Tag
llm
2 published notes.
Tagged llm
Full archive14 Aug 2026The MoE Frontier: Evaluating Runtime Implementations for Consumer LLMsMoE sounds great on paper, but lives or dies on the runtime. This post looks at what it actually takes to run MoE models on consumer GPUs, and where projects like Swiftlet, Colibri, and Soup currently fit in. 24 Mar 2025My Failed Journey: Running Ollama on Synology NASA tale of ambition, Docker, and hardware limitations when trying to run Ollama on a Synology DS920+