0.11.0

What's Changed

• Multi-agent system for dealing with vision tasks by @haesleinhuepf in #115
• Support for llava (vision model)

Backwards compatibility breaking changes

• when using vision models, one must use a new parameter vision_model for specifying which model to use. E.g.:

bob.initialize(model='gpt-4-0125-preview', vision_model='gpt-4-vision-preview')

or

bob.initialize(model='gemini-pro', vision_model='gemini-pro-vision')

Full Changelog: 0.10.2...0.11.0