ST Community Manager

STM32N6 demo workshop

Forum|Forum|10 months ago
July 21, 2025
16 replies
4886 views

This workshop offers a comprehensive walk-through of the STM32N6, STMicroelectronics' high-performance MCU with AI acceleration. The session aims to help you explore its extensive capabilities, starting with an understanding of the basic features and architecture of this new device. A significant portion is dedicated to demonstrating how to deploy AI models using the powerful on-board neural processing unit (NPU). Furthermore, the workshop explores how to leverage the N6's resources for multimedia applications, including camera interfacing and H.264 encoding for computer vision.

What you'll learn

The cutting-edge capabilities of the STM32N6
How to leverage the integrated neural processing unit (NPU) for AI acceleration
Effective power supply management for the STM32N6
Deep insights into the STM32N6's boot chain
Advanced multimedia features (camera pipeline, H.264 video encoding)
Practical hands-on experience using the STM32N6 Discovery kit

Ready to get started?

The videos are available at our YouTube playlist embedded below, or at the URL:
https://www.youtube.com/watch?v=3jnal51NBZQ&list=PLTJzs51NlEICI7ARAHOyUqMhTgJE0UFv0&index=1

Join the discussion

Share your thoughts, ask questions, and engage with fellow developers. Our experts are active in this post and they are excited to hear your feedback.

Additional resources

V

venkatesh_vltk

Associate III

Hi,
I have followed every step that you have explained video to run a model in stm32n6-dk board. network_data.hex also generated and I have flashed .hex file using STM32cubeprog tool but my application file it showing the error after building. But in your demo there no such errors. But I facing the issue.
How to solve the below issue.

I am getting error as below:

D:/STM32N6-GettingStarted-ObjectDetection-main/Model/STM32N6570-DK/network.c: In function 'LL_ATON_EpochBlockItems_Default':

D:/STM32N6-GettingStarted-ObjectDetection-main/Model/STM32N6570-DK/network.c:1024:184:

error:

'EpochBlock_Flags_blob_encrypted' undeclared (first use in this function); did you mean 'EpochBlock_Flags_blob'?

| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

| EpochBlock_Flags_blob

D:/STM32N6-GettingStarted-ObjectDetection-main/Model/STM32N6570-DK/network.c:1024:184: note: each undeclared identifier is reported only once for each function it appears in

make: *** [Application/subdir.mk:111: Application/network.o] Error 1

"make -j8 all" terminated with exit code 2. Build might be incomplete.

N. SANTINI

ST Employee

Hi,

I confirm this behavior when using the version 2.2 of the STEdgeAI-Core that was released recently, after the content of this MOOC was created.

In the meantime I invite you to use the version 2.1 of the tool that can still be installed using the same online installer from this link: https://www.st.com/en/development-tools/stedgeai-core.html

Sorry for this situation which is the drawback of constantly evolving tools.

I have escalated the point to the R&D team for investigation and fix.

Best regards,

Nicolas

V

venkatesh_vltk

Associate III

Hi,

Now I am using the the version 2.1 of the STEdgeAI-Core after that my code is error free after I have flashed the file network_data.hex after I was entered to debugging mode in STM32cubeIDE after start debugging I have getting the issue as below

STMicroelectronics ST-LINK GDB server. Version 7.9.0

Starting server with the following options:

Persistent Mode : Disabled

Logging Level : 1

Listen Port Number : 61234

Status Refresh Delay : 15s

Verbose Mode : Disabled

SWD Debug : Enabled

InitWhile : Enabled

Waiting for debugger connection...

Debugger connected

Waiting for debugger connection...

Debugger connected

Waiting for debugger connection...

Target is not responding, retrying...

Shutting down...

Exit.

after LD3 becoming red after I cannot debug or flash any code.

V

venkatesh_vltk

Associate III

Hi, I have followed each and every step in STM32N6 demo workshop - 07 Neural Processing Unit (NPU) in practice in this demo video you have implemented a aton model from .tflite model and it was worked fine to you but I have followed same steps though I not getting the proper output.

K

Kevin_Woo

Associate II

Please tell me where I can get the lecture materials and sample code for the STM32N6 demo workshop released on YouTube.

V

venkatesh_vltk

Associate III

Hello Kevin_Woo,

There is no specific material for the demo there is only pdf of demo slides are available. Those pdf file is about slides of demo only.

Thankyou.

K

Kevin_Woo

Associate II

Hi Venkatesh,

I'd like to try out the n6 dk that arrived yesterday. Could you tell me where I can find documentation explaining how to download the binaries provided on GitHub for the n6 dk using the stm32cubeprogrammer?

Thanksyou

V

venkatesh_vltk

Associate III

You can use the below link to download bin files and code

https://www.st.com/en/development-tools/stm32n6-ai.html

N. SANTINI

ST Employee

Dear all,

As most of you have noticed the STEdgeAI tool is constantly evolving and may lead to backward incompatibility with the STM32N6 A.I. Demos.

This is why these demos are now available directly on gihub repositories:

STM32N6 A.I. Demos gihub links

Additionally, the STEdgeAI tool is now downloadable through an online installer which allows to select which version to install:

STEdgeAI Core tool download

On a new release of the STEdgeAI tool several situations may occur:

Situation A: The new STEdgeAI core version is impacting a SW application package meaning:

1.impacting the code itself: ex. new feature, API break etc.

2.impacting the build system: ex: new file generated, etc.

3.impacting the performances with the model used by default in the SW appli package: ex: inference time significantly different)

then we will upgrade the SW application package on Github repository and a new tag will be created with release notes.We are targeting to publish this upgrade within the 2 weeks following the public release of the new STEdgeAI core version.

Situation B: If the new STEdgeAI core version is NOT impacting a SW application package

Then we will not upgrade the SW application package on gihub repository.

Instead, we will provide a note explaining the user how the upgrade to the new version of STEdgeAI core (i.e. replace inc folder, rename and replace AI runtime lib, etc.: it will be very similar to what is described here).

Situation C: Deployment via the Getting Started SW application packages in the STM32 Model Zoo.

The assumption is that the Model Zoo is always upgraded to the latest version of the STEdgeAI core (and the upgrade is made public shortly after the public release of a new STEdgeAI core version).

* In Situation A above: no problem since the SW application packages that need to be updated will be updated.

* In Situation B above: no problem since the deployment script should automatically do the needful to copy and compile the right inputs.

As an example, STEdgeAI core V2.2 (release done in July 2025): this version has an impact on the SW application packages which have been updated accordingly, and a dedicated tag has been created.

As a summary, when using an A.I. demo from github always refer to the "tags" tab to check the STEdgeAI tool compatibility.

Best regards,

Nicolas

debugging

Lead

Yeah, only slides. Any invite for a workshop in these many years, I ALWAYS try but ALWAYS get rejected and result in the follow response. Perhaps STM applies some kind of selective process by email, name, partnership, region etc.. Why sending such vapor invites to those who don't make a chance to join anyhow?

Show more replies

What you'll learn

Ready to get started?

Join the discussion

Additional resources

Sign up

Login with SSO

Login to the community

Login with SSO

Scanning file for viruses.

This file cannot be downloaded