Deep Learning

What’s New in MATLAB for Deep Learning?

MATLAB makes deep learning easy and accessible for everyone, even if you’re not an expert. Check out the latest features for designing and building your own models, network training and visualization, and deployment.

Data Preparation and Labeling

  • App to label pixels and regions for semantic segmentation and object detection
  • Video Labeler app: Interactive labeling of ground-truth data in a video or image sequences
  • New Patch image datastore: Extract random image patches to split up large images
  • Pixel label datastore: Store pixel information for semantic segmentation data
  • Augmented image datastore: Create more training samples to augment deep learning training data
  • New Audio Labeler app: Interactively define and visualize ground-truth labels for audio datasets

Network Architectures

  • New Train faster R-CNN object detectors using DAG networks such as ResNet-50 and Inception-v3
  • Regression and bidirectional LSTMs for continuous, time-series outputs
  • New Deep Learning Network Designer app: Graphically design and analyze deep networks
  • Custom layers support: Define new layers and specify loss functions for classification and regression output layers
  • Automatic validation of custom layers to check for data size and type consistency

Accessing the Latest Pretrained Models

  • New ONNX model converter: Import and export models using the ONNX model format for interoperability with other deep learning frameworks
  • Ability to work with ResNet-18, DenseNet-201, Inception-ResNet-v2, and SqueezeNet
  • TensorFlow-Keras model importer and Caffe model importer

See a comprehensive list of pretrained models supported in MATLAB.

Network Training

  • Automatically validate network performance, and stop training when the validation metrics stop improving
  • Perform hyperparameter tuning using Bayesian optimization
  • Additional optimizers for training: Adam and RMSProp
  • Train DAG networks in parallel and on multiple GPUs

Debugging and Visualization

  • DAG activations: Visualize intermediate activations for networks like Inception-ResNet-v2, ResNet-50, ResNet-101, GoogLeNet, and Inception-v3
  • Monitor training progress with plots for accuracy, loss, and validation metrics
  • Network Analyzer: Visualize, analyze, and find problems in network architectures before training

Deployment

  • Automatically convert trained deep learning models in MATLAB® to CUDA using GPU Coder™
  • Integrate the generated CUDA code with NVIDIA® TensorRT
  • Support for DAG networks including GoogLeNet, ResNet-50, ResNet-101, and SegNet
  • Generate code from trained deep learning models for Intel® Xeon and ARM® Cortex-A® processors
  • New Automated deployment to NVIDIA Jetson and DRIVE platforms
  • New Deep learning optimization: Improved performance through auto-tuning, layer fusion, and Thrust library support

Get a Free Trial

30 days of exploration at your fingertips.

Have Questions?

Talk to a deep learning expert.