CoIR

CoIR (Code Information Retrieval Benchmark) is the first comprehensive benchmark designed to evaluate models' abilities in code retrieval. It includes four distinct code retrieval tasks: text-to-code retrieval, code-to-text retrieval, code-to-code retrieval, and hybrid-code retrieval. The benchmark encompasses a total of 10 datasets, covering 14 programming languages and containing 2 million evaluation documents. For comprehensive details, please refer to our GitHub repository and the associated research paper