search

MIND-Lab / OCTIS

OCTIS: Comparing Topic Models is Simple! A python package to optimize and evaluate topic models (accepted at EACL2021 demo track)
MIT License
718 stars 102 forks source link

How do I handle this error #116

Open bl-sun opened 9 months ago

bl-sun commented 9 months ago

Description

This bug appeared in the program I was running. I hope someone can help me analyze the cause of this problem. I think there's something wrong with the parameters I passed in during training.

    142         score : score
    143         """
--> 144         test_labels, predicted_test_labels, self.same_svm = compute_SVM_output(model_output, self, super())
    145         return f1_score(test_labels, predicted_test_labels, average=self.average)
    146 

/home/sbl/opt/anaconda3/envs/py38-torch18/lib/python3.8/site-packages/octis/evaluation_metrics/classification_metrics.py in compute_SVM_output(model_output, metric, super_metric)
    105         test_labels, predicted_test_labels = stored_svm_results
    106     else:
--> 107         test_labels, predicted_test_labels = super_metric.score(model_output)
    108 
    109         stored_average = metric.average

/home/sbl/opt/anaconda3/envs/py38-torch18/lib/python3.8/site-packages/octis/evaluation_metrics/classification_metrics.py in score(self, model_output)
     36     def score(self, model_output):
     37         self._train_document_representations = model_output["topic-document-matrix"].T
---> 38         self._test_document_representations = model_output["test-topic-document-matrix"].T
     39 
     40         if self.use_log:

KeyError: 'test-topic-document-matrix'