Scribd
Upload
14 views4 pages

Image Classifications

This document loads and preprocesses the CIFAR-10 image dataset. It then defines and trains two neural network models - a simple ANN and a CNN - to classify the images. The ANN achieves 49.9% accuracy over 5 epochs, while the CNN achieves 87.58% accuracy. Classification reports for both models on the test set are also presented.

Uploaded by

Anvesh Silagana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
14 views4 pages

Image Classifications

This document loads and preprocesses the CIFAR-10 image dataset. It then defines and trains two neural network models - a simple ANN and a CNN - to classify the images. The ANN achieves 49.9% accuracy over 5 epochs, while the CNN achieves 87.58% accuracy. Classification reports for both models on the test set are also presented.

Uploaded by

Anvesh Silagana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

import 

tensorflow as tf
from tensorflow.keras import datasets, layers, models
import matplotlib.pyplot as plt
import numpy as np

(X_train, y_train),(X_test,y_test) = datasets.cifar10.load_data()

Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz


170498071/170498071 [==============================] - 4s 0us/step

X_train.shape

(50000, 32, 32, 3)

X_test.shape

(10000, 32, 32, 3)

X_test
[[ 12, 25, 6],
[ 20, 37, 7],
[ 24, 36, 15],
...,
[115, 134, 138],
[149, 168, 177],
[104, 117, 131]],

[[ 12, 25, 11],


[ 15, 29, 6],
[ 34, 40, 24],
...,
[154, 172, 182],
[157, 175, 192],
[116, 129, 151]],

...,

[[100, 129, 81],


[103, 132, 84],
[104, 134, 86],
...,
[ 97, 128, 84],
[ 98, 126, 84],
[ 91, 121, 79]],

[[103, 132, 83],


[104, 131, 83],
[107, 135, 87],
...,
[101, 132, 87],
[ 99, 127, 84],
[ 92, 121, 79]],

[[ 95, 126, 78],


[ 95, 123, 76],
[101, 128, 81],
...,
[ 93, 124, 80],
[ 95, 123, 81],
[ 92, 120, 80]]],

[[[ 73, 78, 75],


[ 98, 103, 113],
[ 99, 106, 114],
...,
[135, 150, 152],
[135, 149, 154],
[203, 215, 223]],

[[ 69, 73, 70],


[ 84, 89, 97],
[ 68, 75, 81],
...,
[ 85, 95, 89],
[ 71, 82, 80],
[120, 133, 135]],

plt.imshow(X_train[1])
<matplotlib.image.AxesImage at 0x7fb5ea7f0c10>

print(y_train[:5])

[[6]
[9]
[9]
[4]
[1]]

y_train = y_train.reshape(-1,)
y_train[:5]

array([6, 9, 9, 4, 1], dtype=uint8)

classes =["airplane" , "automobile" , "bird" , "cat", "deer", 'dog','frog',"horse" "ship" , "truck"]

def plot_samples(X,y, index):
  plt.figure(figsize=(15, 2))
  plt.imshow(X[index])
  plt.xlabel(classes[y[index]])

plot_samples(X_train,y_train,0)

X_train = X_train/255
X_test = X_test/255

ann_model = models.Sequential([
    layers.Flatten(input_shape=(32,32,3)),
    layers.Dense(3000,activation='relu'),
    layers.Dense(1000,activation='relu'),
    layers.Dense(10,activation='sigmoid')
])

ann_model.compile(optimizer='SGD',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

ann_model.fit(X_train, y_train, epochs=5)

Epoch 1/5
1563/1563 [==============================] - 13s 5ms/step - loss: 1.8142 - accuracy: 0.3542
Epoch 2/5
1563/1563 [==============================] - 7s 5ms/step - loss: 1.6250 - accuracy: 0.4262
Epoch 3/5
1563/1563 [==============================] - 7s 4ms/step - loss: 1.5436 - accuracy: 0.4563
Epoch 4/5
1563/1563 [==============================] - 7s 5ms/step - loss: 1.4811 - accuracy: 0.4796
Epoch 5/5
1563/1563 [==============================] - 7s 4ms/step - loss: 1.4345 - accuracy: 0.4949
<keras.callbacks.History at 0x7fb56c106580>

from sklearn.metrics import confusion_matrix, classification_report
import numpy as np
y_pred = ann_model.predict(X_test)
y_pred_classes = [np.argmax(element) for element in y_pred]
print("Classification Report:", classification_report(y_test,y_pred_classes))

313/313 [==============================] - 1s 2ms/step


Classification Report: precision recall f1-score support

0 0.66 0.41 0.51 1000


1 0.62 0.61 0.61 1000
2 0.35 0.41 0.38 1000
3 0.36 0.29 0.32 1000
4 0.57 0.24 0.34 1000
5 0.41 0.40 0.41 1000
6 0.45 0.68 0.54 1000
7 0.66 0.44 0.53 1000
8 0.48 0.76 0.59 1000
9 0.50 0.61 0.55 1000

accuracy 0.49 10000


macro avg 0.51 0.49 0.48 10000
weighted avg 0.51 0.49 0.48 10000

cnn_model = models.Sequential([
    
    layers.Conv2D(filters=32, kernel_size=(3,3), activation='relu', input_shape=(32,32,3)),
    layers.MaxPool2D((2,2)),

    layers.Flatten(),
    layers.Dense(3000,activation='relu'),
    layers.Dense(10,activation='softmax')
])

cnn_model.compile(optimizer='Adam',
            loss='sparse_categorical_crossentropy',
            metrics=['accuracy'])

cnn_model.fit(X_train,y_train,epochs=5,batch_size=32)

Epoch 1/5
1563/1563 [==============================] - 18s 8ms/step - loss: 1.3959 - accuracy: 0.5071
Epoch 2/5
1563/1563 [==============================] - 12s 7ms/step - loss: 1.0267 - accuracy: 0.6390
Epoch 3/5
1563/1563 [==============================] - 12s 7ms/step - loss: 0.7931 - accuracy: 0.7206
Epoch 4/5
1563/1563 [==============================] - 12s 7ms/step - loss: 0.5650 - accuracy: 0.8042
Epoch 5/5
1563/1563 [==============================] - 12s 8ms/step - loss: 0.3639 - accuracy: 0.8758
<keras.callbacks.History at 0x7fb4eec543a0>

np.argmax(y_pred[0])

print("Classification Report:\n", classification_report(y_test,y_pred_classes))

Classification Report:
precision recall f1-score support

0 0.66 0.41 0.51 1000


1 0.62 0.61 0.61 1000
2 0.35 0.41 0.38 1000
3 0.36 0.29 0.32 1000
4 0.57 0.24 0.34 1000
5 0.41 0.40 0.41 1000
6 0.45 0.68 0.54 1000
7 0.66 0.44 0.53 1000
8 0.48 0.76 0.59 1000
9 0.50 0.61 0.55 1000

accuracy 0.49 10000


macro avg 0.51 0.49 0.48 10000
weighted avg 0.51 0.49 0.48 10000

lt i h (X t i [9])
plt.imshow(X_train[9])

<matplotlib.image.AxesImage at 0x7fb4c82d1340>

Colab paid products - Cancel contracts here

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy