Module dfdx::losses

Functions

cross_entropy_with_logits_loss

Cross entropy loss. This will call logits.log_softmax(), so make sure logits is not the output from softmax() or log_softmax() already.

mae_loss

Mean absolute error. This is the same as doing (pred - &targ).abs().mean()

mse_loss

Mean Squared Error. This is the same as doing (pred - &targ).abs().mean()