[Other] Add support for Ollama keep_alive parameter.

[loveritsu929re]

By default the ollama models are kept in memory for 5 minutes before being unloaded. Would be helpful if I can set this in model setting in chatbox.

Just need to add it to the request, e.g. curl http://localhost:11434/api/generate -d '{"model": "llama2", "keep_alive": -1}'