Ollama adds /v1/models and /v1/completions OpenAI compatible APIs

- https://github.com/ollama/ollama/pull/5209 - https://github.com/ollama/ollama/pull/2476 Both merged into main today, if you're using their pre-built binaries you'll need to wait for the next release. The models endpoint is nice as you query your models via the OpenAI compatible API and libraries, meaning that tools that only support OpenAI servers can return a list of models: `curl http://localhost:11434/v1/models|jq` ```json { "object": "list", "data": [ { { "id": "phi3-128k:Q8_0", "object": "model", "created": 1714120125, "owned_by": "library" }, { "id": "ollama/sfr-embedding-mistral:q4_k_m", "object": "model", "created": 1713239040, "owned_by": "ollama" } ] } ``` And you individual models to get some basic information: `curl http://localhost:11434/v1/models/phi3-128k:Q8_0|jq` ```json { "id": "phi3-128k:Q8_0", "object": "model", "created": 1714120125, "owned_by": "library" } ```