TensorFlow 2 focuses on simplicity and ease of use, with updates like eager execution, intuitive higher-level APIs, and flexible model building on any platform. Many guides are written as Jupyter notebooks and run directly in Google Colab—a hosted notebook environment that requires no setup .