JPDev@programming.dev to Programmer Humor@programming.dev · 8 个月前Machine Learningprogramming.devimagemessage-square13fedilinkarrow-up1386arrow-down113
arrow-up1373arrow-down1imageMachine Learningprogramming.devJPDev@programming.dev to Programmer Humor@programming.dev · 8 个月前message-square13fedilink
minus-squaremarcos@lemmy.worldlinkfedilinkarrow-up55·8 个月前No, this is because the testing set can be derived from the training set. Overfitting alone can’t get you to 1.
minus-squareVictor@lemmy.worldlinkfedilinkarrow-up10·8 个月前So as an eli5, that’s basically that you have to “ask” it stuff it has never heard before? AI has come after my time in higher education.
minus-squaremarcos@lemmy.worldlinkfedilinkarrow-up20·8 个月前Yes. You train it on some data, and ask it about different data. Otherwise it just hard-codes the answers.
minus-squareArtVandelay@lemmy.worldlinkfedilinkEnglisharrow-up3·8 个月前Yes, it’s called a train test split, and is often 80/20 or there about
minus-squaresevenapples@lemmygrad.mllinkfedilinkarrow-up3·8 个月前It can if you don’t do a train-test split. But even if you consider the training set only, having zero loss is definitely a bad sign.
Is this overfitting?
No, this is because the testing set can be derived from the training set.
Overfitting alone can’t get you to 1.
So as an eli5, that’s basically that you have to “ask” it stuff it has never heard before? AI has come after my time in higher education.
Yes.
You train it on some data, and ask it about different data. Otherwise it just hard-codes the answers.
They’re just like us.
Gotcha, thank you!
Yes, it’s called a train test split, and is often 80/20 or there about
It can if you don’t do a train-test split.
But even if you consider the training set only, having zero loss is definitely a bad sign.
Gotcha!