Neural Network Keras identifies a cat and Dog ACC 71 % in 3 EPOCH

Import Library¶

In [15]:

import tensorflow as tf
import numpy as np
import os
import sys
import cv2
import matplotlib.pyplot as plt
import pickle
import random
import pandas as pd

import tensorflow as tf
from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense,Dropout,Activation, Flatten,Conv2D,MaxPooling2D
import pickle

Declare PATH for TRAIN AND TEST¶

In [2]:

DATA_DIR = "/Users/soumilshah/IdeaProjects/mytensorflow/Dataset/training_set"
CATERGORIES = ['cats','dogs']
IMAGE_SIZE = 50

Load dataset¶

In [3]:

def create_training_data():
    training_date = []
    for categories in CATERGORIES:
        path = os.path.join(DATA_DIR,categories)
        class_num = CATERGORIES.index(categories)
        for img in os.listdir(path):
            try:
                img_array = cv2.imread(os.path.join(path,img),cv2.IMREAD_GRAYSCALE)
                new_array = cv2.resize(img_array,(IMAGE_SIZE,IMAGE_SIZE))
                training_date.append([new_array,class_num])
            except:
                pass
    return training_date

cnvert list to np ndarray¶

In [4]:

data = np.asarray(create_training_data())

In [5]:

x_data = []
y_data = []

for x in data:
    x_data.append(x[0])
    y_data.append(x[1])

Normalize image¶

In [7]:

x_data_np = np.asarray(x_data)/255.0
y_data_np = np.asarray(y_data)

Store the data in pickle file so every time you dont have to load images¶

In [8]:

pickle_out = open('x_data_np','wb')
pickle.dump(x_data_np,pickle_out)
pickle_out.close()

Load the data set which is ND array¶

In [9]:

pickle_out = open('y_data_np','wb')
pickle.dump(y_data_np,pickle_out)
pickle_out.close()

In [10]:

X_Temp = open('x_data_np','rb')
x_data_np = pickle.load(X_Temp)

Y_Temp = open('y_data_np','rb')
y_data_np = pickle.load(Y_Temp)

In [13]:

x_data_np = x_data_np.reshape(-1, 50, 50, 1)

Perform train test split¶

In [14]:

from sklearn.model_selection import train_test_split
X_Train, X_Test, Y_Train, Y_Test = train_test_split(x_data_np, y_data_np, test_size=0.3,random_state=101)

create Model¶

In [18]:

model = Sequential()
model.add(Conv2D(150, (3, 3), input_shape=x_data_np.shape[1:]))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(75, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(64))
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
              optimizer='adam',
              metrics=['accuracy'])

In [19]:

model.fit(x_data_np, y_data_np, batch_size=32, epochs=3, validation_split=0.3)

Train on 5600 samples, validate on 2400 samples
WARNING:tensorflow:From /anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.
Epoch 1/3
5600/5600 [==============================] - 68s 12ms/sample - loss: 0.6125 - acc: 0.7064 - val_loss: 1.1417 - val_acc: 0.0000e+00
Epoch 2/3
5600/5600 [==============================] - 69s 12ms/sample - loss: 0.5900 - acc: 0.7145 - val_loss: 1.0541 - val_acc: 0.0012
Epoch 3/3
5600/5600 [==============================] - 72s 13ms/sample - loss: 0.5637 - acc: 0.7198 - val_loss: 0.9673 - val_acc: 0.2829

Out[19]:

<tensorflow.python.keras.callbacks.History at 0xb4d9c5e10>