> What is the shape of this object?
The shape of the object in the image is an ellipse or an oval. It is elongated horizontally and has a rounded top and bottom.
What the model actually sees is this:
...
> What is the shape of this object?
The image features a yellow circular shape with a simple facial expression.
Now the model only(?) accepts 1344x1344 and there is no perfect solution. But I think there should either be a (textual or visual) hint or an option like
(*) stretch
( ) center crop
( ) add borders
I just noticed this piece of code in
cogvlm2.py
:What the model actually sees is this:
...
Now the model only(?) accepts 1344x1344 and there is no perfect solution. But I think there should either be a (textual or visual) hint or an option like (*) stretch ( ) center crop ( ) add borders