TL;DR:
- Supergate will showcase its Deep-learning Microcontroller (DMC) at the Embedded Vision Summit.
- The DMC utilizes AiM Future’s NeuroMosAIc Processor technology for intelligent end-device applications.
- The DMC is a 32-bit RISC-V-based Microcontroller for low-cost AI inference applications.
- It integrates the NMP-300 processor optimized for neural networks and convolutional neural networks (CNN).
- The DMC is suitable for Tiny Machine Learning (TinyML) and Artificial Intelligence of Things (AIoT) applications.
- The NeuroMosAIc Processor family includes the NMP-300, NMP-500, and NMP-700, catering to various machine learning requirements.
- Supergate’s collaboration with AiM Future enhances the DMC’s capabilities for intelligent edge AI workloads.
- The NeuroMosAIc Studio provides developers with tools for running pre-trained machine learning models on NMP hardware.
- Support for industry-standard frameworks like ONNX, PyTorch, and TensorFlow Lite is available.
- Supergate’s Deep-learning Microcontroller and NeuroMosAIc Processor showcase their commitment to AIoT innovation.
Main AI News:
Supergate, a prominent technology company, is set to showcase its latest innovation, the Deep-learning Microcontroller (DMC), at the upcoming Embedded Vision Summit in Santa Clara, California. This premier event brings together professionals from various industries to explore practical computer vision and visual AI solutions. Supergate’s DMC leverages the advanced NeuroMosAIc Processor technology, licensed from AiM Future, to deliver intelligent end-device applications in areas like robotics, smart home consumer products, toys, and the broad realm of artificial intelligence of things (AIoT).
The DMC, based on a 32-bit RISC-V architecture, integrates the highly optimized NMP-300 processor, enabling efficient neural network and convolutional neural network (CNN) processing. This microcontroller is ideal for Tiny Machine Learning (TinyML) and AIoT applications, allowing the development of intelligent and connected devices that operate with minimal power consumption and latency. For instance, a TinyML model deployed in a smart thermostat can optimize temperature control based on user preferences and patterns, while an AIoT platform collects data from multiple devices to enhance energy efficiency.
Supergate’s collaboration with AiM Future has yielded significant advantages for the DMC. The NMP-300, known for its small size and low power consumption, perfectly aligns with the microcontroller’s target applications, providing a compelling edge over traditional counterparts. The partnership enables Supergate to deliver the advanced AI capabilities required by their customers for intelligent end-devices.
The NeuroMosAIc Processor family offers a range of solutions to cater to diverse machine learning applications. The NMP-300 is designed for ultra-low-power devices like wearables and sensors, delivering up to 0.5 TOPS (tera-operations per second) while consuming only microwatts of power. The NMP-500 targets performance efficiency in applications such as smartphones, drones, AR/VR devices, and high-end home appliances, providing up to 4 TOPS of processing power. For applications demanding higher performance, the NMP-700, with its 8 or 16 TOPS capacity, is ideal for edge gateways, servers, robotics, and UAVs.
To facilitate development, the NeuroMosAIc Studio offers a comprehensive set of tools for running pre-trained machine learning models on NMP hardware. This includes a hardware-aware model converter, mapper, and profiler, ensuring maximum efficiency. The software also supports industry-standard frameworks such as ONNX, PyTorch, and TensorFlow Lite, enabling flexibility and optimized model performance through compilers like GLOW and TensorFlow XLA.
Conlcusion:
The introduction of Supergate’s Deep-learning Microcontroller (DMC) powered by AiM Future’s NeuroMosAIc Processor signifies a significant development in the market for intelligent edge AI applications. This collaboration delivers advanced AI capabilities, enabling low-cost AI inference, Tiny Machine Learning (TinyML), and Artificial Intelligence of Things (AIoT) applications.
The integration of optimized processors and comprehensive software tools showcases the market’s focus on providing efficient solutions for running pre-trained machine learning models. This advancement holds great potential for driving innovation and expanding opportunities in various industries, propelling the market further towards intelligent, connected devices and enhanced AIoT ecosystems.
