Skip to main content

2 posts tagged with "inference"

View All Tags

LLMs on Bare Metal: Quantization, SLMs, and Replacing phi4-mini with Qwen 2.5 7B

· 9 min read
Software Engineer & Cloud Architect

Running LLMs on bare-metal CPU hardware forces you to understand the numbers behind model files. This post documents how we reason about model size, quantization, and inference serving on minicloud — and the concrete change we made: replacing phi4-mini with Qwen 2.5 7B across all three Ollama instances.