This Arduino library is here to simplify the deployment of Tensorflow Lite for Microcontrollers models to Arduino boards using the Arduino IDE.
Including all the required files for you, the library exposes an eloquent interface to load a model and run inferences.
EloquentTinyML is available from the Arduino IDE Library Manager or you can clone this repo in you Arduino libraries folder.
git clone https://github.com/eloquentarduino/EloquentTinyML.gitBe sure you install version 2.4.0 or newer.
To run a model on your microcontroller, you should first have a model.
I suggest you use tinymlgen to complete this step:
it will export your TensorFlow Lite model to a C array ready to be loaded
by this library.
from tinymlgen import port
tf_model = create_tf_network()
print(port(tf_model))#include <EloquentTinyML.h>
#include <eloquent_tinyml/tensorflow.h>
// sine_model.h contains the array you exported from Python with xxd or tinymlgen
#include "sine_model.h"
#define N_INPUTS 1
#define N_OUTPUTS 1
// in future projects you may need to tweak this value: it's a trial and error process
#define TENSOR_ARENA_SIZE 2*1024
Eloquent::TinyML::TensorFlow::TensorFlow<N_INPUTS, N_OUTPUTS, TENSOR_ARENA_SIZE> tf;
void setup() {
Serial.begin(115200);
delay(4000);
tf.begin(sine_model);
// check if model loaded fine
if (!tf.isOk()) {
Serial.print("ERROR: ");
Serial.println(tf.getErrorMessage());
while (true) delay(1000);
}
}
void loop() {
for (float i = 0; i < 10; i++) {
// pick x from 0 to PI
float x = 3.14 * i / 10;
float y = sin(x);
float input[1] = { x };
float predicted = tf.predict(input);
Serial.print("sin(");
Serial.print(x);
Serial.print(") = ");
Serial.print(y);
Serial.print("\t predicted: ");
Serial.println(predicted);
}
delay(10000);
}
Latest version of this library (2.4.0) is compatible with Cortex-M and ESP32 chips and is built starting from:
ESP32 support is stuck at TensorFlow 2.1.1 at the moment.