Compatible Model Scope

Convert

You can automatically convert the AI model's framework to the target framework.

  • Only single-input models are supported.
  • The supported channel ordering format is 'Channel First'. The three-dimensional array structure of images should be organized in the order of Batch, Number of Channels, Height, and Width.
  1. Batch size: The number of combined input datasets that the model processes simultaneously.
  2. Channel: enter 3 for RGB channel and 1 for gray channel.
  3. Input size: In computer vision tasks, input size refers to the size of the input images.

ONNX to TensorRT

Target DeviceSW versionInput datatypeBatch sizeChannelInput sizeOutput datatype
NVIDIA Jetson NanoJetPack 4.4.1,
JetPack 4.6
FP321~4
(Static, Dynamic)
1~4height, widthFP16
NVIDIA Jetson Xavier NXJetPack 5.0.2,
JetPack 4.6
FP321~4
(Static, Dynamic)
1~4height, widthFP16
NVIDIA Jetson TX2JetPack 4.6FP321~4
(Static, Dynamic)
1~4height, widthFP16
NVIDIA Jetson AGX XavierJetPack 4.6FP321~4
(Static, Dynamic)
1~4height, widthFP16
NVIDIA Jetson AGX OrinJetPack 5.0.1FP321~4
(Static, Dynamic)
1~4height, widthFP16
NVIDIA T4-FP321~4 (Static, Dynamic)1~4height, widthFP16

ONNX to OpenVino

Input datatypeBatch sizeChannelInput sizeOutput datatype
FP321~4 (Static, Dynamic)1~4height, widthFP16

ONNX to TFlite

Input datatypeBatch sizeChannelInput sizeOutput datatype
FP321~4 (Static, Dynamic)1~4height, widthFP16
INT8

TensorFlow to TensorFlowLite

Input datatypeBatch sizeChannelInput sizeOutput datatype
FP321~4 (Static, Dynamic)1~4height, widthFP16
INT8

Benchmark

Measure the performance of the AI model on the target device. You can measure the actual performance on the devices listed below without the need to purchase or install the device.

Below is a table for benchmarking compatible model frameworks.

Arm MCU

Target Device.tflite
Renesas RA8D1 (Arm Cortex-M85)O (only INT8)
Renesas RA8D1 (Arm Cortex-M85) with heliumO (only INT8)
Alif Ensemble DevKit Gen2 (Arm Cortex-M85 + Ethos-U55)O (only INT8)
Alif Ensemble DevKit Gen2 (Arm Cortex-M85 + Ethos-U55) with heliumO (only INT8)

NVIDIA

  • When benchmarking on Jetson, it is essential for the model file and target device to match the Jetpack version.
Target Device.tflite.engine.tflite.onnx
Jetson Nano JetPack 4.4.1OOOO
Jetson Nano JetPack 4.6OOOO
Jetson Xavier NX JetPack 4.6OOOO
Jetson Xavier NX JetPack 5.0.2OOOO
Jetson TX2 JetPack 4.6OOOO
Jetson AGX Xavier JetPack 4.6OOOO
Jetson AGX Orin JetPack 5.0.1OOOO
AWS-T4OOOO

Raspberry Pi

Target Device.tflite.onnx
Raspberry Pi ZeroWOO
Raspberry Pi Zero2WOO
Raspberry Pi 2BOO
Raspberry Pi 3BOO
Raspberry Pi 3B+OO
Raspberry Pi 4BOO

Intel

Target Device.zip(bin+xml)
Xeon W-2223O