Femi Oyedokun | Freelancer Portfolio Item #252974

6/26/2020 ConvNet_MultiClass_Classifier Multi-Class Classifier ¶ The goal of this project is to build a multi-class classifier to recognise sign language In [1]: import csv import numpy as np import tensorflow as tf from tensorflow.keras.preprocessing.image import ImageDataGenerator from os import getcwd In [2]: # Function for reading the file passed def get_data(filename): with open(filename) as training_file: reader = csv.reader(training_file, delimiter=',') imgs = [] labels = [] next(reader, None) for row in reader: label = row[0] data = row[1:] img = np.array(data).reshape((28, 28)) imgs.append(img) labels.append(label) images = np.array(imgs).astype(float) labels = np.array(labels).astype(float) return images, labels path_sign_mnist_train = f"{getcwd()}/../tmp2/sign_mnist_train.csv" path_sign_mnist_test = f"{getcwd()}/../tmp2/sign_mnist_test.csv" training_images, training_labels = get_data(path_sign_mnist_train) testing_images, testing_labels = get_data(path_sign_mnist_test) # checking dimensions print(training_images.shape) print(training_labels.shape) print(testing_images.shape) print(testing_labels.shape) (27455, 28, 28) (27455,) (7172, 28, 28) (7172,) file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 1/7 6/26/2020 ConvNet_MultiClass_Classifier In [10]: # Add dimension to the data training_images = np.expand_dims(training_images, -1) testing_images = np.expand_dims(testing_images, -1) # Createing an ImageDataGenerator and doing Image Augmentation train_datagen = ImageDataGenerator( rescale = 1./255, rotation_range=10, width_shift_range=0.1, height_shift_range=0.1, shear_range=0.1, zoom_range=0.1 ) validation_datagen = ImageDataGenerator( rescale = 1./255) # Checking the new dimensions #print(training_images.shape) #print(testing_images.shape) #print(training_labels) #print(testing_labels) file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 2/7 6/26/2020 ConvNet_MultiClass_Classifier In [4]: # ConvNet Model Definition model = tf.keras.models.Sequential([ # This is the first convolution tf.keras.layers.Conv2D(32, (3,3), activation='relu' ), tf.keras.layers.MaxPooling2D(2, 2), # The second convolution tf.keras.layers.Conv2D(32, (3,3), activation='relu'), tf.keras.layers.MaxPooling2D(2,2), # Flatten the results to feed into a DNN tf.keras.layers.Flatten(), #tf.keras.layers.Dropout(0.5), # 512 neuron hidden layer tf.keras.layers.Dense(512, activation='relu'), tf.keras.layers.Dense(26, activation='softmax') ]) # Compile Model. model.compile(loss = 'sparse_categorical_crossentropy', optimizer='adam', metr ics=['accuracy']) # Train the Model history = model.fit_generator(train_datagen.flow(training_images, training_lab els, batch_size=64), validation_data = validation_datagen.flow(testin g_images, testing_labels, batch_size = 64), epochs=30,verbose = 1) model.evaluate(testing_images, testing_labels, verbose=0) file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 3/7 6/26/2020 ConvNet_MultiClass_Classifier Epoch 1/30 429/429 [==============================] - 49s 114ms/step - loss: 1.6637 - ac curacy: 0.2941 - val_loss: 0.6117 - val_accuracy: 0.8115 Epoch 2/30 429/429 [==============================] - 42s 98ms/step - loss: 0.5357 - acc uracy: 0.7991 - val_loss: 0.2904 - val_accuracy: 0.9078 Epoch 3/30 429/429 [==============================] - 41s 96ms/step - loss: 0.2978 - acc uracy: 0.8925 - val_loss: 0.1258 - val_accuracy: 0.9576 Epoch 4/30 429/429 [==============================] - 41s 95ms/step - loss: 0.1913 - acc uracy: 0.9340 - val_loss: 0.0815 - val_accuracy: 0.9810 Epoch 5/30 429/429 [==============================] - 42s 99ms/step - loss: 0.1309 - acc uracy: 0.9546 - val_loss: 0.0465 - val_accuracy: 0.9877 Epoch 6/30 429/429 [==============================] - 41s 96ms/step - loss: 0.0952 - acc uracy: 0.9689 - val_loss: 0.0336 - val_accuracy: 0.9883 Epoch 7/30 429/429 [==============================] - 43s 100ms/step - loss: 0.0784 - ac curacy: 0.9725 - val_loss: 0.0407 - val_accuracy: 0.9866 Epoch 8/30 429/429 [==============================] - 41s 96ms/step - loss: 0.0641 - acc uracy: 0.9815 - val_loss: 0.0333 - val_accuracy: 0.9884 Epoch 9/30 429/429 [==============================] - 43s 100ms/step - loss: 0.0525 - ac curacy: 0.9806 - val_loss: 0.0207 - val_accuracy: 0.9923 Epoch 10/30 429/429 [==============================] - 41s 94ms/step - loss: 0.0464 - acc uracy: 0.9866 - val_loss: 0.0159 - val_accuracy: 0.9958 Epoch 11/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0421 - acc uracy: 0.9867 - val_loss: 0.0220 - val_accuracy: 0.9905 Epoch 12/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0372 - acc uracy: 0.9865 - val_loss: 0.0083 - val_accuracy: 0.9992 Epoch 13/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0331 - acc uracy: 0.9893 - val_loss: 0.0218 - val_accuracy: 0.9918 Epoch 14/30 429/429 [==============================] - 39s 92ms/step - loss: 0.0295 - acc uracy: 0.9904 - val_loss: 0.0223 - val_accuracy: 0.9930 Epoch 15/30 429/429 [==============================] - 40s 93ms/step - loss: 0.0292 - acc uracy: 0.9913 - val_loss: 0.0294 - val_accuracy: 0.9895 Epoch 16/30 429/429 [==============================] - 44s 101ms/step - loss: 0.0302 - ac curacy: 0.9915 - val_loss: 0.0136 - val_accuracy: 0.9953 Epoch 17/30 429/429 [==============================] - 45s 105ms/step - loss: 0.0245 - ac curacy: 0.9923 - val_loss: 0.0106 - val_accuracy: 0.9947 Epoch 18/30 429/429 [==============================] - 41s 95ms/step - loss: 0.0236 - acc uracy: 0.9931 - val_loss: 0.0128 - val_accuracy: 0.9936 Epoch 19/30 429/429 [==============================] - 41s 95ms/step - loss: 0.0192 - acc uracy: 0.9934 - val_loss: 0.0097 - val_accuracy: 0.9967 file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 4/7 6/26/2020 ConvNet_MultiClass_Classifier Epoch 20/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0255 - acc uracy: 0.9927 - val_loss: 0.0097 - val_accuracy: 0.9968 Epoch 21/30 429/429 [==============================] - 41s 96ms/step - loss: 0.0197 - acc uracy: 0.9944 - val_loss: 0.0032 - val_accuracy: 0.9997 Epoch 22/30 429/429 [==============================] - 40s 93ms/step - loss: 0.0180 - acc uracy: 0.9936 - val_loss: 0.0139 - val_accuracy: 0.9937 Epoch 23/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0167 - acc uracy: 0.9952 - val_loss: 0.0112 - val_accuracy: 0.9969 Epoch 24/30 429/429 [==============================] - 42s 98ms/step - loss: 0.0170 - acc uracy: 0.9955 - val_loss: 0.0121 - val_accuracy: 0.9968 Epoch 25/30 429/429 [==============================] - 43s 100ms/step - loss: 0.0178 - ac curacy: 0.9938 - val_loss: 0.0131 - val_accuracy: 0.9927 Epoch 26/30 429/429 [==============================] - 42s 97ms/step - loss: 0.0155 - acc uracy: 0.9965 - val_loss: 0.0079 - val_accuracy: 0.9985 Epoch 27/30 429/429 [==============================] - 40s 93ms/step - loss: 0.0158 - acc uracy: 0.9953 - val_loss: 0.0026 - val_accuracy: 0.9999 Epoch 28/30 429/429 [==============================] - 40s 93ms/step - loss: 0.0092 - acc uracy: 0.9974 - val_loss: 0.0018 - val_accuracy: 1.0000 Epoch 29/30 429/429 [==============================] - 39s 92ms/step - loss: 0.0145 - acc uracy: 0.9958 - val_loss: 0.0196 - val_accuracy: 0.9925 Epoch 30/30 429/429 [==============================] - 41s 96ms/step - loss: 0.0151 - acc uracy: 0.9952 - val_loss: 0.0158 - val_accuracy: 0.9933 Out[4]: [-,-] file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 5/7 6/26/2020 ConvNet_MultiClass_Classifier In [5]: # Plotting the chart for accuracy and loss on both training and validation %matplotlib inline import matplotlib.pyplot as plt acc = history.history['accuracy'] val_acc = history.history['val_accuracy'] loss = history.history['loss'] val_loss = history.history['val_loss'] epochs = range(len(acc)) plt.plot(epochs, acc, 'r', label='Training accuracy') plt.plot(epochs, val_acc, 'b', label='Validation accuracy') plt.title('Training and validation accuracy') plt.legend() plt.figure() plt.plot(epochs, loss, 'r', label='Training Loss') plt.plot(epochs, val_loss, 'b', label='Validation Loss') plt.title('Training and validation loss') plt.legend() plt.show() file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 6/7 6/26/2020 ConvNet_MultiClass_Classifier In [ ]: file:///C:/Users/User/Downloads/utf-8''ConvNet_MultiClass_Classifier.html 7/7