Silicon Valley, CA, 26 June 2024, From the TinyML Summit in Milan Italy, embedUR systems Inc., a leader in embedded systems and Edge AI, and a TinyML sponsor partner, is thrilled to announce the launch of ModelNova, a groundbreaking software hub catering for Edge AI solutions.
ModelNova is a model zoo for pre-trained AI models, optimized for different software frameworks and a variety of low-power hardware platforms with and without native AI acceleration. This innovative platform streamlines Edge AI product development, enabling rapid prototyping and deployment of intelligent applications on edge devices, in a fraction of the time it used to take to develop Edge AI solutions from scratch.
ModelNova addresses a significant challenge faced by developers: the complexity of selecting, creating, training, porting, and optimizing AI models for different hardware platforms, especially low-power IoT devices.
This is a critical process, upon-which the success of a project largely depends. It requires extensive expertise in low-level firmware, plus a deep understanding of the silicon they are trying to optimize for, which is a rare skill set.
ModelNova bridges this gap by offering a diverse collection of pre-trained models for AI Vision, Speech, Sound and more, which have already been fine-tuned to run on multiple combinations of hardware and software architectures.
Most AI models begin as academic research and open source implementations trained with a particular dataset not ready for real end product development. Leveraging our unique expertise in AI and embedded software device integration to maximize performance and minimize power usage, we have adapted the models for use in real-world applications.
“Our platform removes the barriers to entry for AI prototyping and development on low-powered devices, enabling engineers to focus on innovation rather than the intricacies of hardware integration and optimization”, says John Marconi, VP Technology at embedUR
Developers will be able to download models known to run on hardware similar to what they have in mind to use, and quickly evaluate the feasibility of their application. This dramatically reduces the time required to build proof of concepts and validate the hardware for Edge AI applications, allowing developers to move from idea to proof of concept in days not months.
“We are excited to work with our partners and customers to speed up product development, bringing AI solutions to more users. ModelNova expands our offerings providing an easy path to new innovative use-cases”, said Rajesh C. Subramaniam, Founder and CEO of embedUR.
“With its focus on models for resource-constrained devices, ModelNova can accelerate the developer journey for IoT Edge AI, and showcases what can be enabled on Synaptics’ Astra™ AI-Native™ devices”, said Vikram Gupta, Senior Vice President and General Manager of IoT Processors, Chief Product Officer, Synaptics
Key Benefits of ModelNova:
Rapid Prototyping: Developers can download pre-trained models for functions similar to those needed by their application and see how well they run on different platforms of choice.
Streamlined Development: Developers only need to adapt or retrain existing models to perform new functions, which is a lot less work than creating the model, in the first place.
Access to Edge AI Expertise: As integrators of these models to the different platforms they run on, embedUR is available to assist with model optimization, training for application use-cases, and integration to target hardware to accelerate productization.
“This is an exciting resource for the developer community, unlocking new possibilities for developing edge AI solutions at a much faster pace” said Pete Bernard, Executive Director of the TinyML Foundation.
Silicon vendors and platform creators who are building new AI-native hardware are encouraged to reach out to embedUR to have ModelNova’s library of models ported to and optimized on their new devices.
ModelNova enables businesses in all sectors from healthcare to agriculture to explore new market opportunities with minimal upfront investment and risk.