Glass Classification Using Deep Learning¶

In this project, we will examine the data and build a deep neural network that will classify glass based upon certain features.¶

!wget https://cainvas-static.s3.amazonaws.com/media/user_data/cainvas-admin/glass.csv

--2021-07-14 17:24:00--  https://cainvas-static.s3.amazonaws.com/media/user_data/cainvas-admin/glass.csv
Resolving cainvas-static.s3.amazonaws.com (cainvas-static.s3.amazonaws.com)... 52.219.62.32
Connecting to cainvas-static.s3.amazonaws.com (cainvas-static.s3.amazonaws.com)|52.219.62.32|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10053 (9.8K) [text/csv]
Saving to: ‘glass.csv’

glass.csv           100%[===================>]   9.82K  --.-KB/s    in 0s      

2021-07-14 17:24:00 (180 MB/s) - ‘glass.csv’ saved [10053/10053]

Importing the python libraries and packages¶

import pandas as pd
import numpy as np
from sklearn.preprocessing import MinMaxScaler
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
import matplotlib.pyplot as plt 
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense, Dropout, BatchNormalization
from tensorflow.keras.optimizers import Adam
from sklearn.metrics import confusion_matrix,classification_report
from tensorflow.keras.utils import to_categorical, normalize

Reading the CSV file of the dataset¶

Pandas read_csv() function imports a CSV file (in our case, ‘glass.csv’) to DataFrame format.

df=pd.read_csv('glass.csv')

Exploring the Data¶

After importing the data, let's learn more about the dataset.

df.head()

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 214 entries, 0 to 213
Data columns (total 10 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   RI      214 non-null    float64
 1   Na      214 non-null    float64
 2   Mg      214 non-null    float64
 3   Al      214 non-null    float64
 4   Si      214 non-null    float64
 5   K       214 non-null    float64
 6   Ca      214 non-null    float64
 7   Ba      214 non-null    float64
 8   Fe      214 non-null    float64
 9   Type    214 non-null    int64  
dtypes: float64(9), int64(1)
memory usage: 16.8 KB

df.describe()

Null check¶

df.isna().sum()

RI      0
Na      0
Mg      0
Al      0
Si      0
K       0
Ca      0
Ba      0
Fe      0
Type    0
dtype: int64

There’s is no null value in a dataset.

Duplicate Check¶

df[df.duplicated()]

Dropping Duplicate¶

There’re multiple ways to deal with the duplicate records but we have adopted the approach by keeping the last rows and drooping the rows which occurred first in the dataset.

df.drop_duplicates(keep='last',inplace=True)

Pairplot¶

Pairplot shows the relations pairwise among features.

corr_mat=df.corr()
plt.figure(figsize=(16,10))
sns.heatmap(corr_mat,annot=True,fmt='.2f',alpha = 0.7,   cmap= 'coolwarm')
plt.show()

Classes Distribution¶

plt.figure(figsize=(10,10))
sns.countplot(x='Type', data=df, order=df['Type'].value_counts().index)

<AxesSubplot:xlabel='Type', ylabel='count'>

The distribution shows us that the data is imbalanced

Data Manipulation¶

Usually the normalization is performed to bring down all the features on the same scale. By brining down all the features to same scale benefit is that model treat each feature as same.

X=df.drop('Type',axis=1)
X=normalize(X)
y=df['Type']

Class Balancing¶

As above from Distribution of class we can see that the classes are imbalance so if we develop the model of unbalance dataset the model will bias towards the class containing most of the samples so dealing with imbalance classes will help in developing fair model

from imblearn.over_sampling import RandomOverSampler
ros = RandomOverSampler(random_state=42)
x_ros, y_ros = ros.fit_resample(X, y)

Preparing the Data¶

X_train, X_test, y_train, y_test = train_test_split(x_ros,y_ros,test_size=0.2,random_state=42)
y_train=to_categorical(y_train)
y_test=to_categorical(y_test)
print('X_train :',X_train.shape)
print('y_train :',y_train.shape)
print('X_test :',X_test.shape)
print('y_test :',y_test.shape)

X_train : (364, 9)
y_train : (364, 8)
X_test : (92, 9)
y_test : (92, 8)

Building the model¶

We create a Sequential model and add layers one at a time until we are happy with our network architecture.

model = tf.keras.models.Sequential([
       
    tf.keras.layers.Dense(256, input_shape=(9,), activation='relu'),
    tf.keras.layers.BatchNormalization(),
    tf.keras.layers.Dropout(0.3),
    
    
    tf.keras.layers.Dense(256, activation='relu'),
    tf.keras.layers.BatchNormalization(),
    tf.keras.layers.Dropout(0.3),
    
    tf.keras.layers.Dense(512, activation='relu'),
    tf.keras.layers.BatchNormalization(),
    tf.keras.layers.Dropout(0.5),
            
    tf.keras.layers.Dense(8, activation='softmax')
])

Compile the Model¶

Now that the model is defined, we can compile it.

model.compile(loss='categorical_crossentropy',
              optimizer=Adam(0.0001),
              metrics=['acc'])

Model Summary¶

model.summary()

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense (Dense)                (None, 256)               2560      
_________________________________________________________________
batch_normalization (BatchNo (None, 256)               1024      
_________________________________________________________________
dropout (Dropout)            (None, 256)               0         
_________________________________________________________________
dense_1 (Dense)              (None, 256)               65792     
_________________________________________________________________
batch_normalization_1 (Batch (None, 256)               1024      
_________________________________________________________________
dropout_1 (Dropout)          (None, 256)               0         
_________________________________________________________________
dense_2 (Dense)              (None, 512)               131584    
_________________________________________________________________
batch_normalization_2 (Batch (None, 512)               2048      
_________________________________________________________________
dropout_2 (Dropout)          (None, 512)               0         
_________________________________________________________________
dense_3 (Dense)              (None, 8)                 4104      
=================================================================
Total params: 208,136
Trainable params: 206,088
Non-trainable params: 2,048
_________________________________________________________________

Fitting the Model¶

We have defined our model and compiled it ready for efficient computation.

history = model.fit(X_train, y_train,
                    epochs=400,
                    validation_data=(X_test, y_test),
                    verbose=1,
                   )

Epoch 1/400
12/12 [==============================] - 0s 16ms/step - loss: 3.3028 - acc: 0.1429 - val_loss: 2.0722 - val_acc: 0.1848
Epoch 2/400
12/12 [==============================] - 0s 4ms/step - loss: 2.5345 - acc: 0.2500 - val_loss: 2.0587 - val_acc: 0.1848
Epoch 3/400
12/12 [==============================] - 0s 3ms/step - loss: 2.1647 - acc: 0.3269 - val_loss: 2.0453 - val_acc: 0.1848
Epoch 4/400
12/12 [==============================] - 0s 3ms/step - loss: 2.0455 - acc: 0.3626 - val_loss: 2.0335 - val_acc: 0.1848
Epoch 5/400
12/12 [==============================] - 0s 3ms/step - loss: 1.8094 - acc: 0.4478 - val_loss: 2.0225 - val_acc: 0.1848
Epoch 6/400
12/12 [==============================] - 0s 4ms/step - loss: 1.6704 - acc: 0.4533 - val_loss: 2.0131 - val_acc: 0.1848
Epoch 7/400
12/12 [==============================] - 0s 3ms/step - loss: 1.5667 - acc: 0.4808 - val_loss: 2.0046 - val_acc: 0.1848
Epoch 8/400
12/12 [==============================] - 0s 3ms/step - loss: 1.5318 - acc: 0.4945 - val_loss: 1.9951 - val_acc: 0.1848
Epoch 9/400
12/12 [==============================] - 0s 3ms/step - loss: 1.3878 - acc: 0.5495 - val_loss: 1.9867 - val_acc: 0.1848
Epoch 10/400
12/12 [==============================] - 0s 3ms/step - loss: 1.3815 - acc: 0.5220 - val_loss: 1.9791 - val_acc: 0.1848
Epoch 11/400
12/12 [==============================] - 0s 4ms/step - loss: 1.4113 - acc: 0.5467 - val_loss: 1.9717 - val_acc: 0.1848
Epoch 12/400
12/12 [==============================] - 0s 4ms/step - loss: 1.2949 - acc: 0.5467 - val_loss: 1.9628 - val_acc: 0.1848
Epoch 13/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2791 - acc: 0.5659 - val_loss: 1.9537 - val_acc: 0.1848
Epoch 14/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2925 - acc: 0.5907 - val_loss: 1.9434 - val_acc: 0.1848
Epoch 15/400
12/12 [==============================] - 0s 3ms/step - loss: 1.3415 - acc: 0.5440 - val_loss: 1.9350 - val_acc: 0.1848
Epoch 16/400
12/12 [==============================] - 0s 3ms/step - loss: 1.3280 - acc: 0.5577 - val_loss: 1.9265 - val_acc: 0.1848
Epoch 17/400
12/12 [==============================] - 0s 4ms/step - loss: 1.2673 - acc: 0.5742 - val_loss: 1.9165 - val_acc: 0.1848
Epoch 18/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2576 - acc: 0.5797 - val_loss: 1.9058 - val_acc: 0.1848
Epoch 19/400
12/12 [==============================] - 0s 4ms/step - loss: 1.1717 - acc: 0.6126 - val_loss: 1.8986 - val_acc: 0.2500
Epoch 20/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1273 - acc: 0.5879 - val_loss: 1.8892 - val_acc: 0.3370
Epoch 21/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2552 - acc: 0.5824 - val_loss: 1.8739 - val_acc: 0.3587
Epoch 22/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1587 - acc: 0.6154 - val_loss: 1.8559 - val_acc: 0.3587
Epoch 23/400
12/12 [==============================] - 0s 4ms/step - loss: 1.1922 - acc: 0.5604 - val_loss: 1.8384 - val_acc: 0.3804
Epoch 24/400
12/12 [==============================] - 0s 4ms/step - loss: 1.1435 - acc: 0.6264 - val_loss: 1.8161 - val_acc: 0.3804
Epoch 25/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2469 - acc: 0.6044 - val_loss: 1.7916 - val_acc: 0.4457
Epoch 26/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0371 - acc: 0.6319 - val_loss: 1.7667 - val_acc: 0.5000
Epoch 27/400
12/12 [==============================] - 0s 3ms/step - loss: 1.2216 - acc: 0.5659 - val_loss: 1.7470 - val_acc: 0.5109
Epoch 28/400
12/12 [==============================] - 0s 4ms/step - loss: 1.1009 - acc: 0.6484 - val_loss: 1.7196 - val_acc: 0.5543
Epoch 29/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0803 - acc: 0.6209 - val_loss: 1.6847 - val_acc: 0.6087
Epoch 30/400
12/12 [==============================] - 0s 4ms/step - loss: 1.2183 - acc: 0.5962 - val_loss: 1.6491 - val_acc: 0.6413
Epoch 31/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0455 - acc: 0.6071 - val_loss: 1.6136 - val_acc: 0.6522
Epoch 32/400
12/12 [==============================] - 0s 4ms/step - loss: 1.1263 - acc: 0.6209 - val_loss: 1.5782 - val_acc: 0.6522
Epoch 33/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0636 - acc: 0.6264 - val_loss: 1.5414 - val_acc: 0.6522
Epoch 34/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1067 - acc: 0.6291 - val_loss: 1.4988 - val_acc: 0.6304
Epoch 35/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0689 - acc: 0.6071 - val_loss: 1.4486 - val_acc: 0.5978
Epoch 36/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1386 - acc: 0.5824 - val_loss: 1.3872 - val_acc: 0.6196
Epoch 37/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0633 - acc: 0.6099 - val_loss: 1.3273 - val_acc: 0.6304
Epoch 38/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1734 - acc: 0.5989 - val_loss: 1.2739 - val_acc: 0.6304
Epoch 39/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0132 - acc: 0.6319 - val_loss: 1.2343 - val_acc: 0.6304
Epoch 40/400
12/12 [==============================] - 0s 4ms/step - loss: 1.0544 - acc: 0.6016 - val_loss: 1.1822 - val_acc: 0.6304
Epoch 41/400
12/12 [==============================] - 0s 4ms/step - loss: 1.0139 - acc: 0.6016 - val_loss: 1.1386 - val_acc: 0.6196
Epoch 42/400
12/12 [==============================] - 0s 3ms/step - loss: 1.1267 - acc: 0.6209 - val_loss: 1.0870 - val_acc: 0.6196
Epoch 43/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9966 - acc: 0.6429 - val_loss: 1.0473 - val_acc: 0.6196
Epoch 44/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9370 - acc: 0.6319 - val_loss: 1.0180 - val_acc: 0.6196
Epoch 45/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0553 - acc: 0.6264 - val_loss: 0.9718 - val_acc: 0.6304
Epoch 46/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0859 - acc: 0.6044 - val_loss: 0.9232 - val_acc: 0.6413
Epoch 47/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9561 - acc: 0.6346 - val_loss: 0.8932 - val_acc: 0.6413
Epoch 48/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0110 - acc: 0.6429 - val_loss: 0.8627 - val_acc: 0.6630
Epoch 49/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0606 - acc: 0.6236 - val_loss: 0.8364 - val_acc: 0.6630
Epoch 50/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9551 - acc: 0.6456 - val_loss: 0.7867 - val_acc: 0.6739
Epoch 51/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9816 - acc: 0.6291 - val_loss: 0.7539 - val_acc: 0.6630
Epoch 52/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9859 - acc: 0.6401 - val_loss: 0.7111 - val_acc: 0.6848
Epoch 53/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0138 - acc: 0.6291 - val_loss: 0.6850 - val_acc: 0.7174
Epoch 54/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9043 - acc: 0.6676 - val_loss: 0.6704 - val_acc: 0.7283
Epoch 55/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9585 - acc: 0.6374 - val_loss: 0.6438 - val_acc: 0.7609
Epoch 56/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9539 - acc: 0.6181 - val_loss: 0.6131 - val_acc: 0.7826
Epoch 57/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8816 - acc: 0.6566 - val_loss: 0.6044 - val_acc: 0.7935
Epoch 58/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0060 - acc: 0.6291 - val_loss: 0.5905 - val_acc: 0.8152
Epoch 59/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9459 - acc: 0.6456 - val_loss: 0.5834 - val_acc: 0.8152
Epoch 60/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9290 - acc: 0.6676 - val_loss: 0.5995 - val_acc: 0.7826
Epoch 61/400
12/12 [==============================] - 0s 3ms/step - loss: 1.0831 - acc: 0.5824 - val_loss: 0.6044 - val_acc: 0.7391
Epoch 62/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9270 - acc: 0.6621 - val_loss: 0.6276 - val_acc: 0.6957
Epoch 63/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8652 - acc: 0.6374 - val_loss: 0.6062 - val_acc: 0.7609
Epoch 64/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8720 - acc: 0.6731 - val_loss: 0.5785 - val_acc: 0.8043
Epoch 65/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8961 - acc: 0.6374 - val_loss: 0.5623 - val_acc: 0.8043
Epoch 66/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9252 - acc: 0.6731 - val_loss: 0.5523 - val_acc: 0.7935
Epoch 67/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8289 - acc: 0.6731 - val_loss: 0.5557 - val_acc: 0.7717
Epoch 68/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9555 - acc: 0.6758 - val_loss: 0.5655 - val_acc: 0.7500
Epoch 69/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9240 - acc: 0.6538 - val_loss: 0.5733 - val_acc: 0.7500
Epoch 70/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8816 - acc: 0.6346 - val_loss: 0.5804 - val_acc: 0.7500
Epoch 71/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9193 - acc: 0.6429 - val_loss: 0.5698 - val_acc: 0.7609
Epoch 72/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9706 - acc: 0.6429 - val_loss: 0.5479 - val_acc: 0.7717
Epoch 73/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9173 - acc: 0.6566 - val_loss: 0.5447 - val_acc: 0.7609
Epoch 74/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9632 - acc: 0.6346 - val_loss: 0.5383 - val_acc: 0.7500
Epoch 75/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9376 - acc: 0.6374 - val_loss: 0.5425 - val_acc: 0.7935
Epoch 76/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9651 - acc: 0.6154 - val_loss: 0.5490 - val_acc: 0.7826
Epoch 77/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8847 - acc: 0.6538 - val_loss: 0.5342 - val_acc: 0.8043
Epoch 78/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8516 - acc: 0.6566 - val_loss: 0.5155 - val_acc: 0.8261
Epoch 79/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8643 - acc: 0.6593 - val_loss: 0.5080 - val_acc: 0.8152
Epoch 80/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8926 - acc: 0.6429 - val_loss: 0.4945 - val_acc: 0.7609
Epoch 81/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9232 - acc: 0.6209 - val_loss: 0.4994 - val_acc: 0.7500
Epoch 82/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8712 - acc: 0.6429 - val_loss: 0.5019 - val_acc: 0.7826
Epoch 83/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8286 - acc: 0.6731 - val_loss: 0.5086 - val_acc: 0.8043
Epoch 84/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8794 - acc: 0.6676 - val_loss: 0.5202 - val_acc: 0.7935
Epoch 85/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9169 - acc: 0.6621 - val_loss: 0.5100 - val_acc: 0.7935
Epoch 86/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8046 - acc: 0.6951 - val_loss: 0.5110 - val_acc: 0.8261
Epoch 87/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8409 - acc: 0.6813 - val_loss: 0.5300 - val_acc: 0.7935
Epoch 88/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8237 - acc: 0.6758 - val_loss: 0.5462 - val_acc: 0.7935
Epoch 89/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8690 - acc: 0.6538 - val_loss: 0.5417 - val_acc: 0.7826
Epoch 90/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8352 - acc: 0.6648 - val_loss: 0.5292 - val_acc: 0.7935
Epoch 91/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8709 - acc: 0.6951 - val_loss: 0.5230 - val_acc: 0.7935
Epoch 92/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9713 - acc: 0.6291 - val_loss: 0.5143 - val_acc: 0.8152
Epoch 93/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8694 - acc: 0.6648 - val_loss: 0.5124 - val_acc: 0.8152
Epoch 94/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8264 - acc: 0.6758 - val_loss: 0.5076 - val_acc: 0.7935
Epoch 95/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8307 - acc: 0.6813 - val_loss: 0.5097 - val_acc: 0.7935
Epoch 96/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8287 - acc: 0.6758 - val_loss: 0.5167 - val_acc: 0.7935
Epoch 97/400
12/12 [==============================] - 0s 4ms/step - loss: 0.9121 - acc: 0.6374 - val_loss: 0.5223 - val_acc: 0.7935
Epoch 98/400
12/12 [==============================] - 0s 3ms/step - loss: 0.9209 - acc: 0.6511 - val_loss: 0.5298 - val_acc: 0.8152
Epoch 99/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8791 - acc: 0.6401 - val_loss: 0.5346 - val_acc: 0.8152
Epoch 100/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8305 - acc: 0.6648 - val_loss: 0.5331 - val_acc: 0.8043
Epoch 101/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8200 - acc: 0.6896 - val_loss: 0.5244 - val_acc: 0.8152
Epoch 102/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7859 - acc: 0.6621 - val_loss: 0.5401 - val_acc: 0.8043
Epoch 103/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8216 - acc: 0.6786 - val_loss: 0.5321 - val_acc: 0.8152
Epoch 104/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7819 - acc: 0.6758 - val_loss: 0.5216 - val_acc: 0.8152
Epoch 105/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7486 - acc: 0.7088 - val_loss: 0.5381 - val_acc: 0.8261
Epoch 106/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7695 - acc: 0.6813 - val_loss: 0.5427 - val_acc: 0.8152
Epoch 107/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7518 - acc: 0.6868 - val_loss: 0.5444 - val_acc: 0.8261
Epoch 108/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8028 - acc: 0.6511 - val_loss: 0.5521 - val_acc: 0.8261
Epoch 109/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8351 - acc: 0.6538 - val_loss: 0.5679 - val_acc: 0.7935
Epoch 110/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8343 - acc: 0.6566 - val_loss: 0.5954 - val_acc: 0.7500
Epoch 111/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7576 - acc: 0.6896 - val_loss: 0.6016 - val_acc: 0.7391
Epoch 112/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8038 - acc: 0.6511 - val_loss: 0.5781 - val_acc: 0.7717
Epoch 113/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7871 - acc: 0.6648 - val_loss: 0.5685 - val_acc: 0.7609
Epoch 114/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8886 - acc: 0.6538 - val_loss: 0.5648 - val_acc: 0.7609
Epoch 115/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7624 - acc: 0.6978 - val_loss: 0.5557 - val_acc: 0.7935
Epoch 116/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7925 - acc: 0.6648 - val_loss: 0.5457 - val_acc: 0.8043
Epoch 117/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8212 - acc: 0.6676 - val_loss: 0.5335 - val_acc: 0.8043
Epoch 118/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7516 - acc: 0.6868 - val_loss: 0.5282 - val_acc: 0.8043
Epoch 119/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8742 - acc: 0.6484 - val_loss: 0.5221 - val_acc: 0.8152
Epoch 120/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6509 - acc: 0.7170 - val_loss: 0.5061 - val_acc: 0.8043
Epoch 121/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8135 - acc: 0.6676 - val_loss: 0.4974 - val_acc: 0.8043
Epoch 122/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7891 - acc: 0.6841 - val_loss: 0.4909 - val_acc: 0.8043
Epoch 123/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7904 - acc: 0.6621 - val_loss: 0.5065 - val_acc: 0.8043
Epoch 124/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7728 - acc: 0.6786 - val_loss: 0.5071 - val_acc: 0.7935
Epoch 125/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8096 - acc: 0.6868 - val_loss: 0.5215 - val_acc: 0.8152
Epoch 126/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8484 - acc: 0.6593 - val_loss: 0.5501 - val_acc: 0.8043
Epoch 127/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7015 - acc: 0.7170 - val_loss: 0.5323 - val_acc: 0.8043
Epoch 128/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7195 - acc: 0.6868 - val_loss: 0.5285 - val_acc: 0.8043
Epoch 129/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7763 - acc: 0.6593 - val_loss: 0.5165 - val_acc: 0.7935
Epoch 130/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8232 - acc: 0.6676 - val_loss: 0.5110 - val_acc: 0.8043
Epoch 131/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7495 - acc: 0.6813 - val_loss: 0.5111 - val_acc: 0.8043
Epoch 132/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7712 - acc: 0.7143 - val_loss: 0.5379 - val_acc: 0.7935
Epoch 133/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7377 - acc: 0.6868 - val_loss: 0.5181 - val_acc: 0.7935
Epoch 134/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7484 - acc: 0.6951 - val_loss: 0.5208 - val_acc: 0.7935
Epoch 135/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7997 - acc: 0.6813 - val_loss: 0.5268 - val_acc: 0.8152
Epoch 136/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7066 - acc: 0.6923 - val_loss: 0.5341 - val_acc: 0.8043
Epoch 137/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7306 - acc: 0.6648 - val_loss: 0.5266 - val_acc: 0.8152
Epoch 138/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7168 - acc: 0.7363 - val_loss: 0.5028 - val_acc: 0.8152
Epoch 139/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7248 - acc: 0.6841 - val_loss: 0.4903 - val_acc: 0.8152
Epoch 140/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7767 - acc: 0.7143 - val_loss: 0.4794 - val_acc: 0.8043
Epoch 141/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7308 - acc: 0.6868 - val_loss: 0.4947 - val_acc: 0.8043
Epoch 142/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7354 - acc: 0.6896 - val_loss: 0.5094 - val_acc: 0.8261
Epoch 143/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7765 - acc: 0.6923 - val_loss: 0.5212 - val_acc: 0.8043
Epoch 144/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7672 - acc: 0.6621 - val_loss: 0.5064 - val_acc: 0.8043
Epoch 145/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8613 - acc: 0.6648 - val_loss: 0.5068 - val_acc: 0.8043
Epoch 146/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7818 - acc: 0.6841 - val_loss: 0.5233 - val_acc: 0.8043
Epoch 147/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6920 - acc: 0.7308 - val_loss: 0.5469 - val_acc: 0.7826
Epoch 148/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6812 - acc: 0.7280 - val_loss: 0.5516 - val_acc: 0.7935
Epoch 149/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7455 - acc: 0.7033 - val_loss: 0.5502 - val_acc: 0.8152
Epoch 150/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8089 - acc: 0.7060 - val_loss: 0.5306 - val_acc: 0.8261
Epoch 151/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7441 - acc: 0.6813 - val_loss: 0.5141 - val_acc: 0.8370
Epoch 152/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7793 - acc: 0.6923 - val_loss: 0.5017 - val_acc: 0.8370
Epoch 153/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7490 - acc: 0.6841 - val_loss: 0.4949 - val_acc: 0.8261
Epoch 154/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7346 - acc: 0.6978 - val_loss: 0.4951 - val_acc: 0.8152
Epoch 155/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6809 - acc: 0.7198 - val_loss: 0.5071 - val_acc: 0.8261
Epoch 156/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6828 - acc: 0.6978 - val_loss: 0.5374 - val_acc: 0.8261
Epoch 157/400
12/12 [==============================] - 0s 3ms/step - loss: 0.8247 - acc: 0.6538 - val_loss: 0.5407 - val_acc: 0.7826
Epoch 158/400
12/12 [==============================] - 0s 4ms/step - loss: 0.8041 - acc: 0.6731 - val_loss: 0.5460 - val_acc: 0.7826
Epoch 159/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6729 - acc: 0.7308 - val_loss: 0.5038 - val_acc: 0.8043
Epoch 160/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7674 - acc: 0.6841 - val_loss: 0.4975 - val_acc: 0.8043
Epoch 161/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7170 - acc: 0.7033 - val_loss: 0.4898 - val_acc: 0.8043
Epoch 162/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7244 - acc: 0.6896 - val_loss: 0.5140 - val_acc: 0.8043
Epoch 163/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7574 - acc: 0.7170 - val_loss: 0.5606 - val_acc: 0.7935
Epoch 164/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7878 - acc: 0.6593 - val_loss: 0.5544 - val_acc: 0.8043
Epoch 165/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7102 - acc: 0.7143 - val_loss: 0.5390 - val_acc: 0.7826
Epoch 166/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7264 - acc: 0.7170 - val_loss: 0.5291 - val_acc: 0.7935
Epoch 167/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7172 - acc: 0.7115 - val_loss: 0.5305 - val_acc: 0.8152
Epoch 168/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7559 - acc: 0.7060 - val_loss: 0.5421 - val_acc: 0.7717
Epoch 169/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6641 - acc: 0.7225 - val_loss: 0.5353 - val_acc: 0.7935
Epoch 170/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6763 - acc: 0.6951 - val_loss: 0.5127 - val_acc: 0.7935
Epoch 171/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7238 - acc: 0.6978 - val_loss: 0.5070 - val_acc: 0.8152
Epoch 172/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7297 - acc: 0.6813 - val_loss: 0.5033 - val_acc: 0.8261
Epoch 173/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6707 - acc: 0.7225 - val_loss: 0.5119 - val_acc: 0.7935
Epoch 174/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7180 - acc: 0.7033 - val_loss: 0.5080 - val_acc: 0.8261
Epoch 175/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7029 - acc: 0.7060 - val_loss: 0.5052 - val_acc: 0.8152
Epoch 176/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7751 - acc: 0.6484 - val_loss: 0.5141 - val_acc: 0.7826
Epoch 177/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6605 - acc: 0.7582 - val_loss: 0.5045 - val_acc: 0.7826
Epoch 178/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6417 - acc: 0.7198 - val_loss: 0.4919 - val_acc: 0.7717
Epoch 179/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7115 - acc: 0.6978 - val_loss: 0.4816 - val_acc: 0.7935
Epoch 180/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6645 - acc: 0.7225 - val_loss: 0.4763 - val_acc: 0.7935
Epoch 181/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7660 - acc: 0.6896 - val_loss: 0.4562 - val_acc: 0.7935
Epoch 182/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6889 - acc: 0.6951 - val_loss: 0.4478 - val_acc: 0.8043
Epoch 183/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7076 - acc: 0.7170 - val_loss: 0.4520 - val_acc: 0.8370
Epoch 184/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7254 - acc: 0.7088 - val_loss: 0.4505 - val_acc: 0.8370
Epoch 185/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7457 - acc: 0.7060 - val_loss: 0.4699 - val_acc: 0.8261
Epoch 186/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7626 - acc: 0.6951 - val_loss: 0.4774 - val_acc: 0.8152
Epoch 187/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6800 - acc: 0.7088 - val_loss: 0.4767 - val_acc: 0.8043
Epoch 188/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6915 - acc: 0.7005 - val_loss: 0.4868 - val_acc: 0.8043
Epoch 189/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7122 - acc: 0.7088 - val_loss: 0.4950 - val_acc: 0.7935
Epoch 190/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7551 - acc: 0.6841 - val_loss: 0.4795 - val_acc: 0.8043
Epoch 191/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6706 - acc: 0.7253 - val_loss: 0.4743 - val_acc: 0.8152
Epoch 192/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7357 - acc: 0.6841 - val_loss: 0.4824 - val_acc: 0.7935
Epoch 193/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6785 - acc: 0.6978 - val_loss: 0.5139 - val_acc: 0.7717
Epoch 194/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6440 - acc: 0.6978 - val_loss: 0.5349 - val_acc: 0.7500
Epoch 195/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7369 - acc: 0.7033 - val_loss: 0.5342 - val_acc: 0.7609
Epoch 196/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6563 - acc: 0.7253 - val_loss: 0.5294 - val_acc: 0.7826
Epoch 197/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6777 - acc: 0.7088 - val_loss: 0.4968 - val_acc: 0.8043
Epoch 198/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6173 - acc: 0.7527 - val_loss: 0.4978 - val_acc: 0.8043
Epoch 199/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6742 - acc: 0.7060 - val_loss: 0.4901 - val_acc: 0.8043
Epoch 200/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6346 - acc: 0.7363 - val_loss: 0.4881 - val_acc: 0.8152
Epoch 201/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6346 - acc: 0.7335 - val_loss: 0.5256 - val_acc: 0.8043
Epoch 202/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6192 - acc: 0.7363 - val_loss: 0.5441 - val_acc: 0.7826
Epoch 203/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6581 - acc: 0.7280 - val_loss: 0.5456 - val_acc: 0.7826
Epoch 204/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7019 - acc: 0.7225 - val_loss: 0.5384 - val_acc: 0.7935
Epoch 205/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6868 - acc: 0.7225 - val_loss: 0.5376 - val_acc: 0.7935
Epoch 206/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7026 - acc: 0.7060 - val_loss: 0.5264 - val_acc: 0.7826
Epoch 207/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7095 - acc: 0.6841 - val_loss: 0.5306 - val_acc: 0.7826
Epoch 208/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7000 - acc: 0.7088 - val_loss: 0.5201 - val_acc: 0.7826
Epoch 209/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6507 - acc: 0.7418 - val_loss: 0.4889 - val_acc: 0.8043
Epoch 210/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5947 - acc: 0.7363 - val_loss: 0.5000 - val_acc: 0.8043
Epoch 211/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7171 - acc: 0.7198 - val_loss: 0.5038 - val_acc: 0.8043
Epoch 212/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6920 - acc: 0.6923 - val_loss: 0.5368 - val_acc: 0.7609
Epoch 213/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6665 - acc: 0.7170 - val_loss: 0.5321 - val_acc: 0.7609
Epoch 214/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6773 - acc: 0.7170 - val_loss: 0.5308 - val_acc: 0.7935
Epoch 215/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6772 - acc: 0.7005 - val_loss: 0.5478 - val_acc: 0.7826
Epoch 216/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6504 - acc: 0.7198 - val_loss: 0.5272 - val_acc: 0.8043
Epoch 217/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6882 - acc: 0.7170 - val_loss: 0.4993 - val_acc: 0.8152
Epoch 218/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6248 - acc: 0.7390 - val_loss: 0.5047 - val_acc: 0.7935
Epoch 219/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6756 - acc: 0.7170 - val_loss: 0.5315 - val_acc: 0.7717
Epoch 220/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5767 - acc: 0.7390 - val_loss: 0.5338 - val_acc: 0.7717
Epoch 221/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6095 - acc: 0.7280 - val_loss: 0.5481 - val_acc: 0.7717
Epoch 222/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6722 - acc: 0.7198 - val_loss: 0.5410 - val_acc: 0.7826
Epoch 223/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6691 - acc: 0.7253 - val_loss: 0.5249 - val_acc: 0.7826
Epoch 224/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6404 - acc: 0.7198 - val_loss: 0.5209 - val_acc: 0.7935
Epoch 225/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6689 - acc: 0.7198 - val_loss: 0.5302 - val_acc: 0.7826
Epoch 226/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6867 - acc: 0.7198 - val_loss: 0.5461 - val_acc: 0.7717
Epoch 227/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6111 - acc: 0.7418 - val_loss: 0.5400 - val_acc: 0.7935
Epoch 228/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6457 - acc: 0.7500 - val_loss: 0.5526 - val_acc: 0.8152
Epoch 229/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6711 - acc: 0.7445 - val_loss: 0.5541 - val_acc: 0.7826
Epoch 230/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7045 - acc: 0.7033 - val_loss: 0.5440 - val_acc: 0.8043
Epoch 231/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6635 - acc: 0.7143 - val_loss: 0.5338 - val_acc: 0.8043
Epoch 232/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6679 - acc: 0.7308 - val_loss: 0.5175 - val_acc: 0.8152
Epoch 233/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7473 - acc: 0.7088 - val_loss: 0.5127 - val_acc: 0.7935
Epoch 234/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5910 - acc: 0.7280 - val_loss: 0.5185 - val_acc: 0.8043
Epoch 235/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6954 - acc: 0.7143 - val_loss: 0.5345 - val_acc: 0.7826
Epoch 236/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5890 - acc: 0.7363 - val_loss: 0.5174 - val_acc: 0.8261
Epoch 237/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6218 - acc: 0.7088 - val_loss: 0.5085 - val_acc: 0.8152
Epoch 238/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6770 - acc: 0.7198 - val_loss: 0.5141 - val_acc: 0.8261
Epoch 239/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5940 - acc: 0.7363 - val_loss: 0.5349 - val_acc: 0.7935
Epoch 240/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6922 - acc: 0.7005 - val_loss: 0.5563 - val_acc: 0.7609
Epoch 241/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6365 - acc: 0.7143 - val_loss: 0.5517 - val_acc: 0.7717
Epoch 242/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6489 - acc: 0.7170 - val_loss: 0.5449 - val_acc: 0.7717
Epoch 243/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6201 - acc: 0.7060 - val_loss: 0.5578 - val_acc: 0.7609
Epoch 244/400
12/12 [==============================] - 0s 4ms/step - loss: 0.7034 - acc: 0.6786 - val_loss: 0.5738 - val_acc: 0.7391
Epoch 245/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6819 - acc: 0.7363 - val_loss: 0.5835 - val_acc: 0.7283
Epoch 246/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6901 - acc: 0.7033 - val_loss: 0.5670 - val_acc: 0.7283
Epoch 247/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6886 - acc: 0.6978 - val_loss: 0.5415 - val_acc: 0.7717
Epoch 248/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6945 - acc: 0.6978 - val_loss: 0.5050 - val_acc: 0.8152
Epoch 249/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6286 - acc: 0.7088 - val_loss: 0.4998 - val_acc: 0.7935
Epoch 250/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6302 - acc: 0.7143 - val_loss: 0.4908 - val_acc: 0.7935
Epoch 251/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6004 - acc: 0.7253 - val_loss: 0.5105 - val_acc: 0.7717
Epoch 252/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6543 - acc: 0.6868 - val_loss: 0.5071 - val_acc: 0.7717
Epoch 253/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6198 - acc: 0.7253 - val_loss: 0.5067 - val_acc: 0.7717
Epoch 254/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6418 - acc: 0.7280 - val_loss: 0.5092 - val_acc: 0.7826
Epoch 255/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6511 - acc: 0.7143 - val_loss: 0.5333 - val_acc: 0.7826
Epoch 256/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6337 - acc: 0.7445 - val_loss: 0.5262 - val_acc: 0.7935
Epoch 257/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5936 - acc: 0.7390 - val_loss: 0.5318 - val_acc: 0.7935
Epoch 258/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6356 - acc: 0.7253 - val_loss: 0.5326 - val_acc: 0.8043
Epoch 259/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6153 - acc: 0.7390 - val_loss: 0.5158 - val_acc: 0.7826
Epoch 260/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6963 - acc: 0.7225 - val_loss: 0.4982 - val_acc: 0.8152
Epoch 261/400
12/12 [==============================] - 0s 3ms/step - loss: 0.7039 - acc: 0.6978 - val_loss: 0.5086 - val_acc: 0.8261
Epoch 262/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6231 - acc: 0.7143 - val_loss: 0.5126 - val_acc: 0.8261
Epoch 263/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6280 - acc: 0.7335 - val_loss: 0.5172 - val_acc: 0.8261
Epoch 264/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6030 - acc: 0.7473 - val_loss: 0.5214 - val_acc: 0.7826
Epoch 265/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6150 - acc: 0.7637 - val_loss: 0.5087 - val_acc: 0.8043
Epoch 266/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6880 - acc: 0.7253 - val_loss: 0.5127 - val_acc: 0.7826
Epoch 267/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6386 - acc: 0.7033 - val_loss: 0.4951 - val_acc: 0.7935
Epoch 268/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5797 - acc: 0.7473 - val_loss: 0.5054 - val_acc: 0.7826
Epoch 269/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5548 - acc: 0.7445 - val_loss: 0.5166 - val_acc: 0.7826
Epoch 270/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6450 - acc: 0.7363 - val_loss: 0.4974 - val_acc: 0.8370
Epoch 271/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5695 - acc: 0.7747 - val_loss: 0.4909 - val_acc: 0.8370
Epoch 272/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6168 - acc: 0.7610 - val_loss: 0.4967 - val_acc: 0.8261
Epoch 273/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5733 - acc: 0.7418 - val_loss: 0.5068 - val_acc: 0.8261
Epoch 274/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5870 - acc: 0.7335 - val_loss: 0.5057 - val_acc: 0.8152
Epoch 275/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6905 - acc: 0.7308 - val_loss: 0.5044 - val_acc: 0.8152
Epoch 276/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6169 - acc: 0.7445 - val_loss: 0.4804 - val_acc: 0.8152
Epoch 277/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5574 - acc: 0.7473 - val_loss: 0.4868 - val_acc: 0.8370
Epoch 278/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6667 - acc: 0.7143 - val_loss: 0.4961 - val_acc: 0.7826
Epoch 279/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6328 - acc: 0.7143 - val_loss: 0.4930 - val_acc: 0.8043
Epoch 280/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6135 - acc: 0.7582 - val_loss: 0.5168 - val_acc: 0.7826
Epoch 281/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6431 - acc: 0.7445 - val_loss: 0.4906 - val_acc: 0.8043
Epoch 282/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6240 - acc: 0.7198 - val_loss: 0.4642 - val_acc: 0.7935
Epoch 283/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5880 - acc: 0.7582 - val_loss: 0.4572 - val_acc: 0.7935
Epoch 284/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6317 - acc: 0.7418 - val_loss: 0.4559 - val_acc: 0.7935
Epoch 285/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6340 - acc: 0.7143 - val_loss: 0.4634 - val_acc: 0.7935
Epoch 286/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6346 - acc: 0.7308 - val_loss: 0.4623 - val_acc: 0.7935
Epoch 287/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6384 - acc: 0.7390 - val_loss: 0.4653 - val_acc: 0.8043
Epoch 288/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6126 - acc: 0.7198 - val_loss: 0.5068 - val_acc: 0.7935
Epoch 289/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6373 - acc: 0.7225 - val_loss: 0.5153 - val_acc: 0.7826
Epoch 290/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6384 - acc: 0.7253 - val_loss: 0.4951 - val_acc: 0.7935
Epoch 291/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6346 - acc: 0.7418 - val_loss: 0.4825 - val_acc: 0.8261
Epoch 292/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5706 - acc: 0.7473 - val_loss: 0.4514 - val_acc: 0.8152
Epoch 293/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6367 - acc: 0.7225 - val_loss: 0.4569 - val_acc: 0.8152
Epoch 294/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5463 - acc: 0.7527 - val_loss: 0.4760 - val_acc: 0.8261
Epoch 295/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6149 - acc: 0.7500 - val_loss: 0.4867 - val_acc: 0.8152
Epoch 296/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6296 - acc: 0.7335 - val_loss: 0.4933 - val_acc: 0.7935
Epoch 297/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6610 - acc: 0.7115 - val_loss: 0.5008 - val_acc: 0.7609
Epoch 298/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6359 - acc: 0.7170 - val_loss: 0.4967 - val_acc: 0.7935
Epoch 299/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6256 - acc: 0.7225 - val_loss: 0.4968 - val_acc: 0.7935
Epoch 300/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5515 - acc: 0.7637 - val_loss: 0.4961 - val_acc: 0.8043
Epoch 301/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5954 - acc: 0.7198 - val_loss: 0.4800 - val_acc: 0.8370
Epoch 302/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5976 - acc: 0.7527 - val_loss: 0.4803 - val_acc: 0.8370
Epoch 303/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5649 - acc: 0.7637 - val_loss: 0.4773 - val_acc: 0.8478
Epoch 304/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6388 - acc: 0.7390 - val_loss: 0.4882 - val_acc: 0.8261
Epoch 305/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5957 - acc: 0.7637 - val_loss: 0.4889 - val_acc: 0.8152
Epoch 306/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5968 - acc: 0.7610 - val_loss: 0.5021 - val_acc: 0.8043
Epoch 307/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6339 - acc: 0.7473 - val_loss: 0.4740 - val_acc: 0.7935
Epoch 308/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5988 - acc: 0.7445 - val_loss: 0.4707 - val_acc: 0.7935
Epoch 309/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5720 - acc: 0.7582 - val_loss: 0.4648 - val_acc: 0.8152
Epoch 310/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6041 - acc: 0.7308 - val_loss: 0.4604 - val_acc: 0.8152
Epoch 311/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6061 - acc: 0.7308 - val_loss: 0.4628 - val_acc: 0.8152
Epoch 312/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5723 - acc: 0.7445 - val_loss: 0.4660 - val_acc: 0.8043
Epoch 313/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6498 - acc: 0.7280 - val_loss: 0.4800 - val_acc: 0.8043
Epoch 314/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6414 - acc: 0.7527 - val_loss: 0.4629 - val_acc: 0.8043
Epoch 315/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6295 - acc: 0.7115 - val_loss: 0.4793 - val_acc: 0.8152
Epoch 316/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5408 - acc: 0.7555 - val_loss: 0.4985 - val_acc: 0.7935
Epoch 317/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5939 - acc: 0.7198 - val_loss: 0.4904 - val_acc: 0.8152
Epoch 318/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5544 - acc: 0.7445 - val_loss: 0.5016 - val_acc: 0.7826
Epoch 319/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6106 - acc: 0.7170 - val_loss: 0.4799 - val_acc: 0.8261
Epoch 320/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6388 - acc: 0.7115 - val_loss: 0.4800 - val_acc: 0.8370
Epoch 321/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6110 - acc: 0.7363 - val_loss: 0.5041 - val_acc: 0.8043
Epoch 322/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5619 - acc: 0.7582 - val_loss: 0.5015 - val_acc: 0.8152
Epoch 323/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6312 - acc: 0.7280 - val_loss: 0.5051 - val_acc: 0.7826
Epoch 324/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5021 - acc: 0.7747 - val_loss: 0.5007 - val_acc: 0.7717
Epoch 325/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6409 - acc: 0.7170 - val_loss: 0.4702 - val_acc: 0.8370
Epoch 326/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6302 - acc: 0.7390 - val_loss: 0.4695 - val_acc: 0.8370
Epoch 327/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5668 - acc: 0.7637 - val_loss: 0.4677 - val_acc: 0.8370
Epoch 328/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5359 - acc: 0.7418 - val_loss: 0.4610 - val_acc: 0.8478
Epoch 329/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5895 - acc: 0.7308 - val_loss: 0.4673 - val_acc: 0.8370
Epoch 330/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5631 - acc: 0.7665 - val_loss: 0.4660 - val_acc: 0.8370
Epoch 331/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6226 - acc: 0.7363 - val_loss: 0.4917 - val_acc: 0.7935
Epoch 332/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6164 - acc: 0.7225 - val_loss: 0.5031 - val_acc: 0.7826
Epoch 333/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5599 - acc: 0.7610 - val_loss: 0.4856 - val_acc: 0.8043
Epoch 334/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6037 - acc: 0.7747 - val_loss: 0.4948 - val_acc: 0.7935
Epoch 335/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5791 - acc: 0.7335 - val_loss: 0.4760 - val_acc: 0.7826
Epoch 336/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5831 - acc: 0.7198 - val_loss: 0.4783 - val_acc: 0.7826
Epoch 337/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5873 - acc: 0.7527 - val_loss: 0.4737 - val_acc: 0.8043
Epoch 338/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5917 - acc: 0.7665 - val_loss: 0.5104 - val_acc: 0.7717
Epoch 339/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5685 - acc: 0.7747 - val_loss: 0.5234 - val_acc: 0.7717
Epoch 340/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6128 - acc: 0.7527 - val_loss: 0.5108 - val_acc: 0.7826
Epoch 341/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5055 - acc: 0.8077 - val_loss: 0.5079 - val_acc: 0.7717
Epoch 342/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5994 - acc: 0.7500 - val_loss: 0.5088 - val_acc: 0.7717
Epoch 343/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5539 - acc: 0.7582 - val_loss: 0.5232 - val_acc: 0.7609
Epoch 344/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5686 - acc: 0.7363 - val_loss: 0.5333 - val_acc: 0.7717
Epoch 345/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5394 - acc: 0.7610 - val_loss: 0.5546 - val_acc: 0.7500
Epoch 346/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5244 - acc: 0.7720 - val_loss: 0.5612 - val_acc: 0.7283
Epoch 347/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5555 - acc: 0.7582 - val_loss: 0.5773 - val_acc: 0.7283
Epoch 348/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5819 - acc: 0.7445 - val_loss: 0.5182 - val_acc: 0.7717
Epoch 349/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5527 - acc: 0.7610 - val_loss: 0.5213 - val_acc: 0.7717
Epoch 350/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5463 - acc: 0.7610 - val_loss: 0.5598 - val_acc: 0.7174
Epoch 351/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5854 - acc: 0.7335 - val_loss: 0.5395 - val_acc: 0.7283
Epoch 352/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5403 - acc: 0.7830 - val_loss: 0.5430 - val_acc: 0.7391
Epoch 353/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5707 - acc: 0.7582 - val_loss: 0.5287 - val_acc: 0.7609
Epoch 354/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5881 - acc: 0.7308 - val_loss: 0.5438 - val_acc: 0.7609
Epoch 355/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5588 - acc: 0.7555 - val_loss: 0.5415 - val_acc: 0.7391
Epoch 356/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5235 - acc: 0.8077 - val_loss: 0.5370 - val_acc: 0.7609
Epoch 357/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5627 - acc: 0.7527 - val_loss: 0.5095 - val_acc: 0.8043
Epoch 358/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5298 - acc: 0.7775 - val_loss: 0.4923 - val_acc: 0.8261
Epoch 359/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5897 - acc: 0.7610 - val_loss: 0.4564 - val_acc: 0.7935
Epoch 360/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5432 - acc: 0.7692 - val_loss: 0.4764 - val_acc: 0.7935
Epoch 361/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5627 - acc: 0.7747 - val_loss: 0.4864 - val_acc: 0.7935
Epoch 362/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5802 - acc: 0.7225 - val_loss: 0.4645 - val_acc: 0.8043
Epoch 363/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5487 - acc: 0.7637 - val_loss: 0.4759 - val_acc: 0.7609
Epoch 364/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5726 - acc: 0.7582 - val_loss: 0.4761 - val_acc: 0.7935
Epoch 365/400
12/12 [==============================] - 0s 3ms/step - loss: 0.6017 - acc: 0.7527 - val_loss: 0.4642 - val_acc: 0.8152
Epoch 366/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5263 - acc: 0.7527 - val_loss: 0.4747 - val_acc: 0.7826
Epoch 367/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6309 - acc: 0.7445 - val_loss: 0.5056 - val_acc: 0.7935
Epoch 368/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5508 - acc: 0.7473 - val_loss: 0.5276 - val_acc: 0.7500
Epoch 369/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5356 - acc: 0.7445 - val_loss: 0.5351 - val_acc: 0.7500
Epoch 370/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5508 - acc: 0.7610 - val_loss: 0.5481 - val_acc: 0.7283
Epoch 371/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5802 - acc: 0.7830 - val_loss: 0.5481 - val_acc: 0.7174
Epoch 372/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5721 - acc: 0.7418 - val_loss: 0.5284 - val_acc: 0.7717
Epoch 373/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5353 - acc: 0.7500 - val_loss: 0.5407 - val_acc: 0.7609
Epoch 374/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5018 - acc: 0.7747 - val_loss: 0.4955 - val_acc: 0.8152
Epoch 375/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5964 - acc: 0.7555 - val_loss: 0.4816 - val_acc: 0.8043
Epoch 376/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6247 - acc: 0.7115 - val_loss: 0.5064 - val_acc: 0.8261
Epoch 377/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5381 - acc: 0.7473 - val_loss: 0.5472 - val_acc: 0.7609
Epoch 378/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5676 - acc: 0.7445 - val_loss: 0.5542 - val_acc: 0.7283
Epoch 379/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5619 - acc: 0.7555 - val_loss: 0.5305 - val_acc: 0.7500
Epoch 380/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5648 - acc: 0.7418 - val_loss: 0.5156 - val_acc: 0.7500
Epoch 381/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6262 - acc: 0.7445 - val_loss: 0.4942 - val_acc: 0.7826
Epoch 382/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5999 - acc: 0.7308 - val_loss: 0.5110 - val_acc: 0.7717
Epoch 383/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5738 - acc: 0.7363 - val_loss: 0.5161 - val_acc: 0.7826
Epoch 384/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5313 - acc: 0.7665 - val_loss: 0.5097 - val_acc: 0.7826
Epoch 385/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5369 - acc: 0.7637 - val_loss: 0.5367 - val_acc: 0.7609
Epoch 386/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5195 - acc: 0.7830 - val_loss: 0.5759 - val_acc: 0.7065
Epoch 387/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5925 - acc: 0.7527 - val_loss: 0.6008 - val_acc: 0.6848
Epoch 388/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5519 - acc: 0.7637 - val_loss: 0.6551 - val_acc: 0.6848
Epoch 389/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5111 - acc: 0.7720 - val_loss: 0.6327 - val_acc: 0.7174
Epoch 390/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5431 - acc: 0.7830 - val_loss: 0.6298 - val_acc: 0.6957
Epoch 391/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5589 - acc: 0.7418 - val_loss: 0.6388 - val_acc: 0.6739
Epoch 392/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5660 - acc: 0.7280 - val_loss: 0.5978 - val_acc: 0.7391
Epoch 393/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5752 - acc: 0.7308 - val_loss: 0.6122 - val_acc: 0.7609
Epoch 394/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5376 - acc: 0.7665 - val_loss: 0.5887 - val_acc: 0.7717
Epoch 395/400
12/12 [==============================] - 0s 4ms/step - loss: 0.6013 - acc: 0.7527 - val_loss: 0.5698 - val_acc: 0.7935
Epoch 396/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5071 - acc: 0.7775 - val_loss: 0.5531 - val_acc: 0.7717
Epoch 397/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5479 - acc: 0.7775 - val_loss: 0.5075 - val_acc: 0.7717
Epoch 398/400
12/12 [==============================] - 0s 4ms/step - loss: 0.5623 - acc: 0.7582 - val_loss: 0.5373 - val_acc: 0.7609
Epoch 399/400
12/12 [==============================] - 0s 3ms/step - loss: 0.5286 - acc: 0.7692 - val_loss: 0.5299 - val_acc: 0.7717
Epoch 400/400
12/12 [==============================] - 0s 3ms/step - loss: 0.4884 - acc: 0.7940 - val_loss: 0.5313 - val_acc: 0.7609

Accuracy and Loss Plots¶

def plot_learningCurve(history, epoch):
  # Plot training & validation accuracy values
  epoch_range = range(1, epoch+1)
  plt.plot(epoch_range, history.history['acc'])
  plt.plot(epoch_range, history.history['val_acc'])
  plt.title('Model accuracy')
  plt.ylabel('Accuracy')
  plt.xlabel('Epoch')
  plt.legend(['Train', 'Val'], loc='upper left')
  plt.show()

  # Plot training & validation loss values
  plt.plot(epoch_range, history.history['loss'])
  plt.plot(epoch_range, history.history['val_loss'])
  plt.title('Model loss')
  plt.ylabel('Loss')
  plt.xlabel('Epoch')
  plt.legend(['Train', 'Val'], loc='upper left')
  plt.show()

plot_learningCurve(history, 400)

Model Evaluation¶

The evaluate() function will return a list with two values. The first will be the loss of the model on the dataset and the second will be the accuracy of the model on the dataset.

model.evaluate(X_test, y_test)

3/3 [==============================] - 0s 1ms/step - loss: 0.5313 - acc: 0.7609

[0.53132563829422, 0.760869562625885]

Confusion matrix¶

Let's look at the confusion matrix.

y_pred = model.predict_classes(X_test)
y_true = np.argmax(y_test, axis = 1)
cf=confusion_matrix(y_true, y_pred)
sns.heatmap(cf, annot=True)

WARNING:tensorflow:From <ipython-input-22-766a1276ead7>:1: Sequential.predict_classes (from tensorflow.python.keras.engine.sequential) is deprecated and will be removed after 2021-01-01.
Instructions for updating:
Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype("int32")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).

<AxesSubplot:>

Classification Report¶

print(classification_report(y_true,y_pred))

              precision    recall  f1-score   support

           1       0.64      0.53      0.58        17
           2       0.90      0.47      0.62        19
           3       0.36      0.80      0.50        10
           5       0.88      1.00      0.93        14
           6       1.00      1.00      1.00        14
           7       1.00      0.89      0.94        18

    accuracy                           0.76        92
   macro avg       0.80      0.78      0.76        92
weighted avg       0.83      0.76      0.77        92

Making predictions¶

y_pred = model.predict_classes(X_test)
y_pred

array([6, 1, 5, 2, 3, 7, 5, 1, 6, 1, 2, 5, 1, 6, 6, 1, 5, 5, 7, 2, 3, 3,
       3, 2, 3, 7, 2, 7, 3, 3, 6, 3, 7, 3, 7, 2, 5, 5, 3, 2, 6, 3, 5, 2,
       1, 3, 3, 3, 7, 5, 3, 1, 3, 5, 6, 7, 1, 6, 1, 7, 5, 3, 3, 7, 5, 1,
       7, 6, 5, 7, 5, 6, 1, 6, 7, 7, 3, 6, 5, 2, 6, 5, 6, 1, 3, 7, 7, 2,
       1, 3, 1, 3])

Now, let's save the model¶

#saving the model
model.save('Glass_Classification.h5')

DeepCC¶

!deepCC Glass_Classification.h5

[INFO]
Reading [keras model] 'Glass_Classification.h5'
[SUCCESS]
Saved 'Glass_Classification_deepC/Glass_Classification.onnx'
[INFO]
Reading [onnx model] 'Glass_Classification_deepC/Glass_Classification.onnx'
[INFO]
Model info:
  ir_vesion : 4
  doc       : 
[WARNING]
[ONNX]: terminal (input/output) dense_input's shape is less than 1. Changing it to 1.
[WARNING]
[ONNX]: terminal (input/output) dense_3's shape is less than 1. Changing it to 1.
WARN (GRAPH): found operator node with the same name (dense_3) as io node.
[INFO]
Running DNNC graph sanity check ...
[SUCCESS]
Passed sanity check.
[INFO]
Writing C++ file 'Glass_Classification_deepC/Glass_Classification.cpp'
[INFO]
deepSea model files are ready in 'Glass_Classification_deepC/' 
[RUNNING COMMAND]
g++ -std=c++11 -O3 -fno-rtti -fno-exceptions -I. -I/opt/tljh/user/lib/python3.7/site-packages/deepC-0.13-py3.7-linux-x86_64.egg/deepC/include -isystem /opt/tljh/user/lib/python3.7/site-packages/deepC-0.13-py3.7-linux-x86_64.egg/deepC/packages/eigen-eigen-323c052e1731 "Glass_Classification_deepC/Glass_Classification.cpp" -D_AITS_MAIN -o "Glass_Classification_deepC/Glass_Classification.exe"
[RUNNING COMMAND]
size "Glass_Classification_deepC/Glass_Classification.exe"
   text	   data	    bss	    dec	    hex	filename
 969805	   3184	    760	 973749	  edbb5	Glass_Classification_deepC/Glass_Classification.exe
[SUCCESS]
Saved model as executable "Glass_Classification_deepC/Glass_Classification.exe"

	RI	Na	Mg	Al	Si	K	Ca	Type
0	1.52101	13.64	4.49	1.10	71.78	0.06	8.75	1
1	1.51761	13.89	3.60	1.36	72.73	0.48	7.83	1
2	1.51618	13.53	3.55	1.54	72.99	0.39	7.78	1
3	1.51766	13.21	3.69	1.29	72.61	0.57	8.22	1
4	1.51742	13.27	3.62	1.24	73.08	0.55	8.07	1

	RI	Na	Mg	Al	Si	K	Ca	Ba	Fe	Type
count	214.000000	214.000000	214.000000	214.000000	214.000000	214.000000	214.000000	214.000000	214.000000	214.000000
mean	1.518365	13.407850	2.684533	1.444907	72.650935	0.497056	8.956963	0.175047	0.057009	2.780374
std	0.003037	0.816604	1.442408	0.499270	0.774546	0.652192	1.423153	0.497219	0.097439	2.103739
min	1.511150	10.730000	0.000000	0.290000	69.810000	0.000000	5.430000	0.000000	0.000000	1.000000
25%	1.516523	12.907500	2.115000	1.190000	72.280000	0.122500	8.240000	0.000000	0.000000	1.000000
50%	1.517680	13.300000	3.480000	1.360000	72.790000	0.555000	8.600000	0.000000	0.000000	2.000000
75%	1.519157	13.825000	3.600000	1.630000	73.087500	0.610000	9.172500	0.000000	0.100000	3.000000
max	1.533930	17.380000	4.490000	3.500000	75.410000	6.210000	16.190000	3.150000	0.510000	7.000000