To expand the capabilities of development tools and accelerate embedded artificial intelligence (AI) and machine learning (ML) development projects, stmicroelectronics has released the NanoEdge AI Studio and STM32Cub.AI. The updated version of These two development tools help move artificial intelligence and machine learning to application edge devices. After moving to the edge of the network, the benefits of AI and machine learning are outstanding, including native privacy protection, deterministic real-time response, higher reliability, and lower power consumption.
NanoEdge AI Studio is an automated machine learning development tool for applications that do not require the development of neural networks. The tool needs to be used in conjunction with an STM32 microcontroller (MCU) and a MEMS sensor built into stmicroelectronics' unique embedded Smart Sensor Processing Unit (ISPU). Stm32cub. AI is a STM32 artificial intelligence model optimizer and compiler for developers who need to use neural networks. The features of these two new releases help to rapidly develop high-performance AI and machine learning solutions with minimal investment.
NanoEdge AI Studio version 3.2 now includes an automatic data logging generator to improve development efficiency. Inputs include the stmicroelectronics development board and developer-defined sensor parameters, such as data rate, range, sample size, and number of axes. Using this input data, NanoEdge AI Studio can generate binaries for the development board without the developer having to write any code.
Given that data set quality directly affects the performance of machine learning, new data manipulation capabilities in NanoEdge AI Studio allow users to clean and optimize data captured by sensors with a few mouse clicks in NanoEdge AI Studio. A new validation phase is added to help users evaluate the developed algorithm by displaying common performance metrics such as inference time, memory usage and accuracy, and F1-Score. The software also displays more information about the preprocessing and machine learning models involved in the selected library.
NanoEdge AI Studio's latest enhancements add more exception detection and regression algorithm preprocessing techniques and machine learning models to improve application performance. In addition, the tool supports the creation of intelligent libraries that predict future system states with multistep regression models.
Stm32cub. AI version 7.3 is a must-have tool for developing cutting-edge artificial intelligence and machine learning solutions, fully integrated into the STM32 ecosystem, It can convert the well-trained neural network into C language code that can run on the 32-bit MCU of STM32 Arm® Cortex® core, which is popular in the industry. The STM32Cub. AI upgrade adds greater flexibility in optimizing neural networks (NN). The tool can adapt an existing neural network to meet performance requirements or storage capacity constraints, or optimize a neural network for both performance and code volume. The update also adds support for the TensorFlow 2.10 model and new improvements to kernel performance.
