NeptunePossibleLegacyUsageException when logging accuracy

[talhaanwarch]

I want to log accuracy to neptune.ai logger. I tried this way

def test_epoch_end(self, outputs):
        pred=torch.cat([x["pred"] for x in outputs]).detach().cpu().numpy()
        label=torch.cat([x["label"] for x in outputs]).detach().cpu().numpy()
        pred=np.argmax(pred,1)
        acc=accuracy_score(label,pred)
        print('acc',acc)
        self.logger.experiment.log_metric("acc", acc)

and i am getting this error

---------------------------------------------------------------------------
NeptunePossibleLegacyUsageException       Traceback (most recent call last)
<ipython-input-15-a549be9bb6fe> in <module>
----> 1 trainer.test(model)

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in test(self, model, test_dataloaders, ckpt_path, verbose, datamodule)
    577 
    578         # run test
--> 579         results = self._run(model)
    580 
    581         assert self.state.stopped

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in _run(self, model)
    754 
    755         # dispatch `start_training` or `start_evaluating` or `start_predicting`
--> 756         self.dispatch()
    757 
    758         # plugin will finalized fitting (e.g. ddp_spawn will load trained model)

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in dispatch(self)
    791     def dispatch(self):
    792         if self.evaluating:
--> 793             self.accelerator.start_evaluating(self)
    794         elif self.predicting:
    795             self.accelerator.start_predicting(self)

C:\anaconda3\lib\site-packages\pytorch_lightning\accelerators\accelerator.py in start_evaluating(self, trainer)
     97 
     98     def start_evaluating(self, trainer: 'pl.Trainer') -> None:
---> 99         self.training_type_plugin.start_evaluating(trainer)
    100 
    101     def start_predicting(self, trainer: 'pl.Trainer') -> None:

C:\anaconda3\lib\site-packages\pytorch_lightning\plugins\training_type\training_type_plugin.py in start_evaluating(self, trainer)
    146     def start_evaluating(self, trainer: 'pl.Trainer') -> None:
    147         # double dispatch to initiate the test loop
--> 148         self._results = trainer.run_stage()
    149 
    150     def start_predicting(self, trainer: 'pl.Trainer') -> None:

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in run_stage(self)
    802 
    803         if self.evaluating:
--> 804             return self.run_evaluate()
    805         if self.predicting:
    806             return self.run_predict()

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in run_evaluate(self)
   1042 
   1043         with self.profiler.profile(f"run_{self.state.stage}_evaluation"):
-> 1044             eval_loop_results = self.run_evaluation()
   1045 
   1046         # remove the tensors from the eval results

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\trainer.py in run_evaluation(self, on_epoch)
    986 
    987         # lightning module method
--> 988         self.evaluation_loop.evaluation_epoch_end(outputs)
    989 
    990         # hook

C:\anaconda3\lib\site-packages\pytorch_lightning\trainer\evaluation_loop.py in evaluation_epoch_end(self, outputs)
    206             if is_overridden('test_epoch_end', model=model):
    207                 model._current_fx_name = 'test_epoch_end'
--> 208                 model.test_epoch_end(outputs)
    209 
    210         else:

<ipython-input-9-b08c6aa77d3f> in test_epoch_end(self, outputs)
    104         acc=accuracy_score(label,pred)
    105         print('acc',acc)
--> 106         self.logger.experiment.log_metric("acc", acc)

C:\anaconda3\lib\site-packages\neptune\new\run.py in __getattr__(self, item)
    161     def __getattr__(self, item):
    162         if item in LEGACY_METHODS:
--> 163             raise NeptunePossibleLegacyUsageException()
    164         raise AttributeError(f"'{self.__class__.__name__}' object has no attribute '{item}'")
    165 

[tchaton]

Dear @9157,

Netpune is currently in the process of switching their client and I believe you are using their latest version while the currently Logger within Lightning users the old client.
Please, downgrade your netpune version and it should work !

Note: Netpune is currently in the process of updating the logger within Lightning.

Best,
T.C

[talhaanwarch]

any idea which version should i use @tchaton

[tchaton]

If you use Lightning master, the new Netpune Logger has been merged. You could use their latest version.