Loss function e transformações do target (log1p, MSE)

[lucasmunizst]

In no-code AutoML platforms, can you choose the training objective/loss function (e.g. MSE) and apply your own target transforms like log1p, or is all preprocessing and scaling fully automatic?