I recently found out about the STM32Cube AI toolkit, which holds much promises. In the ST blog they specified that it currently only supports mcu's with an FPU but aims to bring support also to fixed-point implementation. It states that a Cortex M4 or higher is currently required. When can we expect support for example Cortex M0?
https://blog.st.com/stm32cubeai-neural-networks/
