Summary of Building Pre-train Llm Dataset For the Indic Languages: a Case Study on Hindi, by Shantipriya Parida and Shakshi Panwar and Kusum Lata and Sanskruti Mishra and Sambit Sekhar
Building pre-train LLM Dataset for the INDIC Languages: a case study on Hindi
by Shantipriya Parida, Shakshi Panwar, Kusum Lata, Sanskruti Mishra, Sambit Sekhar
First submitted to arxiv on: 13 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes a large pre-training dataset in Hindi for building foundation Large Language Models (LLMs) for the Indic language. The dataset, comprising 1.28 billion Hindi tokens, spans several domains and dialects, making it suitable for pre-training LLMs. The authors outline their pipeline for data collection, preprocessing, and availability for LLM pre-training, with potential extensions to other Indic and low-resource languages. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new dataset is being proposed in Hindi, a major Indian language. This dataset will help make Large Language Models (LLMs) better at understanding and responding to human requests. The challenge has been that there isn’t enough good data for building foundation LLMs in Hindi or other Indic languages. To solve this problem, the authors have collected 1.28 billion Hindi words from different sources and domains. This dataset can be used to train LLMs and will be available for free. |
