Understand the architecture of the machine learning application

What you’ll deploy

Topo is an open-source command-line tool developed by Arm that you can use to deploy projects to an Arm-based Linux target over SSH. Topo builds container images on the host, transfers them to the target, and starts the services on the target. Topo Templates are the standardized format for deploying projects with Topo.

In this Learning Path, you’ll deploy the topo-imx93-npu-deployment Topo Template to an NXP FRDM i.MX 93 board, and understand how this Topo Template was created.

The Topo Template builds and deploys a browser-based MobileNetV2 image classifier. The user interface runs on the Cortex-A (Linux) side of the SoC. The inference runner is packaged as Cortex-M33 firmware and is started by remoteproc-runtime . The model is exported to an ExecuTorch .pte file for Ethos-U65 NPU acceleration.

Prerequisites

Before getting started, ensure that your i.MX 93 board is set up with Linux and accessible over SSH. Use the Learning Path Use Linux on the NXP FRDM i.MX 93 board as a guide.

Complete the Learning Path Deploy containerized workloads to Arm-based Linux targets with Topo to learn how to install Topo, run host and target health checks, inspect a target, list compatible Topo Templates, and deploy a containerized workload.

(Optional) Background reading

To learn more about Topo Templates, and how to create a basic Topo Template for a web application, complete the introductory Create and deploy a custom Topo Template Learning Path.

To learn more about the model, firmware, and Ethos-U65 execution flow behind this NPU example, see the Deploy ExecuTorch firmware on NXP FRDM i.MX 93 for Ethos-U65 acceleration Learning Path.

What the template does

Deploying the Topo Template starts two runtime services on the target:

• webapp: Web application running on the Cortex-A Linux host. It receives an image input from the user and outputs the results of the ML image classification.
• cm33-runner: Cortex-M33 firmware that receives the image tensor from the web application, runs the compiled MobileNetV2 ExecuTorch .pte program, delegates supported operators to the Ethos-U65 NPU, and runs non-delegated operators on the Cortex-M33 CPU.

System architecture

The deployed application spans three processing domains on the i.MX 93:

• Cortex-A Linux host: runs Docker, Topo-deployed containers, the Flask web app, and the Linux remoteproc and RPMsg interfaces.
• Cortex-M33 firmware domain: runs the ExecuTorch runner firmware loaded by remoteproc-runtime.
• Ethos-U65 NPU: accelerates delegated neural network operators from the ExecuTorch MobileNetV2 program.

The high-level data flow is:

    

        
        Browser
  |
  v
Flask web application on Cortex-A Linux
  |
  | writes .pte file and input tensor to reserved memory
  | sends RUN over RPMsg
  v
Cortex-M33 ExecuTorch runner firmware
  |
  | loads the .pte program from reserved memory
  | delegates supported operators
  v
Ethos-U65 NPU
  |
  v
Cortex-M33 returns classification results over RPMsg
  |
  v
Browser displays ImageNet top-1 and top-5 results

        
    

What you’ve learned and what’s next

You now understand that the Topo Template deploys a Cortex-A web application, a Cortex-M33 ExecuTorch runner, and Ethos-U65 NPU acceleration as one heterogeneous application. You’ve also seen how inference uses reserved memory for the .pte program and input tensor, with RPMsg carrying commands and results between Cortex-A and Cortex-M33.

Next, you’ll review the toolchains and runtime interfaces used by the Topo Template.