OmniML launches Omnimizer to simplify ML deployment on edge devices

Categories Edge Applications  |  Edge Startups
OmniML launches Omnimizer to simplify ML deployment on edge devices

OmniML, a start-up that aims to make AI more accessible to edge devices, has developed a platform that connects ML models with edge hardware, making machine learning operations (MLOps) simpler and faster.

Omnimizer is designed to make it easier for enterprises to develop and deploy machine learning models on various types of hardware. It does this by abstracting away the complexities of working with different devices. For example, the platform helps businesses overcome inefficiencies that can slow down deployment, increase costs, and adversely impact performance.

“Omnimizer solves this by unifying workstreams of ML development and deployment, enabling enterprises to adapt existing models for their hardware based on their specific business needs,” said OmniML co-founder and CEO Di Wu, Ph.D.

The technology behind Omnimizer is based on “deep compression,” a technique invented by Stanford University researchers that allows for more efficient data storage and processing. The platform is intended to work with any ML model. It supports various devices, including CPUs, GPUs, and AI SoCs.

The company is backed by GGV Capital, Qualcomm Ventures, and other notable investors. Dr. Song Han founded the start-up with MIT EECS Professor and experienced start-up founder, Dr. Di Wu, former Facebook engineer, and Dr. Huizi Mao, co-inventor of the deep compression technology.

The release of Omnimizer comes at a time when the demand for edge AI grows rapidly as more businesses look to deploy AI applications on devices such as autonomous vehicles, robots, and IoT devices. However, cloud-based ML models often do not work well on edge devices because they are not designed to work with them. Such a disconnect severely inhibits AI’s potential and has been a barrier for businesses involved with edge AI.

The company claims that Omnimizer can solve this problem by automatically adjusting and optimizing models for hardware. In this way, ML engineers can focus on model designing and training without worrying about whether the item will deploy properly. For example, a smart camera maker used Omnimizer to significantly reduce the complexity of designing a model for its low-cost chips. This situation considerably reduces deployment time while achieving excellent inference performance on its edge devices.

“Our work with OmniML reflects our continued commitment to developing powerful, next-generation software, advanced machine learning and seamless AI solutions to further advance robotics through platforms such as the Qualcomm Robotics RB5 Platform,” said Dev Singh, senior director of business development and head of autonomous robotics at Qualcomm Technologies, Inc.

Article Topics

 |   |   |   |   | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Edge Analysis & Opinion

Featured Company

Edge Market Reports

Explaining Edge Computing

Edge Computing Events

Latest News