Hugging Face's logo

Spaces:
csuhan
/
LLaMA-Adapter
Runtime error

App Files Community
2

How to do VQA ?

#2
by Jetcombo - opened
Discussion
Jetcombo
•
edited May 3, 2023

How do I provide an image as a context input?
I also followed your code here: https://github.com/ZrrSkywalker/LLaMA-Adapter/blob/main/example.py
How do I provide an image object / image path in prompt_input ?

Jetcombo

I could do it in the hugging face App by just pasting the image url inside context box. Hopefully the image path would work in the code (example.py)

csuhan
 Owner

Hi @Jetcombo , we do not release the training code of VQA / image captioning with LLaMA-Adapter. But you can refer the inference code of coco captioning at here.

Sign up or log in to comment