*Online purchasing is US-only. To buy outside the US, please fill out the form here.

SiMa.ai MLSoC DevKit 2.0

Deploy ML models to production in less time at the edge - today.

MLSoC Platform Overview

Frequently Asked Questions

Can I purchase a DevKit outside the US?

Yes, but online purchasing is US-only at this time. To buy outside the US, please fill out the form here.

What is the SiMa.ai MLSoC DevKit?

The SiMa.ai MLSoC DevKit is designed for ML developers to rapidly evaluate, prototype, and demonstrate computer vision edge ML applications. Palette Software™ (included) provides real-time Key Performance Indicators (KPIs) that provide feedback for developers as they make iterative improvements to their ML models and ML pipelines. Developers can get familiar with the MLSoC by executing pre-configured ML models and pipelines on the developer board out of the box. They can incorporate their own ML models using the included compiler as well as modify and customize the provided example ML pipelines. The DevKit includes a compact Development Board, the most recent version of our Palette Software, power supply and all the essential cables, so you can be up and running in minutes. Read more information on our DevKit here, or view the product brief here.

What do I need to get started?

The DevKit has everything you need to get started on your own desktop PC. For the real-time camera streaming, utilize your own Ethernet based streaming camera or get the camera bundle package available above.  

Following the purchase of a DevKit, please register at the SiMa.ai Developer Portal for access to: 

  • Latest Palette Software Release
  • Online Developer Journey Documentation
  • Hardware and Software Reference Documentation
What is the On-Site Accelerate Session? Is it required to purchase a DevKit?

Although we highly recommend the purchase of an Accelerate session to expedite your deployment, it is not required. On-Site Accelerate customers receive priority access to SiMa support.

During an On-Site Accelerate session, a SiMa.ai technical expert will walk you through your DevKit setup in person (available by adding to your on-line order, above). Accelerate provides a customized face-to face session with a field application engineer at your location to walk through the DevKit bring-up and run out-of-box demonstrations.

The Accelerate session includes up to eight hours of training on Palette Software including demonstrating how to compile, build and deploy your own ML models and pipelines on the Developer Board.

Incorporate your whole team to get up to speed quickly and ask questions as you walk through the development process in an interactive environment.

How can I get additional support?

On-Site Accelerate customers receive priority access to SiMa support. Additional support is available including specialized training on C/C++, Python and GStreamer programming as well as product support, software training and other turn-key services to support the development of proof-of-concept and product development requirements. Following a DevKit purchase and completed registration for our Developer Portal, our team will be in touch as soon as availability permits for those customers who have not purchased an On-Site Accelerate session.

What can I expect when bringing up my SiMa.ai DevKit?

Our Accelerate customers typically enjoy successful DevKit bring-ups in the span of a few minutes. Our Developer Journey Documentation includes detailed instructions to help customers navigate any issues they may encounter as a result of their unique environment with their DevKit deployment. Deployments can be impacted by outdated Docker software, firmware settings, IT restrictions and other unforeseen common occurrences. Customers can quickly mitigate these and more in an On-Site Accelerate session or via our online Developer Journey Documentation on our Developer Portal.

What is the difference between purchasing a standalone DevKit and the DevKit bundle?

The DevKit has everything you need to get started on your own desktop PC. For the real-time camera streaming, utilize your own Ethernet based streaming camera or purchase the DevKit Camera Bundle to receive a camera with a DevKit.

The contents of a SiMa.ai MLSoC DevKit include:

  • SiMa.ai Developer Board (HHHL)
  • Type C Power adaptor & Type C to micro USB – cable
  • Ethernet Cable
  • UART Cable to USB
  • PCIe Mounting Bracket
  • Seat license for SiMa.ai Palette Software

The SiMa.ai MLSoC DevKit camera bundle includes an e-con Systems 1GB RouteCAM_CU20 POE Ethernet camera to demonstrate GStreamer ML pipelines in real-time out of the box.

Camera specifications:

  • Power-over-Ethernet camera with IEEE 802.3af compliance
  • Comes with PoE power adapter and cable
  • Houses Sony® Starvis IMX462 CMOS image sensor
  • Ultra-low light sensitivity & Superior near-infrared performance
  • High Dynamic Range (HDR)
  • On-board high-performance ISP
  • Supports 10Base-T, 100Base-TX and 1000base-T-modes
What is the SiMa.ai MLSoC Platform?

The SiMa.ai MLSoC (Machine Learning System on Chip) enables ML developers to deploy a complete solution for edge ML applications with up to 10X improved power efficiency over preexisting solutions and standalone ML Accelerators.

The SiMa.ai MLSoC is a complete Machine Learning System-on-Chip that comes with support for a complete ML vision pipeline, with industry standard I/O for your data streams, with hardware accelerated video decompression, a computer vision processing subsystem with 4xDSPs, a host ARM subsystem with 4xCPU with L1/L2 coherent cache, and coherent network-on-chip interconnecting to the processing subsystems for application stack hosting and management. 32 MBytes of internal memory is complemented by external DDR and flash memory support to provide for large model and application code. Details can be found in the MLSoC data sheet.

SiMa.ai Palette Software is a unified suite of tools with “pushbutton” software commands to create, build and deploy with ease on multiple devices. Palette Software features both push-button compilation and multi-core code builds to help ML developers scale their ML deployments at the embedded edge with ease. Palette manages all dependencies and configuration issues within a single Docker container while securely communicating to edge devices. This approach enables embedded programmers flexibility to also perform low level optimization within the automated code build environment.

The SiMa.ai MLSoC Developer Kit combines Palette Software and with our MLSoC Developer Board to enable rapid evaluation of ML models, prototyping of proof-of-concept edge ML pipelines and the ability to demonstrate using real-time data streams in a matter of hours.

How can I purchase the products you sell?

We recommend starting with a SiMa.ai MLSoC DevKit to see immediate performance gains in evaluating, prototyping and demonstrating computer vision edge ML applications within hours.

Click “Buy Now” at devkit.sima.ai to get your DevKit today.

How is SiMa.ai software licensed?

Palette Software is licensed as a seat license and provides for use over a specified period of time. Software licenses are subscription based with typical quarterly feature updates, bug fixes, maintenance releases via our Developer Portal.