llama-server development

[brunette69-ruby]

How well is llama-server maintained and developed?
I ask this since I am having issues with embedding thru server.
Is it better to go pure c++ or use python bindings than relying on llama http server and api?

[bandoti]

The server is well maintained and developed—take a look at the history and you'll see it is updated very frequently!

Is it better to go pure c++ or use python bindings than relying on llama http server and api?

This is impossible for anyone to answer, as it depends wholly on what is the intended goal. I recommend coming up with some higher level objectives, and then a general architecture that fits the needs. If you are embedding llama.cpp within an application, perhaps C++ or the Python bindings. If you want to cleanly separate the application from the AI using an API, perhaps the server with OpenAI library. 🙂