chakarapani/arxiv-collection

Arxiv Collection

For the month of february collected the latest arxiv posts data

DatasetDict({
    train: Dataset({
        features: ['id', 'updated', 'published', 'title', 'summary', 'authors', 'pdf_link', 'categories', 'text_content'],
        num_rows: total_training_rows
    })
    validation: Dataset({
        features: ['id', 'updated', 'published', 'title', 'summary', 'authors', 'pdf_link', 'categories', 'text_content'],
        num_rows: total_validation_rows
    })
    test: Dataset({
        features: ['id', 'updated', 'published', 'title', 'summary', 'authors', 'pdf_link', 'categories', 'text_content'],
        num_rows: total_testing_rows
    })
})

Note for REVISION:
main - Scraped the dataset with no limit
V1 - Scraped the dataset with only 130,000 context length
V2 - Scraped the dataset with only 60,000 context length
V3 - Scraped the dataset with only 50,000 context length
V4 - Scraped the dataset with only 30,000 context length
V5 - Scraped the dataset with only 25,000 context length
V6 - Scraped the dataset with only 20,000 context length

Limits on the `text_content` field rather than the `summary`