Main Content

Deep Learning Fundamentals

Import, build, train, tune, visualize, verify, and export deep neural networks

Deep Learning Toolbox™ provides tools for each stage of the deep learning workflow.

  • Preprocess data for deep network training using command-line functions and interactive apps.

  • Import pretrained networks from MATLAB® or external platforms such as TensorFlow™ 2, TensorFlow-Keras, PyTorch®, and ONNX™.

  • Build networks using command-line functions or interactively using the Deep Network Designer app.

  • Select training options and train networks using built-in training functions or custom training loops.

  • Improve network performance by tuning hyperparameters or running multiple trials using the Experiment Manager app.

  • Visualize and verify network behavior during and after training.

  • Export networks to external platforms such as TensorFlow 2 and ONNX.

Deep learning workflow: preprocess data, import or build network, train network, tune network, visualize and verify results, and export network