Видео с ютуба Moe Cache

llama.cpp PR #30637: SLRU Cache for MoE Models

Внутри llama.cpp: как запускать огромные MoE-модели (новый кэш экспертов)

llama.cpp PR #29887: GPU Cache for MoE Experts

Worst clutch in the history of CS by mOE

The LRU Cache That Doubled Local MoE Speeds: llama.cpp Expert Offloading

llama.cpp Just Made 125B MoE Models 4× Faster!

llama.cpp's New MoE Expert Cache: Up To 2.2x Faster on One GPU

Run 125B Models on 8GB Ram: llama.cpp Expert Cache

Janas: Run MoE Models Larger Than RAM from Your SSD?

Your MoE Model Is 2× Slower Without This llama.cpp Flag

The MoE Trap: Why 3B Active Doesn't Mean 3GB of VRAM (Qwen 35B)

llama.cpp's MoE Cache: 3x On 16GB. Off By Default.

llama.cpp Just Got an Expert Cache. Do You Still Need Strata?

Визуальное руководство по составу экспертов в программах магистратуры по праву.

How KV Cache Speeds Up LLMs for Faster AI Models on GPUs

В llama.cpp добавили Expert Cache... Победит ли он Strata на 12 ГБ?!

Broken in 12 Days: How a 1993 Idea Fixed llama.cpp MoE Caching

llama.cpp 0.7: 2x Faster Local AI on One GPU

llama.cpp runs MoE models up to 2.2x faster when VRAM runs out

DeepSeek-V4.1-Flash Explained: 552B MoE, Sub-Kilobyte KV Cache & Massive Agentic Benchmarks!