24/05/2023, 17:37 Assignment No.

9 FINAL - Jupyter Notebook

In [1]:

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

2023-05-17 11:47:28.854513: I tensorflow/core/util/port.cc:110] one

DNN custom operations are on. You may see slightly different numeri
cal results due to floating-point round-off errors from different c
omputation orders. To turn them off, set the environment variable `
TF_ENABLE_ONEDNN_OPTS=0`.
2023-05-17 11:47:28.855506: I tensorflow/tsl/cuda/cudart_stub.cc:2
8] Could not find cuda drivers on your machine, GPU will not be use
d.
2023-05-17 11:47:28.873775: E tensorflow/compiler/xla/stream_execut
or/cuda/cuda_dnn.cc:7704] Unable to register cuDNN factory: Attempt
ing to register factory for plugin cuDNN when one has already been
registered
2023-05-17 11:47:28.873794: E tensorflow/compiler/xla/stream_execut
or/cuda/cuda_fft.cc:609] Unable to register cuFFT factory: Attempti
ng to register factory for plugin cuFFT when one has already been r
egistered
2023-05-17 11:47:28.873798: E tensorflow/compiler/xla/stream_execut
or/cuda/cuda_blas.cc:1520] Unable to register cuBLAS factory: Attem
pting to register factory for plugin cuBLAS when one has already be
en registered
2023-05-17 11:47:28.877004: I tensorflow/tsl/cuda/cudart_stub.cc:2
8] Could not find cuda drivers on your machine, GPU will not be use
d.
2023-05-17 11:47:28.877479: I tensorflow/core/platform/cpu_feature_
guard.cc:182] This TensorFlow binary is optimized to use available
CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX_VNNI FMA, in other o
perations, rebuild TensorFlow with the appropriate compiler flags.
2023-05-17 11:47:29.267976: W tensorflow/compiler/tf2tensorrt/util
s/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT

In [2]:

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

X_train.shape

Out[2]:

(50000, 32, 32, 3)

In [3]:

X_test.shape

Out[3]:

(10000, 32, 32, 3)

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 1/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [4]:

y_train.shape

Out[4]:

(50000, 1)

In [5]:

y_train[:5]

Out[5]:

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

In [6]:

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

Out[6]:

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

In [7]:

y_test = y_test.reshape(-1,)

In [8]:

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

In [9]:

def plot_sample(X, y, index):

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

In [10]:

plot_sample(X_train, y_train, 0)

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 2/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [11]:

plot_sample(X_train, y_train, 1)

In [12]:

X_train = X_train / 255.0

X_test = X_test / 255.0

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 3/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [13]:

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

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

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

2023-05-17 11:47:30.552107: I tensorflow/compiler/xla/stream_execut

or/cuda/cuda_gpu_executor.cc:995] successful NUMA node read from Sy
sFS had negative value (-1), but there must be at least one NUMA no
de, so returning NUMA node zero. See more at https://github.com/tor
valds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-
L355 (https://github.com/torvalds/linux/blob/v6.0/Documentation/AB
I/testing/sysfs-bus-pci#L344-L355)
2023-05-17 11:47:30.552319: W tensorflow/core/common_runtime/gpu/gp
u_device.cc:1960] Cannot dlopen some GPU libraries. Please make sur
e the missing libraries mentioned above are installed properly if y
ou would like to use GPU. Follow the guide at https://www.tensorflo
w.org/install/gpu (https://www.tensorflow.org/install/gpu) for how
to download and setup the required libraries for your platform.
Skipping registering GPU devices...

Epoch 1/5
1563/1563 [==============================] - 35s 22ms/step - loss:
1.8191 - accuracy: 0.3525
Epoch 2/5
1563/1563 [==============================] - 36s 23ms/step - loss:
1.6282 - accuracy: 0.4256
Epoch 3/5
1563/1563 [==============================] - 36s 23ms/step - loss:
1.5428 - accuracy: 0.4542
Epoch 4/5
1563/1563 [==============================] - 36s 23ms/step - loss:
1.4840 - accuracy: 0.4775
Epoch 5/5
1563/1563 [==============================] - 35s 23ms/step - loss:
1.4327 - accuracy: 0.4975

Out[13]:

<keras.src.callbacks.History at 0x7fd814637040>

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 4/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [14]:

from sklearn.metrics import confusion_matrix , classification_report

import numpy as np
y_pred = ann.predict(X_test)
y_pred_classes = [np.argmax(element) for element in y_pred]

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

313/313 [==============================] - 3s 8ms/step

Classification Report:
precision recall f1-score support

0 0.64 0.43 0.52 1000

1 0.72 0.48 0.57 1000
2 0.39 0.33 0.35 1000
3 0.37 0.24 0.29 1000
4 0.48 0.30 0.37 1000
5 0.41 0.38 0.40 1000
6 0.49 0.61 0.54 1000
7 0.35 0.78 0.48 1000
8 0.57 0.69 0.62 1000
9 0.58 0.56 0.57 1000

accuracy 0.48 10000

macro avg 0.50 0.48 0.47 10000
weighted avg 0.50 0.48 0.47 10000

In [15]:

cnn = models.Sequential([
layers.Conv2D(filters=32, kernel_size=(3, 3), activation='relu', input_shape=
layers.MaxPooling2D((2, 2)),

layers.Conv2D(filters=64, kernel_size=(3, 3), activation='relu'),

layers.MaxPooling2D((2, 2)),

layers.Flatten(),
layers.Dense(64, activation='relu'),
layers.Dense(10, activation='softmax')
])

In [16]:

cnn.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 5/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [17]:

cnn.fit(X_train, y_train, epochs=10)

Epoch 1/10
1563/1563 [==============================] - 19s 12ms/step - loss:
1.4756 - accuracy: 0.4655
Epoch 2/10
1563/1563 [==============================] - 19s 12ms/step - loss:
1.1272 - accuracy: 0.6052
Epoch 3/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.9880 - accuracy: 0.6554
Epoch 4/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.8955 - accuracy: 0.6899
Epoch 5/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.8287 - accuracy: 0.7119
Epoch 6/10
1563/1563 [==============================] - 18s 12ms/step - loss:
0.7684 - accuracy: 0.7345
Epoch 7/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.7228 - accuracy: 0.7503
Epoch 8/10
1563/1563 [==============================] - 18s 11ms/step - loss:
0.6735 - accuracy: 0.7650
Epoch 9/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.6355 - accuracy: 0.7797
Epoch 10/10
1563/1563 [==============================] - 19s 12ms/step - loss:
0.5936 - accuracy: 0.7939

Out[17]:

<keras.src.callbacks.History at 0x7fd71df65be0>

In [18]:

cnn.evaluate(X_test,y_test)

313/313 [==============================] - 2s 5ms/step - loss: 0.98

19 - accuracy: 0.6866

Out[18]:

[0.9819189310073853, 0.6866000294685364]

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 6/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [19]:

y_pred = cnn.predict(X_test)
y_pred[:5]

313/313 [==============================] - 1s 4ms/step

Out[19]:

array([[1.5296830e-02, 3.7161630e-04, 2.0790633e-03, 6.7601961e-01,

6.7103736e-04, 2.6613161e-01, 5.2719600e-03, 2.7515048e-03,
3.1405263e-02, 1.5254221e-06],
[5.3905547e-02, 3.9434269e-02, 5.0401150e-07, 9.8367352e-08,
1.2282582e-09, 9.1655368e-09, 6.4237095e-11, 3.0244050e-07,
9.0665001e-01, 9.1841266e-06],
[4.5537379e-01, 4.8484314e-02, 2.0938353e-03, 1.4523463e-02,
8.7027416e-05, 1.2265679e-03, 6.1702876e-06, 5.0438374e-02,
4.2089793e-01, 6.8685529e-03],
[9.8998147e-01, 5.0581153e-03, 1.3736077e-04, 1.1132373e-03,
4.8387710e-06, 1.9321926e-05, 4.4951407e-06, 7.3398351e-06,
3.6501633e-03, 2.3631919e-05],
[7.7832448e-07, 1.1846597e-05, 1.6902879e-02, 6.1331887e-02,
8.4704083e-01, 5.6822794e-03, 6.8861082e-02, 3.2818556e-05,
1.3540218e-04, 1.2205140e-07]], dtype=float32)

In [20]:

y_classes = [np.argmax(element) for element in y_pred]

y_classes[:5]

Out[20]:

[3, 8, 0, 0, 4]

In [21]:

y_test[:5]

Out[21]:

array([3, 8, 8, 0, 6], dtype=uint8)

In [22]:

plot_sample(X_test, y_test,3)

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 7/8

24/05/2023, 17:37 Assignment No. 9 FINAL - Jupyter Notebook

In [23]:

classes[y_classes[3]]

Out[23]:

'airplane'

In [ ]:

localhost:8888/notebooks/Untitled Folder 1/Assignment No. 9 FINAL.ipynb 8/8