New CIDR Dataset Offers 832M Lines of Industrial Code for AI Research

Researchers have introduced the Curated Industrial Developer Repository (CIDR), a substantial dataset containing 4,225 real-world software repositories from industrial partners. This dataset includes code in 75 programming languages, version control history, and engineering practice attributes like continuous integration usage. CIDR is designed to advance research in code intelligence and software engineering tasks, with an initial study showing its effectiveness in fine-tuning a 3-billion-parameter code language model. AI

IMPACT Provides a large-scale, real-world dataset to advance AI research in code intelligence and developer tooling.

RANK_REASON The cluster describes a new dataset released via arXiv for software engineering research. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• CIDR
• Curated Industrial Developer Repository
• DagsHub
• Gotit.pub
• Hugging Face
• ScienceCast
• Vladislav Savenkov

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →