Humaneval

HumanEval is an iconic open-source code evaluation dataset repository originally released by OpenAI on GitHub. It features 164 hand-written python programming tasks designed to measure and score the functional correctness of generative AI code synthesis models.