Early-Signal Detector

method · first seen 6 Oct 2026

expert offloading

Also: MoE offloading, weight offloading, CPU-GPU offloading, MoE expert offloading

Storing inactive expert weights in CPU memory or streaming them to GPU on demand to run large MoE models on limited VRAM.

Mentions per week

0419 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Rank 180 on 9 Nov 2026 · 12 mentions in 28 days · 12 independent groups · code → papers · first seen 5 weeks ago · acceleration +0.77

Earliest sources

Latest sources

Mentioned together with

mixture of expertsspeculative decodingself-hosted LLM deploymentquantizationllama.cppsingle-GPU inferenceQwen3GGUFCPU-GPU collaborative inferenceexpert pruning