Edge AI

Hi,@Miguel CASTRO Related to this article.https://blog.st.com/stm32n6/You mentioned  a machine-vision image signal processor (ISP).do u mean an ISP that can convert the raw output of camera to RGB or YUV? or its something else related to computer vision?Thank you and best RegardsMahmoud Salem

Hello,So far I've been using NN C-files generation and everything works fine. At the moment I'm migrating to a relocatable binary model. At the moment I'm in generation step and for this I've basically added --relocatable and --lib "C:/tools/X-CUBE-AI/8.1.0/Pack/Middlewares/ST/AI/" as arguments in the generate command.However, I'm getting the following error:E600(RelocIOError): Unable to find the PIC lib for stm32h7 in C:\tools\X-CUBE-AI\8.1.0\Pack\Middlewares\ST\AI\Lib\GCC\ARMCortexM7Any ideas why is not getting the NetworkRuntime810_CM7_GCC_PIC.a files? also I would expect it to look for them inside STM32H7 and not ARMCortexM7. Am I missing something?Thanks! 

Hello,Does the AlgoBuiler GUI support STEVAL-MKI109V3 hardware?

Hello,I am looking to use the LSM6DSV16x sensor for position and velocity control. What is the most effective way to achieve the best results? How can I utilize FSM, MLC, and sensor fusion to optimize performance?To clarify, the purposes are:- Trajectory following with mobile robot- Velocity control in both straight and tilted ways.  

I have been able to follow all instructions on the ModelZoo Github page (https://github.com/STMicroelectronics/stm32ai-modelzoo) to successfully download the repository and download all requirements. However, I am getting 2 different errors when I run deploy.py.For context, I was attempting to follow this tutorial: https://github.com/STMicroelectronics/stm32ai-modelzoo/blob/main/object_detection/scripts/deployment/README.mdWhen I set footprints_on_target to false on the user_config.yaml file in order to benchmark my model using the local download of STM32CubeAI, I get the following error:I have followed every other instruction in the tutorial above, making sure that the paths to the model, cube IDE, and STM32CubeAI are correct. However, deploy.py is unable to recognize the STM32CubeAI despite being given the correct path to the executable. I followed the instructions in the tutorial to unzip both the .zip and .pack files that came with the STM32Cube.AI download. I then used STM32C

I am using a STM32WL55JC2 Nucleo Board that mounts a dual core M4/M0+, I changed the linker script in order to have more RAM and FLASH on the M4, but this change is not reflected on the X-CUBE AI interface when adding a new network.Is it possible to do so? Since I am trying to convert a larger network.

Hi,Is it possible to get/study the source code of NetworkRuntime810_CM4_gcc.a? This library is added when X-CUBE-AI is used in CubeMX. I'd like to debug 'ai_platform_network_create' routine. 

Hi, I need help. I checked all the answers to this question in the community, and tried many methods, but still did not solve.when I use the M4 core and validating an ONNX model on target (STM32F747I-DISCO), I got the following error:E200(ValidationError): stm32: Unable to bind the STM AI runtime with "network" c-model: []connection to "serial:115200"/"115200" run-time failsE801(HwIOError): Invalid firmware - COM5:115200​Here are some methods I tried:(1)I use the stm32h747i M7 core successfully "validation on target". (2) when i use the M4 core,I re-selected the cube.ai of the M4 version and reconfigured the usart .It was successful "validation on target" when debugging.But after I run and download successfully,I "validate on target" and got the top error . note:When I reset the development board, the model in the board is not the model just downloaded (downloaded before). Therefore, I think when I using the M4 core to develop, the model is not downloaded into the board(even if it

Hello,After reading the FSM document, I was unable to locate any details on uploading and processing input data for FSM configuration. Could you provide me with more information on this topic?

I have tried to follow the example in the following link to build and deploy the Google tensorflow lite micro speech recognition example on the STM32F469I Discovery board using mbed cli: https://developer.arm.com/solutions/machine-learning-on-arm/developer-material/how-to-guides/build-arm-cortex-m-voice-assistant-with-google-tensorflow-lite/single-pageIn the above link, the example is deployed on the STM32F7 discovery board. To build and compile the micro speech example, you download the Tensorflow lite source code: git clone https://github.com/tensorflow/tensorflow.gitAnd navigate to the project directory and build the example for the target board as follows:cd tensorflowmake -f tensorflow/lite/micro/tools/make/Makefile TARGET=mbed TAGS="disco_f746ng" generate_micro_speech_mbed_project This creates a folder in `tensorflow/lite/micro/tools/make/gen/mbed_cortex-m4/prj/micro_speech/mbed ` that contains the source and header files and Mbed driver files needed to compile the project for th

Hello,Can we utilize any kind of external sensors such as optical flow or range sensors, in addition to a magnetometer or pressure sensor?

Hello,Could you please provide guidance on utilizing LSM6DSV as both a master and a slave in sensor hub mode simultaneously? Kindly refer to the attached image. 

I need help. I am trying to validate on target : STM32L476RG nucleo board.E200(ValidationError): stm32: Unable to bind the STM AI runtime with "network" c-model: []  connection to "serial:COM4:115200"/"COM4:115200" run-time fails  E801(HwIOError): Invalid firmware - COM4:115200 No model runs.

Starting AI validation on target with random data... C:/Users/Admin/STM32Cube/Repository/Packs/STMicroelectronics/X-CUBE-AI/7.0.0/Utilities/windows/stm32ai validate --name network_1627924896993 -m E:/Firmstart/Comtustec/Client_Software/model (1).h5 --type keras --compression 1 --verbosity 1 --workspace C:\Users\Admin\AppData\Local\Temp\mxAI_workspace17156337072000006269500876124151341 --output C:\Users\Admin\.stm32cubemx --mode stm32 --desc COM4:115200  Neural Network Tools for STM32AI v1.5.1 (STM.ai v7.0.0-RC8) Setting validation data...  generating random data, size=10, seed=42, range=default  I[1]: (10, 30, 32, 1)/float32, min/max=[0.000, 1.000], mean/std=[0.494, 0.288], input_0  No 

 I am unable to analyse a trained .onnx model (attached file) in X-Cube-AI Ver 8.0.1. I get an "INTERNAL ERROR: 'label', can someone help?

Good morning, I'm a graduate school student. I'm using STM32CubeAI for inferencing my model. I successfully analyze my model. However, in "Validate on desktop" process, I can see"Assertion failed: n_channel_in == ( (ai_shape_dimension*)((((&((t_scale)->shape))))->data) )[(0x1)], file C:\local\Projects\stm.ai\new_root\targets\common\EmbedNets\src\layers\layers_norm_generic_float.c, line 206" error.My model's input is (1x3x2048) and it consists of only BatchNorm1d(3). I hope you answer my question. Thank you. 

Hello,This article explains how to use sensor fusion and FSM in the UNICO-GUI. How can we utilize FSM and sensor fusion features in Arduino? Could you please clarify how we can effectively utilize unico-GUI in our diverse applications with Arduino? 

Hi,I made an anomaly detector using Nucleo-L412RB (Stm32l412rbt6p) dev board and your nano edge ai library and it was working perfectly(Great job!). After that I made an custom pcb using stm32l412rbt6 mcu mistakenly instead of Stm32l412rbt6p mcu. Now I cant run ml library into it because of different mcu.Please provide me access to nano-edge-ai library for Stm32l412rbt6 so that I can run ml libraries into it and pitch my product.Thanks

I am loading a .onnx model and after i click the analyze button, i get a error pop-up. Is it because cube MX doesn't run on M1 mac?

Hello I have STEVAL-STWINKT1B device. When i try to connect wifi with tera term it gives this error:platform_init L#234 The network interface can not be startedany advice?

HelloI am using NUCLEO-F746ZG. SPI1 and ETH are set with CubeMX and programmed with EWARM.SPI1 will not work if there is an ETH setting ”MX_ETH_Init();”. Do you know the reason?

We are pleased to announce that the latest version of STM32Cube.AI is now available.You can download it here. What is new in X-CUBE-AI 8.1.0? TensorFlow 2.12 and Keras.io 2.12 ONNX Runtime 1.15.0 Host PC system: adding support for MacOS ARM system Include new version of the TensorFlow Lite for Microcontrollers library (SHA-1: 6a1803) Improvements/enhancements Improved inference time for quantized convolutional kernels UI: add specific checkbox allowing to set the “–no-onnx-io-transpose” option CLI: enhance global model format detection, improve the report of the number of operation per layer Core - Code generator optimization of the quantized kernel implementations: CONV2D (151774), ADD (153931) add support for new layers: Einsum (bz121864), BitShift / ONNX Operators Support to the gather with second input variable (TFLITE & ONNX Operators) Support to LSTM with constant inputs not put in initializers (141640) Major bug fixes (129411) TFLITE - support for transpose of batch

Hello, I would like to know what possibilities the STEVAL-MKI109V3 has for connecting to other motherboards. From what I understand, this board (STEVAL-MKI109V3) is used a bridge between a sensor, for example, the LSM6DSOX, and a PC, but I would like to know if it can also be used as a bridge between the sensor and another motherboard in order to develop, for example, a solution involving IoT.

Hello,I am currently utilizing the LSM6DSV16x kit alongside an MKI109V3 evaluation board and UNICO-GUI. I am wondering if there is a way to use UNICO-GUI with wireless communication, as I need to gather data. Any suggestions or advice would be greatly appreciated. 

Hello,I have recently conducted a test on the activity recognition task using the ST library with ESP32 MCU. Although it functions correctly, there are noticeable delays before action recognition occurs. I am seeking advice on how to address this issue.