如何修复分类分类中的"val_accuracy: 0.0000e+00"？

Question

我是深度学习的新手，我有3个类要分类，当我训练我的模型时，我观察到我的"val_loss>val_accuracy">表示我的模型过拟合，我该如何修复？我也得到了"val_accurcy：0.0000e+00〃这样。起初，我一直保持低调。我有更少的数据来训练一个模型。

import tensorflow as tf           
from keras.models import Sequential
from keras.layers import Convolution2D
from keras.layers import MaxPooling2D
from keras.layers import Flatten
from keras.layers import Dropout
from keras.layers import Dense
from keras.preprocessing.image import ImageDataGenerator, image
import numpy as np
import os
from pathlib import Path
country  = "armenia"
cwd = os.getcwd()
print("cwd",cwd)
save_path = r'E://paymentz//'+country+'/'
abc  = os.listdir(r'E:/paymentz/'+country+'/training')
print("list of subfolders in directory:",abc)
model = Sequential()
model.add(Convolution2D(16, 2, 2, input_shape = ( 64, 64, 3), activation = 'relu'))
model.add(MaxPooling2D(pool_size = (2,2)))
model.add(Dropout(0.5))
model.add(Convolution2D(32, 3, 3, activation = 'relu'))
model.add(MaxPooling2D(pool_size = (2,2)))
model.add(Flatten())
model.add(Dropout(0.5))
model.add(Dense(output_dim= 64, activation='relu' ))
output_dim = os.listdir(r'E:/paymentz/'+country+'/training')
print(len(output_dim))
output_dim = len(output_dim)
model.add(Dense(output_dim , activation = 'softmax'))
model.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics =['accuracy'])
batch_size = 16
train_datagen = ImageDataGenerator(rescale = 1./255,
                               shear_range = 0.2,
                               zoom_range = 0.2,
                               horizontal_flip = True,
                               rotation_range = 360)
test_datagen = ImageDataGenerator(rescale = 1./255)
training_set = test_datagen.flow_from_directory(r'E:/paymentz/'+country+'/training',
                                            target_size = (64, 64),
                                            batch_size= batch_size,
                                            class_mode='categorical')
test_set = test_datagen.flow_from_directory(r'E:/paymentz/'+country+'/testing',
                                        target_size= (64, 64),
                                        batch_size= batch_size,
                                        class_mode='categorical')
training_path = Path(fr'E://paymentz//{country}//training')
training_png_count = len(list(training_path.rglob('*.png')))
training_jpg_count = len(list(training_path.rglob('*.jpg')))
training_jpeg_count = len(list(training_path.rglob('*.jpeg')))
          
training_count = training_png_count + training_jpg_count + training_jpeg_count
print("training_count ", training_count)
            
testing_path = Path(fr'E://paymentz//{country}//testing')
testing_png_count = len(list(testing_path.rglob('*.png')))
testing_jpg_count = len(list(testing_path.rglob('*.jpg')))
testing_jpeg_count = len(list(testing_path.rglob('*.jpeg')))
            
testing_count = testing_png_count + testing_jpg_count + testing_jpeg_count
print("testing_count ", testing_count)              
steps_per_epoch = (training_count// batch_size )
print("steps_per_epoch", steps_per_epoch)
validation_steps = ( testing_count // batch_size )
print("validation_steps", validation_steps) 
    
model.fit_generator(
      training_set,
      validation_data = test_set,
      samples_per_epoch = training_count, 
      epochs = 15,
      validation_steps = validation_steps,
      steps_per_epoch = steps_per_epoch)
print("training done.")
score = model.fit(test_set)
score= model.evaluate_generator(test_set)
print("test_set ",score)
#score= model.evaluate_generator(training_set)
#print("training_set ", score)
#if score[0] < 0.05 and score [1] < .85:
save_path = r'E:/paymentz/'+country+'/'
model.save(save_path+country+'.model')
model.save(save_path+country+'.model.h5')
#model.save("StatewiseDLmodel.model.h5")
print("model saved")
abc  = os.listdir(r'E:/paymentz/'+country+'/training')
model_path = r''+country+'model.h5'
#model = tf.keras.models.load_model(country+'.model.h5')
print("model trained to:",score)```

Answer 1

为什么要在测试集上再次拟合模型？

print("training done.")
score = model.fit(test_set) # this should be predict.
score= model.evaluate_generator(test_set)
print("test_set ",score)

请删除以下行，然后重试。

score = model.fit(test_set)