我在混淆矩阵上得到的结果很差，但在最后一个时期的 val_set 上训练时准确率达到 99.44% [重复]

Question

我的数据集有两个文件夹：Test 和 Train。 Test 包含 2 个文件夹，其中包含 Covid X 射线和正常 X 射线，Train 文件夹也是如此。在最后一个训练周期中，我在 val_set 上获得了 99.4% 的准确率，但是在绘制混淆矩阵时，我什至没有获得 60% 的准确率。

而且我的 Y_pred 看起来很奇怪，并且确定性不在 0-1 之间，我需要帮助才能在我的混淆矩阵和分类报告上获得相同的 99.4% 结果。

import os
import numpy as np
import matplotlib.pyplot as plt
import keras
from keras.applications import xception
from keras.layers import *
from keras.models import *
from keras.preprocessing import image

model = xception.Xception(weights='imagenet', include_top=False, input_shape=(224, 224, 3))
for layers in model.layers:
    layers.trainable=False
    
flat1 = Flatten()(model.layers[-1].output)
class1 = Dense(256, activation='relu')(flat1)
output = Dense(1, activation='sigmoid')(class1)

model = Model(inputs = model.inputs, outputs = output)


model.compile(loss = 'binary_crossentropy', optimizer='adam', metrics=['accuracy'])


train_datagen = image.ImageDataGenerator(
    rescale = 1./255,
    shear_range = 0.2,
    zoom_range = 0.2,
    horizontal_flip = True,
    )

test_datagen = image.ImageDataGenerator(rescale = 1./255)

train_generator = train_datagen.flow_from_directory(
    '/Users/xd_anshul/Desktop/Research/Major/CovidDataset/Train',
    target_size = (224,224),
    batch_size = 10,
    class_mode='binary')

validation_generator = test_datagen.flow_from_directory(
    '/Users/xd_anshul/Desktop/Research/Major/CovidDataset/Test',
    target_size = (224,224),
    batch_size = 10,
    class_mode='binary')

#model Fitting

hist = model.fit(
    train_generator,
    epochs=2,
    validation_data=validation_generator)



from sklearn.metrics import classification_report, confusion_matrix

Y_pred = model.predict_generator(validation_generator, steps = np.ceil(validation_generator.samples / validation_generator.batch_size), verbose=1, workers=0)
y_pred = [np.where(predictions>0.5, 1, 0) for predictions in Y_pred]
print('Confusion Matrix')
print(confusion_matrix(validation_generator.classes, y_pred))
print('Classification Report')
target_names = ['Covid', 'Normal']
print(classification_report(validation_generator.classes, y_pred, target_names=target_names))

OUTPUT::

Epoch 1/2
255/255 [==============================] - 464s 2s/step - loss: 1.4076 - accuracy: 0.9342 - val_loss: 0.3753 - val_accuracy: 0.9718
Epoch 2/2
255/255 [==============================] - 558s 2s/step - loss: 0.1218 - accuracy: 0.9885 - val_loss: 0.0181 - val_accuracy: 0.9944
Confusion Matrix
[[310 100]
 [ 97  25]]
Classification Report
              precision    recall  f1-score   support

       Covid       0.76      0.76      0.76       410
      Normal       0.20      0.20      0.20       122

    accuracy                           0.63       532
   macro avg       0.48      0.48      0.48       532
weighted avg       0.63      0.63      0.63       532

Y_pred looks like:

1.05698e-14
2.25061e-13
3.96925e-19
5.53114e-17
3.4461e-32
1.2413e-32
2.29092e-28
4.01138e-17
3.81177e-23
1.88817e-07
1.164e-07
.
.
.
(532 Values)

Answer 1

你的问题在于

y_pred = [np.where(predictions>0.5, 1, 0) for predictions in Y_pred]

打印出来，你会看到它是一个数组列表，因为 Y_pred 是一个数组 try

 y_pred=[ np.argmax(Y_pred[i]) for i in range(val_generator.samples)]