CNN 预测不正确
[英]CNN prediction not correct
问题描述
I experimented with German traffic sign classification using keras.
我使用 keras 对德国交通标志分类进行了试验。 I got 99% val_accuracy when the data set is unbalanced.当数据集不平衡时，我得到了 99% 的 val_accuracy。 Then, I checked the F1-score using the classification report:然后，我使用分类报告检查了 F1 分数：
print(classification_report(y_test1, y_pred_bool))
precision recall f1-score support
0 1.00 0.96 0.98 54
1 0.99 0.99 0.99 618
2 0.98 1.00 0.99 499
3 0.98 0.99 0.99 341
4 1.00 0.98 0.99 472
5 0.98 1.00 0.99 572
6 0.99 1.00 0.99 176
7 1.00 0.98 0.99 176
8 1.00 1.00 1.00 98
9 1.00 1.00 1.00 294
10 0.99 1.00 1.00 313
11 0.96 0.98 0.97 56
12 0.99 0.99 0.99 553
13 0.99 0.96 0.97 95
14 1.00 1.00 1.00 87
15 0.98 1.00 0.99 104
16 0.99 1.00 0.99 133
17 1.00 0.99 0.99 67
18 0.98 1.00 0.99 345
19 1.00 1.00 1.00 151
20 1.00 1.00 1.00 50
21 1.00 0.98 0.99 153
22 0.97 0.99 0.98 73
23 0.99 0.99 0.99 350
24 0.99 0.96 0.98 104
25 1.00 1.00 1.00 217
26 1.00 1.00 1.00 54
27 1.00 0.99 0.99 165
28 0.95 0.98 0.96 93
29 0.99 1.00 0.99 275
30 0.98 1.00 0.99 95
31 1.00 1.00 1.00 58
32 1.00 0.99 1.00 535
33 0.98 1.00 0.99 62
34 0.99 1.00 0.99 497
35 1.00 0.98 0.99 100
36 1.00 1.00 1.00 65
37 0.98 0.98 0.98 49
38 1.00 0.94 0.97 446
39 0.98 1.00 0.99 90
40 0.97 1.00 0.99 368
41 1.00 0.99 1.00 337
42 1.00 0.99 1.00 363
accuracy 0.99 9803
macro avg 0.99 0.99 0.99 9803
weighted avg 0.99 0.99 0.99 9803
Then, I checked the confusion matrix, which is correct.然后，我检查了混淆矩阵，这是正确的。
I saved the model on my disk and loaded it again to predict images from every class with the following code: :我将模型保存在我的磁盘上并再次加载它以使用以下代码预测每个类的图像：
model = load_model('/kaggle/working/models-07-0.9904.h5')
pred = model.predict(images1)
print(pred)
y_pred_bool = np.argmax(pred, axis = 1)
print(y_pred_bool)
Unfortunately, most of the predictions are wrong and I don't know why.不幸的是，大多数预测都是错误的，我不知道为什么。 Any suggestions?有什么建议？
EDIT编辑
It has multi - label classification它具有多标签分类
I took first 10 images from every label and make the prediction我从每个标签中获取前 10 张图像并进行预测
The corresponding labels are对应的标签是
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 19, 19, 19, 19, 19, 19, 19, 19, 19, 19, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 23, 23, 23, 23, 23, 23, 23, 23, 23, 23, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 31, 31, 31, 31, 31, 31, 31, 31, 31, 31, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 33, 33, 33, 33, 33, 33, 33, 33, 33, 33, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 36, 36, 36, 36, 36, 36, 36, 36, 36, 36, 37, 37, 37, 37, 37, 37, 37, 37, 37, 37, 38, 38, 38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42]
The y_pred_bool y_pred_bool
array([ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 23, 23, 23, 23,
23, 23, 23, 23, 23, 23, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 38,
38, 38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39,
39, 39, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 41, 41, 41, 41, 41,
41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 2, 2,
2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3,
3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5,
5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7,
7, 7, 7, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8,
9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, 10,
10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 13, 13, 13, 13,
13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 15,
15, 15, 15, 15, 15, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16,
16, 16, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 18, 18, 18, 18, 18,
18, 18, 18, 18, 18, 19, 19, 19, 19, 19, 19, 19, 19, 19, 19, 20, 20,
20, 20, 20, 20, 20, 20, 20, 3, 21, 21, 21, 21, 21, 21, 21, 21, 21,
21, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 24, 24, 24, 24, 24, 24,
24, 24, 24, 24, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 26, 26, 26,
26, 26, 26, 26, 26, 26, 26, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 29, 29, 29, 29, 29, 29, 29,
29, 29, 29, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 31, 31, 31, 31,
31, 31, 31, 31, 31, 31, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 33,
33, 33, 33, 33, 33, 33, 33, 33, 33, 35, 35, 35, 35, 35, 35, 35, 35,
35, 35, 36, 36, 36, 36, 36, 36, 36, 36, 36, 36, 37, 37, 37, 37, 37,
37, 37, 37, 37, 37])
1 个解决方案
解决方案1
0
If you are saving and loading the model and evaluating on the same dataset, the results must not change before saving and after loading the model.如果您正在保存和加载模型并在同一数据集上进行评估，则在保存之前和加载模型之后结果不得更改。
Example: I have run a simple model and saved using model.save and the did the load with load_model of keras.示例：我运行了一个简单的模型并使用 model.save 进行了保存，并使用 keras 的 load_model 进行了加载。 You can download the pima-indians-diabetes.data.csv dataset from here (If not available here, you can download from kaggle and other sources).你可以从这里下载 pima-indians-diabetes.data.csv 数据集（如果这里没有，你可以从 kaggle 和其他来源下载）。
Build, Evaluate and Save the model -构建、评估和保存模型 -
%tensorflow_version 1.x
# MLP for Pima Indians Dataset saved to single file
import numpy as np
from numpy import loadtxt
from keras.models import Sequential
from keras.layers import Dense
# load pima indians dataset
dataset = np.loadtxt("/content/pima-indians-diabetes.csv", delimiter=",")
# split into input (X) and output (Y) variables
X = dataset[:,0:8]
Y = dataset[:,8]
# define model
model = Sequential()
model.add(Dense(12, input_dim=8, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(1, activation='sigmoid'))
# compile model
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
# Model Summary
model.summary()
# Fit the model
model.fit(X, Y, epochs=150, batch_size=10, verbose=0)
# evaluate the model
scores = model.evaluate(X, Y, verbose=0)
print("%s: %.2f%%" % (model.metrics_names[1], scores[1]*100))
# save model and architecture to single file
model.save("model.h5")
print("Saved model to disk")
Output -输出 -
WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/tensorflow_core/python/ops/resource_variable_ops.py:1630: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.
Instructions for updating:
If using Keras pass *_constraint arguments to layers.
WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/tensorflow_core/python/ops/nn_impl.py:183: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.where in 2.0, which has the same broadcast rule as np.where
Model: "sequential_1"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_1 (Dense) (None, 12) 108
_________________________________________________________________
dense_2 (Dense) (None, 8) 104
_________________________________________________________________
dense_3 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.
accuracy: 75.78%
Saved model to disk
Load and Evaluate the Model -加载和评估模型 -
# load and evaluate a saved model
from numpy import loadtxt
from keras.models import load_model
# load model
model = load_model('model.h5')
# summarize model.
model.summary()
# load dataset
dataset = loadtxt("pima-indians-diabetes.csv", delimiter=",")
# split into input (X) and output (Y) variables
X = dataset[:,0:8]
Y = dataset[:,8]
# evaluate the model
score = model.evaluate(X, Y, verbose=0)
print("%s: %.2f%%" % (model.metrics_names[1], score[1]*100))
Output -输出 -
Model: "sequential_1"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_1 (Dense) (None, 12) 108
_________________________________________________________________
dense_2 (Dense) (None, 8) 104
_________________________________________________________________
dense_3 (Dense) (None, 1) 9
=================================================================
Total params: 221
Trainable params: 221
Non-trainable params: 0
_________________________________________________________________
accuracy: 75.78%
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.
相关问题
使用Tensorflow进行CNN预测 - CNN prediction using tensorflow
CNN模型的预测 - Prediction on CNN model
CNN预测遇到麻烦 - Having trouble with CNN prediction
使用 Tensor Flow CNN 进行预测 - Using Tensor Flow CNN for prediction
预测怎么说？ CNN凯拉斯 - What the prediction says? CNN Keras
CNN模型无法做出预测 - CNN model not able to make prediction
CNN-具有多分类的错误预测 - CNN - Wrong prediction with multiclass classification
3 类预测 CNN 深度学习 - 3 Classes prediction CNN Deep Learning
使用 CNN 进行推文情绪预测 - Tweet emotion prediction using CNN
如何提高CNN的预测准确率 - How to increase the prediction accuracy of CNN
相关标签