issues
search
LLaVA-VL
/
LLaVA-NeXT
1.01k
stars
55
forks
source link
Fix prepare inputs labels for multimodal
#84
Open
khaimt
opened
3 days ago
khaimt
commented
3 days ago
Add assert to make sure number of images == number of image tokens in inputs
Fix the case
where num_images == 0
:
We don't need to use image_features at here
cannot set
cur_image_idx += 1
--> will run into error for many cases. For example, if batch contains 2 data points without containing images in inputs
cur_image_idx += 1
--> will run into error for many cases. For example, if batch contains 2 data points without containing images in inputs