<type>(<scope>): <subject>
| Type | Description | Version Bump |
|---|---|---|
feat |
New feature | Minor (0.x.0) |
fix |
Bug fix | Patch (0.0.x) |
docs |
Documentation only | None |
style |
Code style/formatting | None |
refactor |
Code refactoring | None |
perf |
Performance improvement | Patch |
test |
Adding/updating tests | None |
build |
Build system changes | None |
ci |
CI/CD changes | None |
chore |
Other changes | None |
# Feature (minor bump: 1.6.1 → 1.7.0)
git commit -m "feat(models): add TabNet architecture"
# Bug fix (patch bump: 1.6.1 → 1.6.2)
git commit -m "fix(datamodule): resolve memory leak in batch loading"
# Performance (patch bump)
git commit -m "perf(transformer): optimize attention computation"
# Documentation (no bump)
git commit -m "docs: update API reference for MambaTab"
# Breaking change (major bump: 1.6.1 → 2.0.0)
git commit -m "feat!: remove Python 3.9 support
BREAKING CHANGE: Python 3.10+ is now required"Common scopes in this project:
models: Model implementationsconfigs: Configuration classesdata: Data utilities and dataloadersarch: Architecture utilitiesutils: General utilitiesci: CI/CD relateddeps: Dependencies
# Interactive commit (recommended)
just commit
# Version bump
just bump
# View changelog
cat CHANGELOG.md
# Dry-run semantic release
just release-dryUse ! after type and explain in footer:
feat!: change API signature
BREAKING CHANGE: The `fit()` method now requires `x_train` and `y_train` as separate arguments instead of a tuple.
# Using editor
git commit
# In editor:
feat(models): add multi-head attention support
This commit introduces multi-head attention mechanism
to improve model performance on large datasets.
Closes #123Commits are validated automatically. If rejected:
- Check format:
type(scope): description - Use allowed types only
- Keep header under 72 characters
- Don't end subject with period