requirements file

[weizhou1991]

Hi there,

I just noticed that if I want follow your pipe from generate base to generate caption, it needs employ different projects, each projects rely on its own environment, do you have a universal requirements file to help me to build the environment ? I really appreciate your help !

Best,

[Mayakshanesht]

I think, you could run the code sequentially. For example, you could run the generate_base.py by activating the environment for the stp3 and commenting out llava and minigpt modules imports, to generate the object info json and cropper images json file.
Then you could turn on the llava environment, and run the generate_caption.py file to generate the captions for the objects and scenes.
For the click2chat, we need both and I dont know how to resolve that issue.