I couldn't find any clear answers elsewhere, hence I'm asking here. I have a TensorFlow/TensorFlow Lite model which I have already trained on some sensor data. Now, how can I utilize this model on my Arduino Nano/Uno to perform some other functions like moving a motor?
Any help would be appreciated!
Thanks
The documentation of Tensorflow lite states:
The core runtime just fits in 16 KB on an Arm Cortex M3 and can run many basic models.
So the Tensorflow lite runtime itself needs 16kB of - I guess - RAM. That would be 8 times of what the Uno/Nano actually has. If they mean program memory, then it would fit, since Uno/Nano has 32kB of flash.
Though it would still not work:
TensorFlow Lite for Microcontrollers is written in C++ 17 and requires a 32-bit platform.
The Uno/Nano uses the Atmega328p, which is an 8-bit microcontroller.
So all in all: No, you cannot use Tensorflow lite on an Uno/Nano. You should look at the "Supported Platforms" section of the linked Tensorflow Lite documentation to choose a board, that fits your needs.
