Execution-based Evaluation for Data Science Code Generation Models

Junjie Huang, Chenglong Wang, Jipeng Zhang +6

Code generation models can benefit data scientists' productivity by automatically generating code from context and text descriptions. An important measure of the modeling progress…