Record 007 · Runtimes · 2023
50%23.07.U
Ollama
Ollama
“The one-liner that made `ollama run llama3` the way developers start local AI.”
Ollama wrapped llama.cpp in a brew-installable daemon, a model library, and a REST API. It is how local models left the makefile.
Made local LLMs a default developer tool.
- Model library + Modelfile
- Local HTTP API
- Desktop + server daemon
Docker for models.
