FairyFuse: Multiplication-Free LLM Inference on CPUs via Fused Ternary Kernels
hackernews · оригинал
Заголовок: FairyFuse: Multiplication-Free LLM Inference on CPUs via Fused Ternary Kernels
hackernews · оригинал
Заголовок: FairyFuse: Multiplication-Free LLM Inference on CPUs via Fused Ternary Kernels