I uploaded the Windows XP desktop wallpaper into the image classifier. Just the raw image file. It gave me the labels "monitor", "computer screen", "desktop". "Field", "sky", grass", that kind of thing were nowhere to be found.
I know this is more of a comment on the state of AI models than Transformers.js. It's probably not even representative of state-of-the-art image classifier models. Just a fun example of how these things learn.
Haha very interesting! I assume it's because that type of image is only found on computer screens, so, the model thinks the grass "contributes to it's idea of what a computer screen is".
... and of course, the library only ports those models to the browser; if you train a better model, you can always convert it to the ONNX format, then use it with the library.
I know this is more of a comment on the state of AI models than Transformers.js. It's probably not even representative of state-of-the-art image classifier models. Just a fun example of how these things learn.