keras missing label unlabeled 未标注类 如何训练 如何自定义损失函数 去除未标注类 缺少标签

我从这里找到的答案：

def ignore_unknown_xentropy(ytrue, ypred):
    return (1-ytrue[:, :, :, 0])*categorical_crossentropy(ytrue, ypred)

然后下面是我的代码：

我感觉看到这个帖子的应该都看得懂吧（其实就是懒得解释了，请看注释） 

from __future__ import print_function
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
import os
import keras
import PIL
from PIL import Image
from keras import Model, Input, optimizers
from keras.applications import vgg16, inception_v3, resnet50, mobilenet
from keras.layers import Conv2D, Lambda, Reshape
from keras.losses import categorical_crossentropy
from keras.preprocessing.image import ImageDataGenerator, load_img

#TODO: make more variable not using number in funtions
def load_Img(imgDir, imgFoldName, k):
    imgs = os.listdir(imgDir + imgFoldName)
    imgNum = len(imgs)