• @hoshikarakitaridia
    link
    English
    6
    edit-2
    5 days ago

    Probably because in the training dataset all racer povs have both hands on steering wheel so it doesn’t know how it looks when they are holding something.

    At least that’s my guess.

    • LastOP
      link
      fedilink
      English
      35 days ago

      Oh, yeah, I understand why. I just thought it’d look cooler if it were possible.