Imagine your AI vision application running on an edge device within a few clicks. You would need a place where complex, high-performance AI applications for Edge inference are simple, visual, and seamless to evaluate and deploy on cutting-edge hardware. That’s why we launched Palette Edgematic, a powerful online tool that lets you experience models and applications on SiMa.ai’s high-performance MLSoC right from your browser in a matter of minutes.
Palette Edgematic, developed to leverage all of the powerful capabilities of Palette, is designed for you to meet industry demands for high-performance AI applications by redefining edge AI development. With Palette Edgematic, developers move from concept to real-world deployment with a proven machine learning system-on-a-chip (MLSoC) in minutes. Let’s explore the multitude of features that make this software platform a game-changer, and look ahead at the powerful new capabilities coming soon.
Feature Highlights
Available on AWS Marketplace. Palette Edgematic is now integrated with AWS Marketplace. As part of this integration, Edgematic received AWS Foundational Technical Review (FTR) Approval, elevating it as an industry-trusted, secure, and highly accessible tool. AWS FTR Approval is a stamp of quality, indicating Edgematic has been rigorously reviewed by AWS for security, reliability, and performance. This designation assures users that Edgematic meets AWS technical standards and is optimized for seamless integration with AWS services, such as Amazon SageMaker.
Signing up via AWS Marketplace simplifies access and with just a few clicks, you can explore the full capabilities of Palette Edgematic directly from your browser. For anyone looking to accelerate their edge AI projects, this AWS-backed availability makes deploying Edgematic easy, secure, and streamlined—so you can focus on innovation, not setup.
Once you are logged in, you can create and manage projects tailored to your needs. Each project organizes your applications and models in a sleek, intuitive interface, where your progress is automatically saved and stored. Whether you are finetuning a model in the cloud and updating your application pipeline for evaluation, so you’re ready to jump back in whenever inspiration strikes!
Automatic MLSoC Device Assignment. With Palette Edgematic automatic MLSoC device assignment, users are provided immediate access to MLSoC development kits, removing the overhead typically associated with hardware setup. This feature is a game-changer for developers eager to dive right into testing and deployment. By instantly assigning MLSoC devices, Edgematic allows users to focus quickly on their applications rather than managing infrastructure, ensuring a smooth, uninterrupted workflow. This ready-to-use hardware availability accelerates project timelines, making it easy to experiment, iterate, and deploy applications faster than ever. To aid in testing and exploration, every board on the Edgematic platform is equipped with power meters, giving users full board and chip-level power readings. This setup enables extremely accurate KPI tracking across key metrics, including total power consumption, making it invaluable for developers focused on efficiency and sustainability. The ability to monitor power usage for the entire board brings a deeper level of insight, an essential edge for real-world deployment.
Drag-and-Drop Application Evaluation. Edgematic lets you bring your AI applications to life on the ‘Canvas’ with a simple drag-and-drop action. Want to check the source code of a plugin or see how it’s configured within GStreamer? Just a few clicks away! Edgematic gives you the power to view, modify, and run from sample applications effortlessly. It provides everything you need for a comprehensive, end-to-end workflow—from input source integration to video stream outputs, all optimized for real-time inference on SiMa.ai MLSoC. And it is not a static interface—each application’s processing flow is visually represented, allowing you to trace data paths and monitor activity at each stage of the pipeline.
One-click Real-time Performance Estimation. Before you even run your applications, Edgematic estimates the latency and throughput for each plugin across every processor. This immediate insight saves developers need for manual testing, providing a clear guidance of how your application might perform on device. This powerful feature streamlines your workflow, giving you the assurance that every part of your pipeline is running at peak efficiency.
One-click Deployment on the SiMa.ai MLSoC. With Edgematic, deploying applications to real-world devices is as simple as clicking a button. Your application streams video from AWS to your allocated MLSoC-hosted board, delivering a seamless, single click experience. Imagine being able to play your video feed with live object detections, processing results in real-time, all available to you in under a minute! This rapid, high-fidelity deployment is what sets Edgematic apart from other tools on the market and provides developers with a preview of how their applications perform under real-world conditions.
Live KPI Monitoring – Your Pulse on Real-Time Performance. Once your application is up and running, the magic of Live KPI takes center stage. With just one click, you can get real-time latencies and throughput readings from each processor on the device. Instead of post-execution reports or complex data gathering, Edgematic empowers you to stay informed as your application performs. Adjust, iterate, and optimize in real time—live KPI lets you keep your finger on the pulse of your application’s performance, ensuring it’s delivering exactly what you need and when you need it.
Model Catalog: Over 250 AI Models at Your Fingertips. Edgematic isn’t just a tool for running applications; it’s also a place where you will find a library of always growing high-performance AI models. With over 250 models available in the Model Catalog, you’re only seconds away from choosing and deploying any model you need. Each model comes with the capability to provide live KPIs on the pre-allocated board—so you’ll know exactly how well it’s performing in terms of throughput, latency, efficiency, and power consumption. Need to compare results? Switch models in just a few clicks, and let Palette Edgematic do the heavy lifting.
Bring Your Own Model. Innovation doesn’t stop at Palette Edgematic Model Catalog. Bring your own ONNX model, and Palette Edgematic will automatically compile and deploy it onto one of our MLSoC devices. It provides seamless integration, allowing you to take advantage of real-time KPIs, just as you would with any other model in the library. This flexibility is perfect for developers looking to customize and fine-tune their model performance or those experimenting with new architectures. While it does not provide accuracy metrics, Edgematic hardware-optimized compilation ensures your models achieve top-notch performance, making the platform ideal for pushing the limits of edge AI through the model exploration phase of your development.
A Glimpse into the Future
SiMa.ai is committed to continuously enhancing Palette Edgematic’s capabilities. Here’s a look at some of the features that are in development and will soon transform the way you interact with Edge AI.
Edge MLOps with Cloud Integration. With MLOps Cloud Integration, Edgematic will bring the power of cloud-native AI model management to the edge. This feature will streamline collaboration between developers, data scientists, and IT teams, allowing them to manage the models and monitor their performance seamlessly from the cloud. Imagine the convenience of orchestrating model deployment, updates, and monitoring across multiple edge devices—this feature will make Edgematic a hub for scalable edge AI solutions.
Application Local Deployment. The upcoming Application Local Deployment feature will give users the flexibility to deploy applications directly on local hardware, adding flexibility to run either on the cloud allocated MLSoC or your local MLSoC devkit. This option will be a game-changer for use cases where real-time data processing, privacy, or bandwidth limitations make local deployment ideal. It’s all about bringing Palette Edgematic incredible edge AI power closer to the source, enabling quick, high-performance solutions wherever you need them.
Application Development. Expanding on Palette Edgematic capabilities, the Application Development feature will allow developers to design custom AI applications from the ground up. Instead of working with predefined models and configurations, you’ll be able to craft pipelines, fine-tune plugins, and customize processing logic for your specific needs. This development capability will position Edgematic as a versatile platform, empowering users to fully leverage the capabilities of SiMa’s MLSoC in entirely new ways.
Why Edgematic? It’s About Empowering Your AI Vision
Building the future of Edge AI. Palette Edgematic marks a turning point for developers, ML practitioners, and domain experts in the world of edge AI. It’s about giving you the tools to visualize, iterate, and deploy with unmatched ease and speed. Every feature, from drag-and-drop applications to real-time KPI monitoring, is designed to enhance and accelerate your development process. Whether you’re a seasoned developer, or new to edge AI, Edgematic is committed to deliver an intuitive and powerful experience that will transform the way you create.
Edgematic is more than a platform; it’s a toolkit for building the future of AI at the edge. Ready to take your AI journey to the next level? Join us and experience Palette Edgematic groundbreaking features for yourself. With Edgematic, the future of edge AI development is here, and it’s waiting for you!
Sign up for our Early Access Program to get access to exclusive features, the opportunity to influence SiMa.ai’s roadmap, and access to our growing developer community. Secure your spot today!