[Nvidia] NemoClaw – Enterprise AI Agents, Redefinednemoclaw.bot4 points·by weirdcat·il y a 4 mois·0 comments
Computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haikuanthropic.com1,454 points·by weirdcat·il y a 2 ans·735 comments
Training Language Models to Self-Correct via Reinforcement Learningarxiv.org230 points·by weirdcat·il y a 2 ans·92 comments
Linear, symmetric, self-selecting 14-bit molecular memristors (2023)researchgate.net109 points·by weirdcat·il y a 2 ans·55 comments
weirdcat·le mois dernier·discussHow 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).