ag_news: class_names: - World - Sports - Business - Technology description: News categorization with 4 classes, known for similar content across categories name: AG News Classification num_classes: 4 original_test_samples: 7600 original_train_samples: 120000 quality_issues: - redundancy - similar_content - topic_overlap target_column: label task_type: multi_classification test_samples: 7600 text_columns: - text total_samples: 127600 train_samples: 90000 validation_samples: 30000 amazon_polarity: class_names: - negative - positive description: Amazon reviews with noisy sentiment labels name: Amazon Product Reviews num_classes: 2 original_test_samples: 400000 original_train_samples: 3600000 quality_issues: - label_noise - rating_inconsistency target_column: label task_type: binary_classification test_samples: 400000 text_columns: - text total_samples: 4000000 train_samples: 2700000 validation_samples: 900000 emotion: class_names: - sadness - joy - love - anger - fear - surprise description: Twitter emotion classification with text length outliers name: Emotion Classification num_classes: 6 original_test_samples: 41681 original_train_samples: 333447 quality_issues: - length_outliers - text_anomalies target_column: label task_type: multi_classification test_samples: 41681 text_columns: - text total_samples: 375128 train_samples: 250085 validation_samples: 83362 imdb: class_names: - negative - positive description: Movie reviews with subjective sentiment labels and borderline cases name: IMDB Movie Reviews num_classes: 2 original_test_samples: 25000 original_train_samples: 25000 quality_issues: - label_noise - subjective_labels - borderline_cases target_column: label task_type: binary_classification test_samples: 25000 text_columns: - text total_samples: 50000 train_samples: 18750 validation_samples: 6250 twenty_newsgroups: class_names: - alt.atheism - comp.graphics - comp.os.ms-windows.misc - comp.sys.ibm.pc.hardware - comp.sys.mac.hardware - comp.windows.x - misc.forsale - rec.autos - rec.motorcycles - rec.sport.baseball - rec.sport.hockey - sci.crypt - sci.electronics - sci.med - sci.space - soc.religion.christian - talk.politics.guns - talk.politics.mideast - talk.politics.misc - talk.religion.misc description: Newsgroup posts with overlapping topics and cross-posting name: 20 Newsgroups num_classes: 20 original_test_samples: 7532 original_train_samples: 11314 quality_issues: - redundancy - cross_posting - similar_topics target_column: label task_type: multi_classification test_samples: 7532 text_columns: - text total_samples: 18846 train_samples: 8485 validation_samples: 2829 yelp_polarity: class_names: - negative - positive description: Yelp reviews with positive/negative sentiment, naturally imbalanced name: Yelp Review Polarity num_classes: 2 original_test_samples: 38000 original_train_samples: 560000 quality_issues: - moderate_imbalance - rating_bias target_column: label task_type: binary_classification test_samples: 38000 text_columns: - text total_samples: 598000 train_samples: 420000 validation_samples: 140000