Sagemaker Neo compilation on FreeRTOS enabled devices like Cortex M7 device

0

Hi, I am looking for any support on how to compile and deploy machine learning model to FreeRTOS target devices like Cortex M7. It is clear that compilation jobs are only supported for Windows and Linux platform devices in Sagemaker. Can we customize the target device?

I also need any tutorials on how we can use Amazon FreeRTOS and AWS Green Grass for such devices? Like can we add FreeRTOS device as core Green Grass device? If not, what are the alternative options?

2 回答
0

Hi, FreeRTOS is not a supported Operating System of SageMaker Neo.

Greengrass and FreeRTOS are fundamentally different types of products - Amazon FreeRTOS is an operating system that powers embedded Microcontroller Units (MCU) for sensors and smart lightbulbs (with no CPU). Greengrass is AWS's IOT edge computing platform designed for edge devices running on Linux such as Raspberry Pi, IMX6/IMX8 devices and PC's.

You can connect FreeRTOS client devices to Greengrass devices, but you cannot run Greengrass on FreeRTOS. You can read more about this process from here and here.

AWS
支持工程师
已回答 1 年前
0

For the compilation part, please be aware of STM32Cube.AI. You can train your model with SageMaker and export in a format like TensorFlow Lite or ONNX. Then use STM32Cube.AI to generate optimized C code for your MCU. STM32Cube.AI has a CLI, so you should be able to integrate it into a pipeline workflow to automate training all the way through to firmware build and deployment. Deployment could be handled by AWS IoT Jobs or OTA Updates.

I agree with @Cadence_L's answer. For STM32M7, you would use FreeRTOS and not involve Greengrass.

profile pictureAWS
专家
Greg_B
已回答 1 年前

您未登录。 登录 发布回答。

一个好的回答可以清楚地解答问题和提供建设性反馈,并能促进提问者的职业发展。

回答问题的准则