update model in sample codes

[RanjithRemasan]

Hi 

I have downloaded some face detection sample code from the following repository. My doubt is that how can i replace the models in this project opened in cube IDE. which are the files need to be replaced and what are the steps need to be performed.

GitHub - STMicroelectronics/STM32N6-GettingStarted-FaceDetection: AI software application package demonstrating simple implementation of face detection use case on STM32N6 products.​

Best Regards

Ranjith Remasan

[Julian E.]

Hi @RanjithRemasan,

 

It is explained here:

STM32N6-GettingStarted-FaceDetection/Doc/Deploy-your-Quantized-Model.md at main · STMicroelectronics/STM32N6-GettingStarted-FaceDetection · GitHub

 

Have a good day,

Julian

[RanjithRemasan]

Hi @Julian E. 

Thank you for the input. 

I’m following the document you shared, but I’m having a little trouble locating the st_ai_output/ folder within the project. It doesn't seem to be there—is there a specific step I should take to generate that directory

Project i have selected is : (GitHub - STMicroelectronics/STM32N6-GettingStarted-FaceDetection: AI software application package demonstrating simple implementation of face detection use case on STM32N6 products.​). 

 

I couldn't find st_ai_output/, but I do see a Model/ folder. Is this where the generated files are stored in this version of the project?

Also, how can I start a edgeAI project from scratch to run my own image classification model on STM32N6570-DK?

[Julian E.]

Hi @RanjithRemasan,

 

You are trying to replace the model from the default binary firmware.

To do that, you need the st edge ai core v3.0 as stated here: GitHub - STMicroelectronics/STM32N6-GettingStarted-FaceDetection: AI software application package demonstrating simple implementation of face detection use case on STM32N6 products.​

 

After installing the st edge ai core, you need to generate the C code model based on a python model.

I would suggest that you first try to do that with the original model in /model: blazeface_front_128_quant_pc_ff_od_wider_face.tflite

 

the folder st_ai_output is the result of the generate command using the core.

 

After getting this folder, please continue with the rest of the tutorial.

 

Once you can successfully deploy this model, you can try to deploy yours.

Take note that depending on your model input and output, you may have to edit the application code.

 

Have a good day,

Julian