weirdcat

How are you running that GGUF, and how many tokens/sec are you getting without MTP? My M1 Max gives me 65 t/s for non-MTP unsloth/gemma-4-26B-A4B-it-qat-GGUF (UD-Q4_K_XL), but with MTP that actually goes down to 56 t/s (at 63% accepted drafts).

I had the exact same experience, only with one added bonus thought: "OK, so I don't see any noise. Middle of the image... Maybe they consider this weird shape to be noise for some reason? Hmmm, what is it anyway?" :)