Fetching the paper…

SparAMX: Accelerating Compressed LLMs Token Generation on AMX-powered CPUs · Around