Open Wenbobobo opened 1 year ago
As far as I understood, the prompts defined in the 'alignment.txt' file are randomly sampled in this line here: https://github.com/Vision-CAIR/MiniGPT-4/blob/22d8888ca2cf0aac862f537e7d22ef5830036808/minigpt4/models/mini_gpt4.py#L118 So in order to use specific prompts for specific image-text pairs, you would need to change a lot there. I would be interested in using specific prompts as well..
I thibk that this may provide custom prompt, so I tried to modify it, but it was useless. Coule anyone teach me or explain it in the document? Thanks