Ollama

[Amuwa]

Add local ollama model support

Usage:

host = "localhost:11434"
key    = "sk-anystring"
model = "ollama:deepseek-r1:14b"

[RayZhhh]

Hello! Thank you for your contribution, and I am sorry for my late revision.

Would you mind leaving the original file unchanged, and creating a new file in the same directory to implement the interface for Ollama?

[RayZhhh]

Hi! I have tested your commit. It works! 👍

Please also take a look at my minor suggestion and see if it's reasonable: What about leaving a default value for the argument key and host in def __init__() ? For example:

def __init__(self, model, host='localhost:11434', key='' , timeout=60, **kwargs):
    ...