TypeError: on_validation_end() missing 1 required positional argument: 'outputs'

[dimgag]

Hello,

When I train my model, I get this error:

TypeError: on_validation_end() missing 1 required positional argument: 'outputs'

This is the function inside my model class:

def on_validation_end(self, outputs):
     loss = torch.stack([x['loss'] for x in outputs]).mean().cpu().numpy().round(2)
     metric = torch.stack([x['metric'] for x in outputs]).mean().cpu().numpy().round(2)
     self.log('val_loss', loss)
     self.log('val_metric', metric)

[AzulGarza]

hey @dimgag! I had the same problem recently. As of lightning>=2.0.0 you have to define the validation_step_outputs in the init clause of your module to store the validation outputs of each epoch. Here's an example (from this PR).

 class MyLightningModule(L.LightningModule):
    def __init__(self):
        super().__init__()
        self.validation_step_outputs = []

     def validation_step(self, ...):
         loss = ...
         self.validation_step_outputs.append(loss)
         return loss

    def on_validation_epoch_end(self):
        epoch_average = torch.stack(self.validation_step_outputs).mean()
        self.log("validation_epoch_average", epoch_average)
        self.validation_step_outputs.clear()  # free memory

[menglimengjianmengbujian]

Thank you very much, my same problem has been solved！

[kekeche]

So how do you solve it at last