Hello everyone, when I read the examples code of tflite-micro, I saw that different examples use different values of kTensorArenaSize. What is the basis for determining this value? For example, in hello_world, the value of kTensorArenaSize is 2000, and in magic_wand, the value of kTensorArenaSize is 64*1024.
In the comment, there are such lines, so the only way is to try again and again to get minist value.
// Create an area of memory to use for input, output, and intermediate arrays.
// The size of this will depend on the model you’re using, and may need to be
// determined by experimentation.