Question 1

Can the AMD Ryzen AI Max (Strix Halo) run llava 7B?

Accepted Answer

llava 7B needs about 6.7GB working set (weights 4.7GB + KV cache 2GB at 4k context). The AMD Ryzen AI Max (Strix Halo) offers ~30GB usable — verdict: full. 6.7 GB needed of 30 GB usable — headroom for context

Question 2

How fast is llava 7B on the AMD Ryzen AI Max (Strix Halo)?

Accepted Answer

~33–46 tok/s on the AMD Ryzen AI Max (Strix Halo). computed roofline: 256 GB/s × 0.60–0.85 efficiency window / 4.7GB (Q4_K_M) — bands, never points.

Question 3

What should I run llava 7B on instead?

Accepted Answer

Tesla M60 (16GB, used) runs it fully ($35).

Can the AMD Ryzen AI Max (Strix Halo) run llava 7B?

Run llava 7B on instead