import warnings

# 忽视警告

warnings.filterwarnings('ignore')

import os

import matplotlib

import cv2 as cv

import numpy as np

import matplotlib.pyplot as plt

from tensorflow.keras.callbacks import ModelCheckpoint, ReduceLROnPlateau, EarlyStopping

from tensorflow.keras.applications.imagenet_utils import preprocess_input

from tensorflow.keras import backend as K

from tensorflow.keras.optimizers import Adam

K.image_data_format() == 'channels_last'

from keras_py.utils import get_random_data

from keras_py.face_rec import mask_rec

from keras_py.face_rec import face_rec

from keras_py.mobileNet import MobileNet

from tensorflow.keras.preprocessing.image import ImageDataGenerator

# 数据集路径

basic_path = "./datasets/5f680a696ec9b83bb0037081-momodel/data/"

def letterbox_image(image, size):  # 调整图片尺寸，返回经过调整的照片

     new_image = cv.resize(image, size, interpolation=cv.INTER_AREA)

     return new_image

read_img = cv.imread("test1.jpg")

print(" 调整前图片的尺寸 :", read_img.shape)

read_img = letterbox_image(image=read_img, size=(50, 50))

print(" 调整前图片的尺寸 :", read_img.shape)

def processing_data(data_path, height, width, batch_size=32, test_split=0.1):  # 数据处理， batch_size 默认大小为 32

     train_data = ImageDataGenerator(

             # 对图片的每个像素值均乘上这个放缩因子，把像素值放缩到 0 和 1 之间有利于模型的收敛

             rescale=1. / 255,  

             # 浮点数，剪切强度（逆时针方向的剪切变换角度）

             shear_range=0.1,  

             # 随机缩放的幅度，若为浮点数，则相当于 [lower,upper] = [1 - zoom_range, 1+zoom_range]

             zoom_range=0.1,

             # 浮点数，图片宽度的某个比例，数据提升时图片水平偏移的幅度

             width_shift_range=0.1,

             # 浮点数，图片高度的某个比例，数据提升时图片竖直偏移的幅度

             height_shift_range=0.1,

             # 布尔值，进行随机水平翻转

             horizontal_flip=True,

             # 布尔值，进行随机竖直翻转

             vertical_flip=True,

             # 在 0 和 1 之间浮动。用作验证集的训练数据的比例

             validation_split=test_split  

     )

     # 接下来生成测试集，可以参考训练集的写法

     test_data = ImageDataGenerator(

             rescale=1. / 255,

             validation_split=test_split)

     train_generator = train_data.flow_from_directory(

             # 提供的路径下面需要有子目录

             data_path,

             # 整数元组 (height, width) ，默认： (256, 256) 。 所有的图像将被调整到的尺寸。

             target_size=(height, width),

             # 一批数据的大小

             batch_size=batch_size,

             # "categorical", "binary", "sparse", "input" 或 None 之一。

             # 默认： "categorical", 返回 one-hot 编码标签。

             class_mode='categorical',

             # 数据子集 ("training" 或 "validation")

             subset='training',

             seed=0)

     test_generator = test_data.flow_from_directory(

             data_path,

             target_size=(height, width),

             batch_size=batch_size,

             class_mode='categorical',

             subset='validation',

             seed=0)

     return train_generator, test_generator

# 数据路径

data_path = basic_path + 'image'

# 图像数据的行数和列数

height, width = 160, 160

# 获取训练数据和验证数据集

train_generator, test_generator = processing_data(data_path, height, width)

# 通过属性 class_indices 可获得文件夹名与类的序号的对应字典。

labels = train_generator.class_indices

print(labels)

# 转换为类的序号与文件夹名对应的字典

labels = 外汇跟单gendan5.comdict((v, k) for k, v in labels.items())

print(labels)

pnet_path = "./datasets/5f680a696ec9b83bb0037081-momodel/data/keras_model_data/pnet.h5"

rnet_path = "./datasets/5f680a696ec9b83bb0037081-momodel/data/keras_model_data/rnet.h5"

>

# 加载 MobileNet 的预训练模型权重

weights_path = basic_path + 'keras_model_data/mobilenet_1_0_224_tf_no_top.h5'

# 图像数据的行数和列数

height, width = 160, 160

model = MobileNet(input_shape=[height,width,3],classes=2)

model.load_weights(weights_path,by_name=True)

print(' 加载完成 ...')

def save_model(model, checkpoint_save_path, model_dir):  # 保存模型

     if os.path.exists(checkpoint_save_path):

         print(" 模型加载中 ")

         model.load_weights(checkpoint_save_path)

         print(" 模型加载完毕 ")

     checkpoint_period = ModelCheckpoint(

         # 模型存储路径

         model_dir + 'ep{epoch:03d}-loss{loss:.3f}-val_loss{val_loss:.3f}.h5',

         # 检测的指标

         monitor='val_acc',

         # ‘ auto ’，‘ min ’，‘ max ’中选择

         mode='max',

         # 是否只存储模型权重

         save_weights_only=False,

         # 是否只保存最优的模型

         save_best_only=True,

         # 检测的轮数是每隔 2 轮

         period=2

     )

     return checkpoint_period

checkpoint_save_path = "./results/last_one88.h5"

model_dir = "./results/"

checkpoint_period = save_model(model, checkpoint_save_path, model_dir)

# 学习率下降的方式， acc 三次不下降就下降学习率继续训练

reduce_lr = ReduceLROnPlateau(

                         monitor='accuracy',  # 检测的指标

                         factor=0.5,     # 当 acc 不下降时将学习率下调的比例

                         patience=3,     # 检测轮数是每隔三轮

                         verbose=2       # 信息展示模式

                     )

early_stopping = EarlyStopping(

                             monitor='val_accuracy',  # 检测的指标

                             min_delta=0.0001,         # 增大或减小的阈值

                             patience=3,         # 检测的轮数频率

                             verbose=1            # 信息展示的模式

                         )

# 一次的训练集大小

batch_size = 64

# 图片数据路径

data_path = basic_path + 'image'

# 图片处理

train_generator, test_generator = processing_data(data_path, height=160, width=160, batch_size=batch_size, test_split=0.1)

# 编译模型

model.compile(loss='binary_crossentropy',  # 二分类损失函数   

               optimizer=Adam(lr=0.001),   # 优化器

               metrics=['accuracy'])        # 优化目标

# 训练模型

history = model.fit(train_generator,    

                     epochs=20, # epochs: 整数，数据的迭代总轮数。

                     # 一个 epoch 包含的步数 , 通常应该等于你的数据集的样本数量除以批量大小。

                     steps_per_epoch=637 // batch_size,

                     validation_data=test_generator,

                     validation_steps=70 // batch_size,

                     initial_epoch=0, # 整数。开始训练的轮次（有助于恢复之前的训练）。

                     callbacks=[checkpoint_period, reduce_lr])

# 保存模型

model.save_weights(model_dir + 'temp.h5')

plt.plot(history.history['loss'],label = 'train_loss')

plt.plot(history.history['val_loss'],'r',label = 'val_loss')

plt.legend()

plt.show()

plt.plot(history.history['accuracy'],label = 'acc')

plt.plot(history.history['val_accuracy'],'r',label = 'val_acc')

plt.legend()

plt.show()