t4-devkit

A toolkit to load and operate T4 dataset.

Getting started

🏠 Documentation | 📝 Dataset Schema | ⚙️ Tutorial | 💻 CLI | 🧰 API Reference

Installation

[For Users] Install via GitHub

Note that the following command installs the latest main branch:

# e.g) with pip
pip install git+https://github.com/tier4/t4-devkit.git

By specifying @<TAG_OR_BRANCH>, you can install the particular version of t4-devkit:

# e.g) with pip
pip install git+https://github.com/tier4/t4-devkit.git@main

[For Developers] Install from source

You need to install uv. For details, please refer to OFFICIAL DOCUMENT.

git clone [email protected]:tier4/t4-devkit.git
cd t4-devkit
uv sync --python 3.10

The virtual environment can be activated with:

source .venv/bin/activate

Feature supports

Visualization

t4-devkit provides a set of visualization tools to help you understand the data.

More details, please refer to t4viz CLI or API usage.

Feature	Task	Support
3D	3D Boxes	✅
	PointCloud Segmentation
	Raw PointCloud	✅
	3D Trajectories	✅
	TF Links	✅
2D	2D Boxes	✅
	Image Segmentation	✅
	Raw Image	✅
	Raw PointCloud on Image	✅
Map	Vector Map	✅
	Ego Position on Street View	✅

Sanity Checks

t4-devkit provides a set of sanity checks to ensure the correctness of the data.

More details, please refer to t4sanity CLI or API usage.