Supported Devices, Chip Architectures, and Systems - Amazon SageMaker AI
DevicesSystems and Chip Architectures
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Supported Devices, Chip Architectures, and Systems

Amazon SageMaker Neo supports the following devices, chip architectures, and operating systems.

Devices

You can select a device using the dropdown list in the Amazon SageMaker AI console or by specifying the TargetDevice in the output configuration of the CreateCompilationJob API.

You can choose from one of the following edge devices:

Device List System on a Chip (SoC) Operating System Architecture Accelerator Compiler Options Example
aisage None Linux ARM64 Mali None
amba_cv2 CV2 Arch Linux ARM64 cvflow None
amba_cv22 CV22 Arch Linux ARM64 cvflow None
amba_cv25 CV25 Arch Linux ARM64 cvflow None
coreml None iOS, macOS None None {"class_labels": "imagenet_labels_1000.txt"}
imx8qm NXP imx8 Linux ARM64 None None
imx8mplus i.MX 8M Plus Linux ARM64 NPU None
jacinto_tda4vm TDA4VM Linux ARM TDA4VM None
jetson_nano None Linux ARM64 NVIDIA {'gpu-code': 'sm_53', 'trt-ver': '5.0.6', 'cuda-ver': '10.0'}

For TensorFlow2, {'JETPACK_VERSION': '4.6', 'gpu_code': 'sm_72'}
jetson_tx1 None Linux ARM64 NVIDIA {'gpu-code': 'sm_53', 'trt-ver': '6.0.1', 'cuda-ver': '10.0'}
jetson_tx2 None Linux ARM64 NVIDIA {'gpu-code': 'sm_62', 'trt-ver': '6.0.1', 'cuda-ver': '10.0'}
jetson_xavier None Linux ARM64 NVIDIA {'gpu-code': 'sm_72', 'trt-ver': '5.1.6', 'cuda-ver': '10.0'}
qcs605 None Android ARM64 Mali {'ANDROID_PLATFORM': 27}
qcs603 None Android ARM64 Mali {'ANDROID_PLATFORM': 27}
rasp3b ARM A56 Linux ARM_EABIHF None {'mattr': ['+neon']}
rasp4b ARM A72 None None None None
rk3288 None Linux ARM_EABIHF Mali None
rk3399 None Linux ARM64 Mali None
sbe_c None Linux x86_64 None {'mcpu': 'core-avx2'}
sitara_am57x AM57X Linux ARM64 EVE and/or C66x DSP None
x86_win32 None Windows 10 X86_32 None None
x86_win64 None Windows 10 X86_32 None None

For more information about JSON key-value compiler options for each target device, see the CompilerOptions field in the OutputConfig API data type.

Systems and Chip Architectures

The following look-up tables provide information regarding available operating systems and architectures for Neo model compilation jobs.

Linux
Accelerator X86_64 X86 ARM64 ARM_EABIHF ARM_EABI
No accelerator (CPU) Yes No Yes Yes Yes
Nvidia GPU Yes No Yes No No
Intel_Graphics Yes No No No No
ARM Mali No No Yes Yes Yes
Android
Accelerator X86_64 X86 ARM64 ARM_EABIHF ARM_EABI
No accelerator (CPU) Yes Yes Yes No Yes
Nvidia GPU No No No No No
Intel_Graphics Yes Yes No No No
ARM Mali No No Yes No Yes
Windows
Accelerator X86_64 X86 ARM64 ARM_EABIHF ARM_EABI
No accelerator (CPU) Yes Yes No No No