Post #3229235
2026-06-01 17:38 UTC
16GB is plenty for even older model setups. Now they've got a few models designed so you load just parts of the model onto the GPU (Mixture of Experts) and use the CPU for less referenced sections, so you get both reasonable speed and a much more complex model.
Replies (1)
-
@onlinepersona@programming.dev 2026-06-02 05:13
Oh nice. Does that depend on just the model or are there other requirements like CUDA or something?