Deep Learning Toolbox

Design, train, analyze, and simulate deep learning networks

Deep Learning Toolbox provides functions, apps, and Simulink blocks for designing, implementing, and simulating deep neural networks. The toolbox provides a framework to create and use many types of networks, such as convolutional neural networks (CNNs) and transformers. You can visualize and interpret network predictions, verify network properties, and compress networks with quantization, projection, or pruning.

With the Deep Network Designer app, you can design, edit, and analyze networks interactively, import pretrained models, and export networks to Simulink. The toolbox lets you interoperate with other deep learning frameworks. You can import PyTorch®, TensorFlow™, and ONNX™ models for inference, transfer learning, simulation, and deployment. You can also export models to TensorFlow and ONNX.

You can automatically generate C/C++, CUDA®, and HDL code for trained networks.

Discover new features in the release notes.

Open and Explore an Interactive Example in MATLAB Online

What Is Deep Learning Toolbox? (1:55)

Photo of a PCB titled “Predicted Defects” with three annotations labeled “missing_hole.”

Deep Learning for Engineers

Create and use explainable, robust, and scalable deep learning models for automated visual inspection, reduced order modeling, wireless communications, computer vision, and other applications.

Deep Learning for Engineers: Building an AI System (3:40)

Documentation | Examples

Three screenshots that show virtual sensor modeling using deep learning and one screenshot of a line graph plotting the variables truth, EKF, deep learning - FNN, and deep learning - LSTM.

Deep Learning in Simulink

Use deep learning with Simulink to test the integration of deep learning models into larger systems. Simulate models based on MATLAB or Python to assess model behavior and system performance.

Integrate TensorFlow Model into Simulink for Simulation and Code Generation (5:47)

Documentation | Examples

Flowchart showing that you can import models from TensorFlow, ONNX, and PyTorch, and you can export models to TensorFlow and ONNX.

Integration with PyTorch and TensorFlow

Exchange deep learning models with Python-based deep learning frameworks. Import PyTorch, TensorFlow, and ONNX models, and export networks to TensorFlow and ONNX with a single line of code. Co-execute Python-based models in MATLAB and Simulink.

Interoperability Between Deep Learning Toolbox, TensorFlow, PyTorch, and ONNX

Documentation | Examples

Code Generation and Deployment

Automatically generate optimized C/C++ code (with MATLAB Coder) and CUDA code (with GPU Coder) for deployment to CPUs and GPUs. Generate synthesizable Verilog^® and VHDL^® code (with Deep Learning HDL Toolbox) for deployment to FPGAs and SoCs.

Generate Generic C/C++ Code for Deep Learning Networks in Simulink (5:10)

Documentation | Examples

Four images of the same road scene that represent the test image, semantic segmentation, Grad-CAM of the road, and Grad-CAM of the pavement.

Explainability and Verification

Visualize training progress and activations of deep neural networks. Use Grad-CAM, D-RISE, and LIME to explain network results. Verify the robustness and reliability of deep neural networks.

What Is Explainable AI? (3:45)

Documentation | Examples

The start page of the Deep Network Designer app showing options for importing pretrained models, including models from PyTorch and TensorFlow, and image networks, including SqueezeNet, GoogLeNet, and Res-Net-50.

Network Design and Training

Use deep learning algorithms to create CNNs, LSTMs, GANs, and transformers, or perform transfer learning with pretrained models. Automatically label, process, and augment image, video, and signal data for network training.

Using Transfer Learning | Deep Learning for Engineers, Part 4 (16:18)

Documentation | Examples

Screenshot of the Deep Network Designer app depicting a network with dozens of layers.

Low-Code Apps

Accelerate network design, analysis, and transfer learning for built-in and Python-based models by using the Deep Network Designer app. Tune and compare multiple models using the Experiment Manager app.

Interactively Build, Visualize, and Edit Deep Learning Networks (3:54)

Documentation | Examples

Screenshot of the Deep Network Quantizer showing three separate sections: a net layer graph, calibration statistics, and a validation summary.

Deep Learning Compression

Compress your deep learning network with quantization, projection, or pruning to reduce its memory footprint and increase inference performance. Assess inference performance and accuracy using the Deep Network Quantizer app.

Quantizing a Deep Learning Network in MATLAB (2:50)

Documentation | Examples

Two graphs of training progress show accuracy and loss for training and validation. Accuracy has an upward trajectory and loss has a downward trajectory.

Scaling Up Deep Learning

Speed up deep learning training using GPUs, cloud acceleration, and distributed computing. Train multiple networks in parallel and offload deep learning computations to run in the background.

Speed Up Deep Neural Network Training

Documentation | Examples

Product Resources:

Documentation Examples Videos Technical articles Functions Requirements Release notes

Mercedes-Benz Simulates Hardware Sensors with Deep Neural Networks

“This was the first time we were simulating sensors with neural networks on one of our powertrain ECUs. Without MATLAB and Simulink, we would have to use a tedious manual coding process that was very slow and error-prone.”

View more customer stories