A Multi-modal Large-scale Scene Dataset and A Versatile Toolchain for Scene Prediction - View it on GitHub