Скачать или смотреть 3.1 Customizing Models with Keras: Introduction to Keras API

The Keras API stands as a prominent high-level neural networks tool written in Python, compatible with TensorFlow, Theano, or CNTK backends, facilitating swift model prototyping and deployment. Its features include a high-level interface for model development, modular construction with layers, activations, and optimizers, and compatibility with multiple backends. Key components encompass a wide range of pre-defined layers, activation functions, loss functions, optimizers, metrics, and callbacks, offering flexibility and ease of use. An illustrative example demonstrates image classification network construction. Training proceeds by compiling and utilizing the fit() method with training data and labels. Post-training, model evaluation on separate datasets is accomplished via the evaluate() method. Overall, Keras furnishes a user-friendly, versatile framework for building, training, and deploying deep learning models, catering to beginners and seasoned practitioners alike, and enabling rapid experimentation and development of state-of-the-art machine learning models.