Multiple Images
#52
by
RonanMcGovern
- opened
Is there a way to pass in multiple images?
Understandable if it makes most sense to focus on just one for now.
Even if it's something not going to be supported, how would you recommend doing two images? just concatenate them perhaps?
It depends on the use case, if your task is extraction related, then the best way would be to simply infer on each image separately, then if you need them aligned you can use a pretty lightweight llm to do so without the need for the vision component.