TensorFlow Lite Micro 官方文档

Welcome to the official documentation for TensorFlow Lite Micro! This guide provides an in-depth look at how to get started with TensorFlow Lite Micro, which is designed for running machine learning models on microcontrollers.

快速开始

术语解释

TensorFlow Lite Micro: A lightweight version of TensorFlow designed for microcontrollers.
Machine Learning Model: A mathematical model that allows computers to learn from data.

资源链接

示例

Here's a simple example of a TensorFlow Lite Micro model in use:

#include "tensorflow/lite/c/common.h"
#include "tensorflow/lite/micro/all_ops_common.h"
#include "tensorflow/lite/micro/kernels/micro_ops.h"
#include "tensorflow/lite/micro/micro_mutable_op_data.h"
#include "tensorflow/lite/micro/micro_error_reporter.h"
#include "tensorflow/lite/micro/micro_interpreter.h"
#include "tensorflow/lite/micro/micro_allocator.h"
#include "tensorflow/lite/micro/micro_utils.h"
#include "tensorflow/lite/micro/tensor_c.h"
#include "tensorflow/lite/micro/tensor_c_data.h"

// ... (code to initialize and run the model)

// Output tensor
tflite::Tensor* output_tensor = interpreter->GetOutputTensor(0);

// Process the model's output
for (int i = 0; i < output_tensor->dims->size; i++) {
  for (int j = 0; j < output_tensor->dims->data[i]; j++) {
    float value = tflite::micro::GetTensorData<float>(output_tensor)[j];
    // Do something with the value
  }
}