Marathi Non-STEM Textbook Corpus with 37 Million Words

37.36 million words from 584 Marathi textbooks form this large-scale educational corpus. InfoBayAI created this dataset to support the development of advanced NLP systems for Marathi language understanding and reasoning. The dataset was last updated on June 8, 2026.

Use Cases

• Train language models for Marathi based on the large-scale textbook corpus.
• Develop educational AI assistants using the dataset's general knowledge content.
• Benchmark reasoning and comprehension models on non-STEM Marathi text.
• Support multilingual NLP research by incorporating this Marathi educational data.

Strengths

• Contains 37.36 million words, providing substantial textual data for model training.
• Comprises 584 distinct books, suggesting diversity in source material.
• Part of a larger multilingual corpus of over 2.6 billion words, indicating a broader context.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The description mentions interwoven images, but the data type is listed as text, suggesting potential missing multimodal components.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

InfoBayAI via Hugging Face.

Collection Method

Likely collected and compiled from Marathi non-STEM educational textbooks.

Freshness

Last updated 2026-06-08 07:51:18; freshness should be verified.