• @[email protected]
    link
    fedilink
    English
    40
    edit-2
    3 months ago

    If you are inferring 32 pixels from 1 pixel that is because the model has been trained on billions of computed pixels. You cannot infer data in vacuum. The statement is bullshit.