Deploy PaddlePaddle on Arm Cortex-M with Arm Virtual Hardware
Introduction
Overview of OCR
Deploy the PaddleOCR model
Next Steps
Deploy PaddlePaddle on Arm Cortex-M with Arm Virtual Hardware
Who is this for?
This is an introductory topic for software developers interested in using PaddlePaddle for Arm Cortex-M processors.
What will you learn?
Upon completion of this Learning Path, you will be able to:
- Export Paddle inference model
- Compile Paddle inference model with TVMC
- Deploy on the AVH Corstone-300 platform with Arm Cortex-M55
Prerequisites
Before starting, you will need the following:
- Some familiarity with embedded programming
- Some familiarity with AI/ML software development
- An Amazon Web Services(AWS) account to subscribe Arm Virtual Hardware Amazon Machine Image(AMI)
Summary
This summary was drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.
You’ll export a PaddlePaddle OCR text recognition inference model, compile it with TVMC, and deploy the result to the Arm Corstone-300 Fixed Virtual Platform (FVP) with Cortex-M55 processors using Arm Virtual Hardware. Starting with a pre-trained PaddlePaddle model, you’ll produce a TVMC-compiled output for the target, and run the workload in a bare-metal environment on the FVP. Along the way, you’ll make the key export, compilation, and deployment choices required for Cortex-M, then validate the workflow by executing the model on the virtual platform.
Frequently asked questions
These FAQs were drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.
No. The steps use a pre-trained PaddlePaddle model for text recognition, so you export and deploy the inference model without training.
You’ll export a Paddle inference model that TVMC can compile.
Compile the model for the Arm Corstone-300 Fixed Virtual Platform with Cortex-M55 processors. The workflow prepares an output intended for that target.
It runs on the Corstone-300 FVP included with Arm Virtual Hardware. You’ll begin by launching the Arm Virtual Hardware Amazon Machine Image (AMI) on AWS.
The Learning Path focuses on text recognition, the stage after text detection that converts an image region into text. Text detection and model training are out of scope.