Can the gpt caption/prompt process be shared?

#2
by Vigilence - opened

I would love to study the method you used with chat gpt so that I can recreate your caption methods and use them to create Lora’s for this model. I would also like to test your method using an alternative captioning model, llava 1.6.

So the prompting method used for gpt and maybe a couple examples of the images used and their proper outcomes would be great. That way I know I’m using the same method when expanding on this model.

Vigilence changed discussion title from Can the gpt caption/prompt prcoess be shared? to Can the gpt caption/prompt process be shared?

I would love to study the method you used with chat gpt so that I can recreate your caption methods and use them to create Lora’s for this model. I would also like to test your method using an alternative captioning model, llava 1.6.

So the prompting method used for gpt and maybe a couple examples of the images used and their proper outcomes would be great. That way I know I’m using the same method when expanding on this model.

see https://huggingface.co/dataautogpt3/ProteusV0.3/discussions/1

Thank you so much!

Do you have a few examples of the images you used for gpt4v and their prompt output? I would like to compare it against a prompt I am working in with llava 1.6

Sign up or log in to comment