Intro to Ollama workshop at GoLab Florence 2025, 2025-10-05, Martin Czygan
We have Slides and examples, which are mostly self contained.
Ollama is a popular tool to run LLM (and multimodal and embedding models) on your own machine. As of 09/2025 it has over 150K stars on GitHub.
In 2025, there are numerous other tools to run models locally:
- llama.cpp (wrapped by ollama)
- llamafile (single file distribution)
- vLLM (used by various cloud services, #1 for prod)
- Docker Model Runner (cf. How OCI Artifacts Will Drive Future AI Use Cases)
- OpenLLM
- Lemonade (AMD)
- Jan.ai
- yzma (lightweight, ffi, Go)
- ggml (used by ollama)
- and more ...
Even more user interfaces of various kinds exist.
As of 09/2025, of the 25809 repositories on GitHub tagged [llm], ollama seems to be among the top ten.
Note: ollama is both open source (MIT licensed) and VC funded.
- please install ollama on your laptop (ok, if it only has a cpu)
After installation, please run the following commands to download a few models files onto your laptop (order of preference):
ollama pull embeddinggemma
ollama pull llama3.2
ollama pull gemma3:270m
ollama pull gemma3
ollama pull qwen2.5vl
Warning: These file models may occupy over 10GB of disk space.