Dataset is imported from CodeXGLUE and pre-processed using their script.
Where to find in Semeru:
The dataset can be found at /nfs/semeru/semeru_datasets/code_xglue/code-to-text/python in Semeru
CodeXGLUE -- Code-To-Text
Task Definition
The task is to generate natural language comments for a code, and evaluted by smoothed bleu-4 score.
The dataset we use comes from CodeSearchNet and we filter the dataset as the following:
Remove… See the full description on the dataset page:
https://huggingface.co/datasets/semeru/code-text-python.
