I work in researcher. In my project, I'm testing Himax WE-1 board.
I am curious about the reason that the version of Tensorflow used in the Arduino IDE is because the current version of Tensorflow does not support the Hello World example of the Himax WE-1 board. However, the Arduino IDE supports the Hello World example of the Himax WE-1 board.
I would like to know the Tensorflow lite micro version of the example built-in on the Arduino.
The latest version of the Arduino_TensorFlowLite is "2.4.0-ALPHA". That seems to match too closely with the versioning of TensorFlow (Tags · tensorflow/tensorflow · GitHub) to be a coincidence, so I'd guess it's from TensorFlow 2.4.0. As for the "ALPHA" part, I have not clue. I see that's on all the versions, so probably just slapped on by the developers as a weasle clause "hey, don't blame me if it doesn't work, it's alpha after all".