I installed your docker version and really like.
I reviewed your roadmap and what I would looking forward to is the option to use different models for label recognition and chat task (such as Sommelier chat).
Many cheap models are very good at vision capabilities, however those lack the reasoning capabilities to provide a detailed and differentiated advice for example for food pairing.
To be as cost and resource-efficient as possible adding a model for AI label scanning and another model for sommelier would be awesome.
I installed your docker version and really like.
I reviewed your roadmap and what I would looking forward to is the option to use different models for label recognition and chat task (such as Sommelier chat).
Many cheap models are very good at vision capabilities, however those lack the reasoning capabilities to provide a detailed and differentiated advice for example for food pairing.
To be as cost and resource-efficient as possible adding a model for AI label scanning and another model for sommelier would be awesome.