For a school project i need to build a simple gesture recogniser and send this over Wi-Fi to the user, but for so far I can find, the only two boards that support TinyML are Nano 33 BLE and Nano 33 BLE Sense. Unfortunately they don't have a build in Wi-Fi shield. In short:
Is the Arduino Nano 33 IoT compatible with TinyML models? If not, why?
The Arduino Nano 33 IoT uses a ARM Cortex-M0+ processor @48MHz while the Arduino Nano 33 BLE uses a ARM Cortex-M4 @64MHz.
The Cortex-M4 has a slight advantage per clock cycle for control code plus higher clock frequency.
However for computation there is a substantial difference between the two processors. The Cortex-M4 has SIMD instructions and a hardware single-precision floating point unit (FPU). That means for single-precision floating point (float not double) there is about a 10x higher throughput compared to software floating point depending the compiler and library used. And with the SIMD instructions you can get 2x performance for 16-bit integer computation.