Add support for remote ollama hosts

[Fried-Squid]

Duplicates

• I have searched the existing issues

Summary 💡

The LLM call blocks should support an additional input for an ollama host which is not the localhost.

Will self-assign and PR later this week.

Examples 🌈

No response

Motivation 🔦

Allowing remote ollama servers would allow self-hosted users to run autogpt+ollama across multiple hosts which may be more suited to tasks, e.g. running ollama on a jetson board.

[Fried-Squid]

PR #8234