【深度学习】tflite模型预测

一、在tensorflow官网下载tflite模型和标签。以分类模型为例：

二、使用下载的模型和标签实现分类预测

  整个项目结构：
                  

import tensorflow as tf
import numpy as np
import cv2 as cv2

#读取标签文件，返回列表
def read_label_list():
    with open('E:\PycharmProject\learning\classification03\labels_mobilenet_quant_v1_224.txt', 'r',encoding="utf8") as f:
        data = f.read().splitlines()
    return data
#图片处理，
def image_process(image_path):
    image=cv2.imread(image_path)
    image=cv2.resize(image,(224,224))
    image=tf.convert_to_tensor(image)
    image=tf.reshape(image,[1,224,224,3])
    image = tf.cast(image, dtype=np.uint8)
    sess = tf.Session()
    sess.run(tf.global_variables_initializer())
    image = image.eval(session=sess)  # 转化为numpy数组
    return image
def main():
    # 加载模型
    interpreter = tf.lite.Interpreter(model_path="E:\PycharmProject\learning\classification03\mobilenet_v1_1.0_224_quant.tflite")
    interpreter.allocate_tensors()
    
    
    # 模型输入和输出细节
    input_details = interpreter.get_input_details()
    #print(input_details)
    #[{'name': 'input', 'index': 88, 'shape': array([  1, 224, 224,   3]), 'dtype': <class 'numpy.uint8'>, 'quantization': (0.0078125, 128)}]
    output_details = interpreter.get_output_details()
    #print(output_details)
    #[{'name': 'MobilenetV1/Predictions/Reshape_1', 'index': 87, 'shape': array([   1, 1001]), 'dtype': <class 'numpy.uint8'>, 'quantization': (0.00390625, 0)}]
    

    #图片传入与处理
    image_path=input('输入图片地址：')
    image=image_process(image_path)

    #模型预测
    interpreter.set_tensor(input_details[0]['index'], image)#传入的数据必须为ndarray类型
    interpreter.invoke()
    output_data = interpreter.get_tensor(output_details[0]['index'])

    #标签预测
    w = np.argmax(output_data)#值最大的位置
    lable_list = read_label_list()#读取标签列表
    print(lable_list[w])

if __name__ == '__main__':
    main()