cp_examples/mip_finetune/mip_model.py [207:224]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        target = batch["labels"]
        # calculate loss
        loss_val = self.loss(output, target)
        # metrics
        self.log("train_metrics/loss", loss_val)
        for i, path in enumerate(self.val_pathology_list):
            j = self.label_list.index(path)
            logits, labels = filter_nans(output[:, j], target[:, j])
            self.train_acc[i](logits, labels)
            self.log(
                f"train_metrics/accuracy_{path}",
                self.train_acc[i],
                on_step=True,
                on_epoch=False,
            )
        return loss_val

    def validation_step(self, batch, batch_idx):
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



cp_examples/sip_finetune/sip_finetune.py [169:189]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        target = batch["labels"]

        # calculate loss
        loss_val = self.loss(output, target)

        # metrics
        self.log("train_metrics/loss", loss_val)
        for i, path in enumerate(self.val_pathology_list):
            j = self.label_list.index(path)
            logits, labels = filter_nans(output[:, j], target[:, j])
            self.train_acc[i](logits, labels)
            self.log(
                f"train_metrics/accuracy_{path}",
                self.train_acc[i],
                on_step=True,
                on_epoch=False,
            )

        return loss_val

    def validation_step(self, batch, batch_idx):
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -