It should be more aware of its own capabilities.
The model keeps telling me its capable of doing text-to-image, stating that theres a feature where there actually isnt one.
29
votes
Melvin
shared this idea