Nuvoton Technology, a leading global semiconductor provider, has announced the launch of “NuML Studio”. This is a graphic user interface (UI) tool designed specifically for machine learning applications on Nuvoton microcontrollers (MCUs). NuML Studio helps developers solve common problems when building Endpoint AI, providing a clear path from real-time data collection to automatic firmware project generation. This allows developers to focus more on improving their AI models and creating new applications.
Easy Setup to Start AI Development Immediately NuML Studio optimises Windows and provides a “ready-to-use” version that does not require users to install Python or complex software libraries, making it significantly easier for beginners to set up their development environment.
This “download and run” approach allows developers to bypass tedious setup processes and immediately access intuitive project management features, where they can easily create projects for data collection, machine learning deployment, or a combination of both to facilitate rapid iteration.
Strong Data Collection and Conversion Features Accurate data is the foundation of any AI model. NuML Studio provides full support for sensors and automatic data conversion:
- Support for Many Sensors: It supports 3-axis G-sensors, 16KHz Audio, and Image collection using the NuMaker-M55M1 board.
- Automatic Data Conversion: Collected raw data can be converted into standard formats like .csv (for sensors), .wav (for audio), or .jpg (for images) with one click.
- Cloud Integration: With built-in machine learning platform API support, developers can upload their collected data directly to a cloud platform for model training.
Automatic Project Generation for Fast Deployment The core technology of NuML Studio can automatically create firmware projects that follow industry standards:
- Support for Popular Models: It works with the TensorFlow Lite Micro (TFLM) framework and supports quantised models.
- Automatic Firmware Creation: It can automatically generate Keil MDK and VS Code CMSIS projects for tasks like image classification, object detection, and keyword spotting (KWS).
- Hardware Optimisation: For chips with an Arm Ethos-U55 NPU (such as the NuMicro M55M1), it provides special library support to get the best performance from the hardware.
With the launch of NuML Studio, Nuvoton reinforces its commitment to lowering the barriers to Endpoint AI development. By providing an integrated path from real-time data collection to automatic firmware generation, this tool allows developers to bypass complex environment setups and focus on AI model optimisation and innovation. Supporting industry standards and providing specialised library support for hardware acceleration on chips like the Arm Ethos-U55 NPU, NuML Studio empowers developers to deliver high-performance intelligent edge applications with unprecedented speed.
