Ask questions and find answers on how to deploy, debug, and optimize AI models on ST microcontrollers, microprocessors, and smart sensors.
I have been following the FP-AI-VISION1: Computer Vision application video tutorial by Thibaut Vercueli and he uses Google Colab and a notebook called "TransferLearning" to train a new Tensor Flow Lite model on custom images, before loading this into...
The standard way to generate a header file with MLC register settings is to generate from decision-tree .txt file (exportet from WEKA or python).In our project we want to avoid tool interaction.The register settings (ucf or header file) should be gen...
I am working on the deployment of the CNN model on an STM32F29 board with 8-bits precision.I am able to generate the validation code, but the problem is that the model data are in u64 format (as illustrated in the image below)please help?thanks,
Following is my data , I am trying to validate on option of 'validate on desktop'. I am getting following errorLoading: /home/censdata/workspace/har/har_ai/har_ai.ioc projectInitializing: STM32F429ZITx [AI:network] Analyzing model [AI:Generator] Comm...
Hello,I used the example DCMI_CaptureMode for the STM32F769I-EVAL and the CubeMX generator with CubeAI in order to make an object classifier on board. It worked well with the internal RAM for 28*28*1 pictures but now that I want it to work for 128*12...
I'm trying to deploy a machine learning model to an STM32H743ZIT6 through cube-ai extension. The model is in .tflite format. When I try to analyze the model this is the result:The tool states that the MCU has 512KB total available and the model excee...
I convert .ucf file to .h by using unico GUI. But I couldn't set the register values via HAL I2C functions. I couldn't get any decision tree result./* ****************************************************************************** * @file LSM6DSO...
FP-AI-VISION1 has many interesting examples. However, all of them are based on STM32H747I-DISCO. Can anyone provide some instructions and ideas about how to extend the examples in FP-AI-VISION1 to other boards, such as STM32H7B3I or even STM32F7 and ...
This talk at TinyML Summit explains some of the key techniques for optimizing CNN models. The focus in on audio models, but some also apply to other multi-variate time-series and some to image processing.https://www.youtube.com/watch?v=ks5kq1R0aws&t=...
