@ppatel Oh I understand that, I am just testing their limitations and seeing how well they maintainers respond to feedback, especially of the negative type. Besides, arguing with BME is much cheaper than therapy! (grin)
these are the tools that are available, so I believe it is in our best interest to test them & push them to their limits. It also is a reminder to those who would put the entire burden of image description on such automated processes of their fallibility, prone as they are to hallucinations & heir as they are to all the shortcomings/limitations/prejudices of the underlying LLM
Particularly frustrating is that with this app, one should be able to submit the output of BME's description app to BME's large pool of volunteers in order to get a reality/veracity check.