Edge AI

..

Hello,I need a sample code to implement AI using ONNX file. I have imported ONNX file into my project. Can anyone help me on the next step to use the model in my program for analyzing data.-Regards

Hello,After enabling the cube-ai module, I imported the .onnx file model in the .ioc file. When I click on Analyze option, I get the below error.[8701] Error loading Python lib '/home/hrish/STM32Cube/Repository/Packs/STMicroelectronics/X-CUBE-AI/7.3.0/Utilities/linux/libpython3.7m.so.1.0': dlopen: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.25' not found (required by /home/hrish/STM32Cube/Repository/Packs/STMicroelectronics/X-CUBE-AI/7.3.0/Utilities/linux/libpython3.7m.so.1.0)I verified the files in their respective paths and they were present at the specified location. I am not able to figure out the issue. Could you please assist me with in resolving the above error.-Regards

WE build computer vision application using STM32H7B3. we want use TouchGFX together with xCube-AI but for some reason i can not enable it the CUBE IDE.

Hi Madam,Mister, I generated code using STM32CubeMX for an STM32H7A3 microcontroller, and I have searched through various files such as main.c, but I cannot find the part of the code where the "network_[name]_val_m_outputs_1.csv" file is created. I would like to modify this file programmatically. Could you help me please?Thank you in advance, Best regards,Mary MICHEL

Hi, I want to implement a Keras Model into STM32 by using CubeAI tool. I successfully upload the model to CubeAI but the fact is that CubeAI doesn't fill the inside of the functions such as ai_platform_network_process(). Could anyone suggest me resource? How can learn to fill the functions? How to use ML models in STM32? #STM32CubeIDE​ #[STM32Cube.AI]​ #X-CUBE-AI​ 

Hi ST member,I would like to create an image classification application on STM32 with Teachable Machine like https://wiki.st.com/stm32mcu/wiki/AI:How_to_use_Teachable_Machine_to_create_an_image_classification_application_on_STM32In first time, I create a model with these parameters :After, I select Tensorflow Lite, check Quantized and export my model.I try different parameters.After, I want to add my model in CubeMX (like https://www.youtube.com/watch?v=NDshmSH7WnA&list=PLnMKNibPkDnG9IC5Nl9vJg1CKMAO1kODW&t=1s). I select these parameters (X-Cube-AI version 6.0.0) :Then, I analyze my model. The analysis returns an error. The model is too big :And if I integrate network.c, network_data.c, network.h and network_data.h into FP-AI-VISION1, this project doesn't compile :c:\st\stm32cubeide_1.6.0\stm32cubeide\plugins\com.st.stm32cube.ide.mcu.externaltools.gnu-tools-for-stm32.9-2020-q2-update.win32_1.5.0.202011040924\tools\arm-none-eabi\bin\ld.exe: STM32H747I_DISCO_PersonDetect_Google_CM

I trained a model of reinforcement learning. In practical applications, I want to fine-tune the model on the embedded system according to the different environment. Does the cube-AI support this operation?

ai_input[0].n_batches = 1; 'ai_buffer' {aka 'struct ai_buffer_'} has no member named 'n_batches'network.hAI_DEPRECATED#define AI_NETWORK_IN \ ai_network_inputs_get(AI_HANDLE_NULL, NULL)This comes up invalid initializer error.Any ideas? The initial code compiled ok with the acc data apart from a bug in cubemx where the lsm6dsl.h could not be found.

詳�?��??調�?��?��?��?�際�?�, �?明�?�点�?��?��?��?��?��?��?��?��?�, �?��?�度�?�連絡致�?��?��?��?�。以下�?�3点�?��?��?��?��?�回答を�?�願�?�致�?��?��?�。​NanoedgeAIStudio�?�必�?メモリ�?��?��??ら�?�?時系列データ�?�対応�?��?��?�る�?�?​目安�?�良�?��?��?�?一�?�当�?�り�?�価格を�?��?��?��?��?�?​2月24日(金)�?��?��?��?�回答頂�?��?��?��?�幸�?��?��?�。�?�忙�?��?��?��?��?�??縮�?��?��?�,�?�回答�?�程よ�?�?��??�?�願�?�致�?��?��?�。​​Excuse me.​I would like to know more about performance of NanoEdgeAIStudio​.I would appreciate it if you could tell me three things about it.What memory do you want to need for operation?Could I please support the time series data ? ​How much would it be ?I would like a response by 2 day later, is that possible?​Looking forward to hearing from you.

Hi,I am trying to deploy the MobileNet NN on a NUCLEO, and I am using X-Cube-AI (7.3) software pack.When using the keras model (H5) and opting for compression with different settings like "high", "medium", "low", the size of model is not decreased at all. (By size, I mean compression in weights and activations)Is there any specific reason why the compression is not applied? The logs and analysis report do not mention anything about this.Thanks,Anirudhan

I designed my own network and completed the training on Pytorch. When I was preparing to use STM32cubeAI to implement .onnx file analysis, I encountered an error as shown in the title. But there is no other valuable information to prompt me to repair it.I successfully checked the network structure contained in .onnx file with the help of netron.app. Does this mean that the file is OK? Or did I use operations that cannot be parsed by cubeAI module? Looking forward to getting help!

..

I am trying AI-MONITOR1 using STWIN Starter Kit.I am referring to the URL below.AI:FP-AI-MONITOR1 �?��?��?�?� - STM32MCUSection 2.3.2 states that the application is flashed on the sensor board, but it cannot be programmed because the memory capacity is not enough for the binary.Is it necessary to prepare an external memory? 

Hello,Taking a look at the new STM32Cube.AI Developer Cloud service, here it is said that it has an API for integration, but the API documentation seems to have just login-related functionality.Is the full API available yet?, if not, when will it be?Thanks.

I am trying to use the STM AI by following the instructions in this document.I get the above mentioned error while trying to download it from the CubeMX.I tried downloading it from this website and it simple does not start downloading.I tried downloading it from its Github repository and it gives a network error. Could you please let me know how to resolve this issue?Thank you.

..

Hello everyoneI have the SensorTile board and I am trying to follow the Mooc in FPAISENING1 in order to understand the example with the audio scene clasification:https://www.youtube.com/watch?v=_s9GMJVwYIg&t=6130sOn this video, when the teacher explained the configuration with the CUBEMx used the board B-L475E-IOT01A. The configuration is pre-loaded on the CUBEMx but I have the board STEVAL-STLKT01V1 and this one is not available in CUBEMX, therefore I have to do the configuration manually and I am not an expert.Could anyone send me the .ioc file with the pins and clocks configuration to execute the Audio Scene Clasification like the video do? I understand that someone of ST has this configuration clear because the audio scene clasification can be executed in  STEVAL-STLKT01V1. Thank you very muchKind regards

I am trying to implement hello world application for mcu but I have I have a card other than the suggested card (the card I have NUCLEO U575ZI Q AT2) I'm trying to do it using Cube.al, but I can't get a complete result and the errors I get are from the files created by cube.al in the platform h file. What can I do about it? and how to perform motion detection on STM32L4 IoTnode? I read the article and wanted to try it, but in the requirements section of the article, a different nucleo card is suggested than the one I have. Can I directly implement this project on my own card by following the steps in the article? If I can't, what changes and configurations do I need to make?

When using cube-ai, the following dialog box will pop up, and it will be stuck at 0%, and trying to manually install the GNU tool will not work. How to solve this?

I want integrate the FP-AI-Vision pack into my own custom project based on STM32H7A3.

I used ai_network_get_error() and got the following results:-> AI ai_network_run error - type=19 code=23​What does type and code mean here?Where are the documents related to this?I want to find the reason why the ai_network_run() is not working.

Could you send me the .ioc for the STEVAL-STLCR01V1. If it is not possible how I have to do the cube configuration? If I do not have the .ioc archive is difficult to follow the video example because I can not generate the C code.Thank you

..

I'm using X-CUBE-AI version 7.3. I'm trying to generate code for my .tflite model using STM32Cube.AI runtime.During the analysis of my model I get the following error:INTERNAL ERROR: Mismatch in channel position.Screenshot:Input dimensions of my model:uint8[1,64,1]Output dimensions:uint8[1,2,64]Unfortunately, I am not allowed to share my .tflite model with you, so just an explanation of this error would be very helpful for me.As well as any feedback on what might be potentially wrong with the dimensions, data types, whatever, will be much appreciated.Just want to add that my .tflite model runs normally in TFLite Micro runtime.