MisconfigurationException in 2nd Tutorial (Classification)

[mrpositron]

I am going over 2nd Torchdyn Tutorial and I encountered this error:

This error happens when I run this command:

I think that this error is caused by PytorchLighting. However, I am a newbie to PytorchLighting. Thus, I don't know how to debug it

[Zymrael]

Hi @MrPositron, thank you for pointing this out. Could you let me know which version of torch you're currently using?

[mrpositron]

PyTorch version is 1.6. I ran it on colab. Furthermore, changing

sched = {'scheduler': torch.optim.lr_scheduler.ReduceLROnPlateau(opt),
                 'monitor': 'loss', 
                 'interval': 'step',
                 'frequency': 10  }

to

sched = {'scheduler': torch.optim.lr_scheduler.ReduceLROnPlateau(opt),
                 'monitor': 'train_loss', 
                 'interval': 'step',
                 'frequency': 10  }

does not result in error.

[Zymrael]

Make sure to have the metric you choose to monitor available to the LightningModule after each step. On my end the training loop works with both loss and train_loss (pytorch-lightning==0.9.0).

The following step method:

 def training_step(self, batch, batch_idx):
        self.iters += 1.
        x, y = batch   
        x, y = x.to(device), y.to(device)
        y_hat = self.model(x)   
        loss = nn.CrossEntropyLoss()(y_hat, y)
        epoch_progress = self.iters / self.loader_len
        acc = accuracy(y_hat, y)
        nfe = model[1].nfe ; model[1].nfe = 0
        tqdm_dict = {'train_loss': loss, 'accuracy': acc, 'NFE': nfe}
        logs = {'train_loss': loss, 'epoch': epoch_progress}
        return {'loss': loss, 'progress_bar': tqdm_dict, 'log': logs} 

Should have loss available. Could you confirm that the MisconfigurationException is raised even with the above?

[Zymrael]

The new version of pytorch-lightning has fixed the issue, closing.