3 ms·
> if you have enough lines you can approximate any curve within a margin How avoid overfitting the training data?
by ngrilly 6y ago
> if you have enough lines you can approximate any curve within a margin
How avoid overfitting the training data?
- RandyRanderson 6y agoYou're right - a LOT of work goes into this (eg GANs) because these large nNets just start "remembering" the training data.