A dataset for training models to generate descriptive commit messages from git diffs. Contains 13,811 examples with git diffs and corresponding commit messages in chat format suitable for instruction tuning.

Dataset Statistics

Number of examples: 13,811 Average text length: 960 characters

Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("hks350d/commit-message-generation")

Example Usage

Load the dataset

dataset =… See the full description on the dataset page: https://huggingface.co/datasets/hks350d/commit-message-generation.