The module makes edge AI faster and cheaper. It is easy to use, works with many systems, and supports robots, machines, and drones.
SiMa.ai, a company focused on software-driven embedded edge machine learning SoCs (MLSoC), has launched its system-on-module (SoM) built on the MLSoC Modalix 50 TOPs chip. This SoM was developed with Enclustra, a company in SoM and FPGA design. The module expands SiMa.ai’s offerings to meet the needs of organizations working on multi-modal and Gen AI applications at the embedded edge.
The SoM uses the MLSoC Modalix chip in a format that works with systems from the GPU SoM provider in edge AI. This allows companies to integrate it into systems and carrier boards, reducing development time and complexity. When used with SiMa.ai’s Palette and Palette Edgematic tools, the SoM provides flexibility and speeds up development and deployment of Gen AI applications.
The companies say they have made a solution that cuts the cost of using AI at the edge. It works in small environments and supports AI uses such as robotics, automation, and unmanned aerial systems.
Some of the key features the SoM include:
- 50 TOPS chip made for edge AI work
- Many input and output options
- Small size fits into many systems
- Software helps build and run AI easily
- Supports Python, PyTorch, and OpenCV
- Handles heat well and works in tough conditions
- Monitors performance in real-time
- Works in both normal and industrial temperatures
“The introduction of our Modalix SoM represents a critical expansion that addresses the diverse needs of the fragmented edge AI market,” said Krishna Rangasayee, CEO and founder of SiMa.ai. “Our partnership with Enclustra is removing integration barriers that typically slow companies down, making it remarkably simple for companies to upgrade from legacy edge compute solutions to Modalix’s superior performance.”
“At Enclustra, our mission is to drive high-performance, energy-efficient computing solutions for the most demanding edge AI applications,” said Philipp Baechtold, CEO of Enclustra.
For more information, click here.
