Trained on 29 images; captions generated by ChatGPT (much more detailed using ecological terms)
This model doesn't have a readme.