Domain-specific adaptation of CodeT5 for Python docstring generation. Compares pre-trained vs fine-tuned performance using NLP metrics (ROUGE, METEOR, BERTScore), showing significant improvements after targeted training.
nlp machine-learning deep-learning transformers code-generation rouge fine-tuning code-summarization bertscore code-comments-generator codet5
-
Updated
Sep 22, 2025 - Jupyter Notebook