diff --git a/Framework/Keras/Keras Assingment/RitikJain_Task1/RitikJain_wihLogs.ipynb b/Framework/Keras/Keras Assingment/RitikJain_Task1/RitikJain_wihLogs.ipynb new file mode 100644 index 0000000..e9ef497 --- /dev/null +++ b/Framework/Keras/Keras Assingment/RitikJain_Task1/RitikJain_wihLogs.ipynb @@ -0,0 +1,266 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 32, + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np\n", + "import keras\n", + "from keras.models import Sequential\n", + "from keras.layers import Dense\n", + "from keras.optimizers import Adam\n", + "from keras.datasets import mnist\n", + "from keras.utils.np_utils import to_categorical # For hot encoding y labels\n", + "import random\n", + "import matplotlib.pyplot as plt" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": 33, + "metadata": {}, + "outputs": [], + "source": [ + "(X_train, y_train), (X_test, y_test) = mnist.load_data() # loading the dataset from mnist" + ] + }, + { + "cell_type": "code", + "execution_count": 34, + "metadata": {}, + "outputs": [], + "source": [ + "num_of_samples = [] # number of images of each label\n", + "num_classes = 10 # 10 digits from 0 to 9\n", + "cols = 5 " + ] + }, + { + "cell_type": "code", + "execution_count": 35, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "# Visualizing the data\n", + "\n", + "fig, axs = plt.subplots(nrows=num_classes, ncols=cols, figsize=(5,10)) # Creating grid\n", + "fig.tight_layout()\n", + "\n", + "for i in range(cols):\n", + " for j in range(num_classes):\n", + " x_selected = X_train[y_train == j]\n", + " axs[j][i].imshow(x_selected[random.randint(0, len(x_selected) - 1)])\n", + " axs[j][i].axis('off')\n", + " \n", + " if i == 2:\n", + " axs[j][i].set_title(str(j))\n", + " num_of_samples.append(len(x_selected)) " + ] + }, + { + "cell_type": "code", + "execution_count": 36, + "metadata": {}, + "outputs": [], + "source": [ + "# Hot encoding the data\n", + "y_train = to_categorical(y_train)\n", + "y_test = to_categorical(y_test)" + ] + }, + { + "cell_type": "code", + "execution_count": 37, + "metadata": {}, + "outputs": [], + "source": [ + "# Normalizing the dataset\n", + "X_train = X_train / 255\n", + "X_test = X_test/ 255" + ] + }, + { + "cell_type": "code", + "execution_count": 38, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "(60000, 28, 28)\n" + ] + } + ], + "source": [ + "# Formatting the dataset\n", + "num_pixels = 28 * 28\n", + "print(X_train.shape)\n", + "X_train = X_train.reshape(X_train.shape[0], num_pixels)\n", + "X_test = X_test.reshape(X_test.shape[0], num_pixels)" + ] + }, + { + "cell_type": "code", + "execution_count": 41, + "metadata": {}, + "outputs": [], + "source": [ + "def create_model():\n", + " model = Sequential()\n", + " model.add(Dense(512, input_dim=num_pixels, activation='relu'))\n", + " model.add(Dense(10, activation='softmax'))\n", + " model.compile(Adam(lr=0.001), loss='categorical_crossentropy', metrics=['accuracy'])\n", + " return model" + ] + }, + { + "cell_type": "code", + "execution_count": 43, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train on 54000 samples, validate on 6000 samples\n", + "Epoch 1/10\n", + "54000/54000 [==============================] - 4s 65us/step - loss: 0.3237 - acc: 0.9099 - val_loss: 0.1338 - val_acc: 0.9648\n", + "Epoch 2/10\n", + "54000/54000 [==============================] - 3s 59us/step - loss: 0.1349 - acc: 0.9619 - val_loss: 0.0931 - val_acc: 0.9738\n", + "Epoch 3/10\n", + "54000/54000 [==============================] - 3s 60us/step - loss: 0.0888 - acc: 0.9745 - val_loss: 0.0786 - val_acc: 0.9775\n", + "Epoch 4/10\n", + "54000/54000 [==============================] - 3s 59us/step - loss: 0.0639 - acc: 0.9817 - val_loss: 0.0725 - val_acc: 0.9800\n", + "Epoch 5/10\n", + "54000/54000 [==============================] - 3s 57us/step - loss: 0.0491 - acc: 0.9862 - val_loss: 0.0701 - val_acc: 0.9787\n", + "Epoch 6/10\n", + "54000/54000 [==============================] - 3s 60us/step - loss: 0.0375 - acc: 0.9894 - val_loss: 0.0664 - val_acc: 0.9810\n", + "Epoch 7/10\n", + "54000/54000 [==============================] - 3s 60us/step - loss: 0.0283 - acc: 0.9926 - val_loss: 0.0626 - val_acc: 0.9818\n", + "Epoch 8/10\n", + "54000/54000 [==============================] - 3s 60us/step - loss: 0.0229 - acc: 0.9941 - val_loss: 0.0640 - val_acc: 0.9817\n", + "Epoch 9/10\n", + "54000/54000 [==============================] - 3s 56us/step - loss: 0.0172 - acc: 0.9963 - val_loss: 0.0574 - val_acc: 0.9845\n", + "Epoch 10/10\n", + "54000/54000 [==============================] - 3s 64us/step - loss: 0.0129 - acc: 0.9975 - val_loss: 0.0633 - val_acc: 0.9827\n" + ] + } + ], + "source": [ + "model = create_model()\n", + "# Fitting the model\n", + "h = model.fit(x=X_train, y=y_train, verbose=1, validation_split=0.1, epochs=10, batch_size=200, shuffle=1)" + ] + }, + { + "cell_type": "code", + "execution_count": 44, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "Text(0.5, 0, 'epochs')" + ] + }, + "execution_count": 44, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "plt.plot(h.history['acc'])\n", + "plt.plot(h.history['val_acc'])\n", + "plt.legend(['acc', 'val_acc'])\n", + "plt.title('Accuracy')\n", + "plt.xlabel('epochs')" + ] + }, + { + "cell_type": "code", + "execution_count": 46, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "10000/10000 [==============================] - 0s 29us/step\n", + "The LOSS is : 0.061867225000972394\n", + "Accuracy of the model : 98.1%\n" + ] + } + ], + "source": [ + "# Evaluating the model\n", + "\n", + "val_loss, val_acc = model.evaluate(X_test, y_test)\n", + "print(\"The LOSS is : \" + str(val_loss))\n", + "print(\"Accuracy of the model : \" + str(val_acc * 100) + '%')" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.7.3" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +}