Skip to content
local-ai

Running Models

Quantisation, inference engines, model formats, and getting good performance.

  1. Intermediate

    Inference engines explained

    Ollama, llama.cpp, vLLM, MLX, LM Studio. They all run models, but they are not the same thing. Here is what an inference engine does and how to pick one.

  2. Beginner

    LM Studio vs Ollama: which should you use?

    Both run models on your own machine, but they are built for different first steps: LM Studio is a graphical app, Ollama is a command line and a local API. Here is what separates them, and why you might use both.

  3. Intermediate

    Understanding quantisation

    Quantisation is what makes local AI possible on consumer hardware. What the formats mean, what you give up, and how to choose one for your memory budget.

  4. Beginner

    Ollama commands: a practical reference

    The Ollama commands you actually reach for: listing, running, and removing models, finding and changing where they are stored, the default port and how to change it, updating, and uninstalling cleanly on macOS, Windows, and Linux.

  5. Intermediate

    llama.cpp vs Ollama vs vLLM: which engine?

    Ollama is built on llama.cpp, and vLLM is a different kind of tool altogether. Here is what you gain by using llama.cpp directly, when vLLM is the right answer, and when it is overkill.