Deep Learning Toolbox R2024a: Major Update and New Examples

著者 Sivylla Paraskevopoulou, March 25, 2024

144 ビュー (過去 30 日間) | 0 いいね | 2 コメント

On March 20^th, MATLAB R2024a was released with many updates for Deep Learning Toolbox. Exciting new features for deep learning help engineers create and use explainable, robust, and scalable deep learning models for automated visual inspection, wireless communications, computer vision, and many more applications.

MATLAB Deep Learning Toolbox product page

Some of the new Deep Learning Toolbox capabilities are:

Simulink co-execution blocks to simulate Python®-based (PyTorch®, TensorFlow™, ONNX™, and custom) models in a system-wide context.
Explainability and verification tools to explain network results and verify the reliability of deep neural networks.
Support for more deep learning architectures, including transformers, and training options.

Find out more about Deep Learning Toolbox features and new capabilities:

In this blog post, we present a few new examples that show how to use the new features and apply deep learning. In future blog posts, we will dive deeper into individual new features.

New Examples on Features


Model Design	Explainability and Verification	Integration with Python

Create and Train Network with Nested Layers	Verification of Neural Networks	Predict Responses Using PyTorch Model Predict Block

New Examples on Applications


Computer Vision	Natural Language Processing	Wireless Communications

Interactively Segment Image Using Segment Anything Model	Out-of-Distribution Detection for LSTM Document Classifier	AI for Positioning Accuracy Enhancement

Segment CT Scan Using MONAI Label	Classify Documents Using Document Embeddings	Data Preparation for Neural Network Digital Predistortion Design