Accelerate your AI project development on STM32 MCUs
MCUs are constrained devices. This means that deploying artificial neural networks at the far-edge requires tailor-made solutions and tools to address the specific challenges of embedded AI on Arm-based MCUs.
In this Tech Talk hosted by Arm, ST experts will present a set of AI tools, which allows users to execute optimized artificial neural networks on STM32 in an efficient way. They will show you how these tools can be easily merged with current practices and workflows of embedded software developers and AI data scientists alike.
The STM32Cube.AI software accelerates inference time on Arm® Cortex®-M-based MCUs, offering up to 60% performance improvement and 20% memory footprint reduction. ST experts will explain:
- How various neural networks can be easily benchmarked on real STM32 silicon using our online STM32Cube.AI Developer Cloud
- How AI projects can be developed in hours thanks to the STM32 model zoo, Python wrappers and Jupyter notebooks.
- Introduction to AI solutions on STM32 MCUs
- Create a flower recognition application with an STM32H7
- Easily optimize and benchmark AI neural network models with STM32Cube.AI Developer Cloud
- Pick your model from STM32 Model Zoo and manage the complete workflow in Python with Jupyter Notebook
Nicolas is a product marketing manager in the artificial intelligence solutions team at STMicroelectronics. He is responsible for the ecosystem and tools enabling the democratization of AI on microcontrollers. He is passionate about technology and has been working for several years to develop real-world cases where AI is a game changer.
An AI enthusiast, Muhammad is a senior data science and algorithm engineer in the artificial intelligence solutions team at STMicroelectronics. He has a PhD in acoustics and signal processing from Politecnico di Milano, Italy. Muhammad has experience working in different international research labs. Since 2019, Muhammad has been focusing on AI, working on various topics related to data science, TinyML, sensing, and computer vision. His day work revolves around generating and optimizing different AI and machine learning models, deploying them on resource-constrained devices such as microcontrollers and microprocessors.