MA-tokenweights/all-the-news-2-tfidf-invfreq-topic-stratified-v1-articles
收藏资源简介:
--- dataset_info: features: - name: article_id dtype: int64 - name: source_split dtype: string - name: source_line_start dtype: int64 - name: source_line_end dtype: int64 - name: source_line_count dtype: int32 - name: article_word_count dtype: int32 - name: words list: string - name: tf_idf_scores list: float32 - name: token_ids list: int32 - name: token_weights list: float32 - name: article_token_count dtype: int32 - name: article_text dtype: string - name: meta_idx dtype: string - name: meta_article_idx dtype: string - name: meta_title dtype: string - name: meta_publication dtype: string - name: meta_date dtype: string - name: meta_url dtype: string - name: meta_section dtype: string splits: - name: train_articles num_bytes: 2587829773 num_examples: 162393 - name: validation_articles num_bytes: 40515683 num_examples: 987 download_size: 1384843986 dataset_size: 2628345456 configs: - config_name: default data_files: - split: train_articles path: data/train_articles-* - split: validation_articles path: data/validation_articles-* ---
This is a news article dataset containing article IDs, source splits, start and end line numbers, line counts, word counts, word lists, TF-IDF scores, token IDs, token weights, token counts, full article texts, and metadata (such as index, article index, title, publication, date, URL, and section). The data is divided into training articles and validation articles, suitable for natural language processing tasks like text analysis, keyword extraction, and information retrieval.




