3 ms·That just sounds like a 3090.by wmf 4mo agoThat just sounds like a 3090.cyanydeez 4mo agonot at the vram sizes that control how much context to load; also, GPUs arn't as effiecient as direct inference.wmf 4mo agoOK, B70.