thanks!

Krishna Mohan1 год назад
Amazing

⛓️☆ ilex ☆⛓️1 год назад
I like how this feels in my brain

losh1 год назад
latest

ueaj1 год назад
It might look cooler if experts are separate squares, like a grid or something

Louis1 год назад
wow

deep Manifold1 год назад
Really cool

Karan Lokchandani1 год назад
im never deleting this app


