2023-04-12 01:44 AM
Hi,
I am working on the implementation of a AI algorithm in a heat pump. My first choice was the SPEAr320 MPU: it is outdated though, it doesn't have enough resources and any FPU. Which MPU should I choose for this application ? that is to say inference and training of the AI algo.
2023-04-13 05:10 AM
Hi @Community member ,
I'm not AI expert, but all STM32MP1 products offer SW AI capabilities with Linux.
I think you could try using an existing development board to figure out what will be the inference rate with your AI algorithm.
https://wiki.st.com/stm32mpu/wiki/X-LINUX-AI_OpenSTLinux_Expansion_Package
Btw, the model training is not done on MPU target but rather on a PC during model development phases.
Regards.