dataset_stats.py

from statistics import mean
import pandas as pd
from datasets import load_dataset


def count_word(text):
    return len(text.split())


if __name__ == '__main__':
    data = ["chemprot", "citation_intent", "hyperpartisan_news", "rct_sample", "sciie", "amcd", 'yelp_review',
            'tweet_eval_irony', 'tweet_eval_hate', 'tweet_eval_emotion']
    stats = {}
    for d in data:
        _data = load_dataset('asahi417/multi_domain_document_classification', d)
        stats[d] = {
            'word/validation': mean([count_word(k['text']) for k in _data['validation']]),
            'word/test': mean([count_word(k['text']) for k in _data['test']]),
            'word/train': mean([count_word(k['text']) for k in _data['train']]),
            'instance/validation': len(_data['validation']),
            'instance/test': len(_data['test']),
            'instance/train': len(_data['train'])
        }
    df = pd.DataFrame(stats).astype(int)
    df.to_csv('stats.csv')
    print(df.to_markdown())