3 ms·
Or maybe models that are much more task-focused? Like models that are trained on just math & coding?
by UncleOxidant 11mo ago
Or maybe models that are much more task-focused? Like models that are trained on just math & coding?
- agoodusername63 11mo agoisn't that what the mixture of experts trick that all the big players do is? Bunch of smaller, tightly focused models
- irthomasthomas 11mo agoNot exactly. MoE uses a router model to select a subset of layers per token. This makes them faster but still requires the same amount of RAM.