使用Tensorflow2.0和Keras训练分类模型实例

1.训练过程：

from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.applications.xception import Xception
from tensorflow.keras.layers import GlobalAveragePooling2D, GlobalMaxPooling2D, Dense, Flatten, Input, Concatenate, Dropout
from tensorflow.keras.models import Model

import tensorflow as tf
from tensorflow.keras.callbacks import ModelCheckpoint,ReduceLROnPlateau,TensorBoard, EarlyStopping
from sklearn.utils import class_weight
import numpy as np
# 下面的一部分是进行数据预处理
import os
import shutil

def image_number(dir_path):
    path = []
    for dir, file, images in os.walk(dir_path):
        for image in images:
            if image.endswith(".bmp") or image.endswith(".jpg")or image.endswith(".jpeg"):
                imagepath = os.path.join(dir, image)
                path.append(imagepath)
    return path

training_path=r".\training"
eval_path =r".\validation"
test_path=r".\test"
MODEL_path=r".\model"
log_path=r".\log"
training_image_path = image_number(training_path)  # 训练集数据路径列表

val_image_path = image_number(eval_path)
#test_image_path=image_number(test_path)
label=["defect","dust","ok"]
#label=["defect","ok"]
training_image_number = len(training_image_path)  # 训练集数据量
val_image_number = len(val_image_path)
#test_image_number=len(test_image_path)
classfication_number = len(label)  # 分类数
batch=16
train_epoch = (training_image_number // batch)

val_epoch = (val_ima