13/basic-number-classifier.ipynb

{"metadata":{"_change_revision":0,"_is_fork":false,"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Basic classifier using a derivation of AlexNet","metadata":{"_uuid":"536a107437da08d03e358e0ea038420483435ea9"}},{"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nfrom subprocess import check_output\nprint(check_output([\"ls\", \"../input\"]).decode(\"utf8\"))","metadata":{"_uuid":"e660aea3fe92927c4f7ba86abe943099f9f1dce3","_cell_guid":"09f38ed4-b8be-89f0-b564-46228d9fa70d","execution":{"iopub.status.busy":"2023-04-15T02:15:29.305629Z","iopub.execute_input":"2023-04-15T02:15:29.305869Z","iopub.status.idle":"2023-04-15T02:15:29.328427Z","shell.execute_reply.started":"2023-04-15T02:15:29.305820Z","shell.execute_reply":"2023-04-15T02:15:29.327716Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stdout","text":"character\ndigit-recognizer\n\n","output_type":"stream"}]},{"cell_type":"code","source":"import matplotlib.pyplot as plt\nimport matplotlib.image as mpimg\nimport seaborn as sns\n%matplotlib inline\n\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import confusion_matrix\nimport itertools\n\nfrom keras.utils.np_utils import to_categorical # convert to one-hot-encoding\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPool2D\nfrom keras.optimizers import RMSprop\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.callbacks import ReduceLROnPlateau\n\n\nsns.set(style='white', context='notebook', palette='deep')","metadata":{"_uuid":"d1503399cf42bd1007087d30036485ee7ea7be54","_cell_guid":"612d6044-57af-a0a6-28c0-43a57f01c33d","execution":{"iopub.status.busy":"2023-04-15T02:15:29.329988Z","iopub.execute_input":"2023-04-15T02:15:29.330306Z","iopub.status.idle":"2023-04-15T02:15:29.348087Z","shell.execute_reply.started":"2023-04-15T02:15:29.330242Z","shell.execute_reply":"2023-04-15T02:15:29.347435Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"markdown","source":"Using data loading and prep from https://www.kaggle.com/yassineghouzam/introduction-to-cnn-keras-0-997-top-6","metadata":{"_uuid":"911b9bfc61b78ef2f8bb324eeca095766d61c776"}},{"cell_type":"code","source":"train = pd.read_csv(\"../input/character/Train.csv\")\ntest = pd.read_csv(\"../input/character/Test.csv\")","metadata":{"_uuid":"cdb298ec3743cce8e6ee65303cdf8a809c3aded5","execution":{"iopub.status.busy":"2023-04-15T02:15:52.217339Z","iopub.execute_input":"2023-04-15T02:15:52.217650Z","iopub.status.idle":"2023-04-15T02:16:11.783448Z","shell.execute_reply.started":"2023-04-15T02:15:52.217597Z","shell.execute_reply":"2023-04-15T02:16:11.782683Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"Y_train = train[\"label\"]\n\n# Drop 'label' column\nX_train = train.drop(labels = [\"label\"],axis = 1) \n\n# free some space\ndel train \n\ng = sns.countplot(Y_train)\n\nY_train.value_counts()","metadata":{"_uuid":"7a5dae9231d7b8cda9e310c0b0258443e1696a05","execution":{"iopub.status.busy":"2023-04-15T02:16:11.784377Z","iopub.execute_input":"2023-04-15T02:16:11.784646Z","iopub.status.idle":"2023-04-15T02:16:12.962353Z","shell.execute_reply.started":"2023-04-15T02:16:11.784601Z","shell.execute_reply":"2023-04-15T02:16:12.961595Z"},"trusted":true},"execution_count":5,"outputs":[{"execution_count":5,"output_type":"execute_result","data":{"text/plain":"0     6000\n8     5999\n14    5999\n13    5999\n12    5999\n11    5999\n10    5999\n9     5999\n7     5999\n16    5999\n6     5999\n5     5999\n4     5999\n3     5999\n2     5999\n1     5999\n15    5999\n34    5999\n33    5999\n26    5999\n32    5999\n31    5999\n30    5999\n29    5999\n28    5999\n27    5999\n25    5999\n18    5999\n24    5999\n23    5999\n22    5999\n21    5999\n20    5999\n19    5999\n17    5999\nName: label, dtype: int64"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"<matplotlib.figure.Figure at 0x7de606454ef0>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAYwAAAEGCAYAAAB2EqL0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nosAQERFTFBgiImKKAkNEREz5/4Inn/Zmd+p9AAAAAElFTkSuQmCC\n"},"metadata":{}}]},{"cell_type":"code","source":"# Normalize the data\nX_train = X_train / 255.0\ntest = test / 255.0","metadata":{"_uuid":"2471d2b77947184dbb549698ff97a04a0605d08d","execution":{"iopub.status.busy":"2023-04-15T02:16:12.963662Z","iopub.execute_input":"2023-04-15T02:16:12.964006Z","iopub.status.idle":"2023-04-15T02:16:15.054184Z","shell.execute_reply.started":"2023-04-15T02:16:12.963927Z","shell.execute_reply":"2023-04-15T02:16:15.053299Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"code","source":"# Reshape image in 3 dimensions (height = 28px, width = 28px , canal = 1)\nX_train = X_train.values.reshape(-1,28,28,1)\ntest = test.values.reshape(-1,28,28,1)","metadata":{"_uuid":"ebacd1c8eef29e04ed702eba2b6cf6485559bbb7","execution":{"iopub.status.busy":"2023-04-15T02:16:20.114449Z","iopub.execute_input":"2023-04-15T02:16:20.114778Z","iopub.status.idle":"2023-04-15T02:16:20.119840Z","shell.execute_reply.started":"2023-04-15T02:16:20.114717Z","shell.execute_reply":"2023-04-15T02:16:20.119016Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"# Encode labels to one hot vectors (ex : 2 -> [0,0,1,0,0,0,0,0,0,0])\nY_train = to_categorical(Y_train, num_classes = 35)","metadata":{"_uuid":"8dd50a996ef3267a4c41864444f0a563295fd62b","execution":{"iopub.status.busy":"2023-04-15T02:16:26.409666Z","iopub.execute_input":"2023-04-15T02:16:26.409982Z","iopub.status.idle":"2023-04-15T02:16:26.445387Z","shell.execute_reply.started":"2023-04-15T02:16:26.409914Z","shell.execute_reply":"2023-04-15T02:16:26.444750Z"},"trusted":true},"execution_count":8,"outputs":[]},{"cell_type":"code","source":"# Split the train and the validation set for the fitting\nX_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size = 0.1, random_state=42)","metadata":{"_uuid":"84960a426ed1993c61ab61ad2c3dfd44e93312b1","execution":{"iopub.status.busy":"2023-04-15T02:16:29.274406Z","iopub.execute_input":"2023-04-15T02:16:29.274728Z","iopub.status.idle":"2023-04-15T02:16:32.859962Z","shell.execute_reply.started":"2023-04-15T02:16:29.274658Z","shell.execute_reply":"2023-04-15T02:16:32.859213Z"},"trusted":true},"execution_count":9,"outputs":[]},{"cell_type":"code","source":"# Some examples\ng = plt.imshow(X_train[0][:,:,0])","metadata":{"_uuid":"c0434bacc9e8bdc895cb4c52911e0a7babb24bfd","execution":{"iopub.status.busy":"2023-04-15T02:16:32.860846Z","iopub.execute_input":"2023-04-15T02:16:32.861093Z","iopub.status.idle":"2023-04-15T02:16:32.960318Z","shell.execute_reply.started":"2023-04-15T02:16:32.861049Z","shell.execute_reply":"2023-04-15T02:16:32.959583Z"},"trusted":true},"execution_count":10,"outputs":[{"output_type":"display_data","data":{"text/plain":"<matplotlib.figure.Figure at 0x7de5e2f8b320>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n"},"metadata":{}}]},{"cell_type":"markdown","source":"Defining the model architecture.","metadata":{"_uuid":"db1c0df642f4930ceaae7a49589e63755f808c2a"}},{"cell_type":"code","source":"model = Sequential()\n\nmodel.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n                 activation ='relu', input_shape = (28,28,1)))\nmodel.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n                 activation ='relu'))\nmodel.add(MaxPool2D(pool_size=(2,2)))\nmodel.add(Dropout(0.25))\n\n\nmodel.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n                 activation ='relu'))\nmodel.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n                 activation ='relu'))\nmodel.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))\nmodel.add(Dropout(0.25))\n\nmodel.add(Conv2D(filters = 128, kernel_size = (1,1),padding = 'Same', \n                 activation ='relu'))\nmodel.add(Conv2D(filters = 128, kernel_size = (1,1),padding = 'Same', \n                 activation ='relu'))\nmodel.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))\nmodel.add(Dropout(0.25))\n\nmodel.add(Flatten())\nmodel.add(Dense(256, activation = \"relu\"))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(35, activation = \"softmax\"))","metadata":{"_uuid":"a84c0f8158e866f0c1e5922f2e0d2358e527bf1a","execution":{"iopub.status.busy":"2023-04-15T02:16:38.573398Z","iopub.execute_input":"2023-04-15T02:16:38.573689Z","iopub.status.idle":"2023-04-15T02:16:38.760810Z","shell.execute_reply.started":"2023-04-15T02:16:38.573633Z","shell.execute_reply":"2023-04-15T02:16:38.760176Z"},"trusted":true},"execution_count":11,"outputs":[]},{"cell_type":"code","source":"# Define the optimizer\noptimizer = RMSprop(lr=0.001, rho=0.9, epsilon=1e-08, decay=0.0)\n\n# Compile the model\nmodel.compile(optimizer = optimizer , loss = \"categorical_crossentropy\", metrics=[\"accuracy\"])","metadata":{"_uuid":"07990eb73670246759ff1281d663b4b45894666f","execution":{"iopub.status.busy":"2023-04-15T02:16:42.064784Z","iopub.execute_input":"2023-04-15T02:16:42.065093Z","iopub.status.idle":"2023-04-15T02:16:42.100661Z","shell.execute_reply.started":"2023-04-15T02:16:42.065034Z","shell.execute_reply":"2023-04-15T02:16:42.100073Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"code","source":"# Set a learning rate annealer\nlearning_rate_reduction = ReduceLROnPlateau(monitor='val_acc', \n                                            patience=3, \n                                            verbose=1, \n                                            factor=0.5, \n                                            min_lr=0.00001)","metadata":{"_uuid":"2aae5e7469868aa3336167206eb5ce19688cabc9","execution":{"iopub.status.busy":"2023-04-15T02:16:47.941721Z","iopub.execute_input":"2023-04-15T02:16:47.942052Z","iopub.status.idle":"2023-04-15T02:16:47.946601Z","shell.execute_reply.started":"2023-04-15T02:16:47.941987Z","shell.execute_reply":"2023-04-15T02:16:47.945791Z"},"trusted":true},"execution_count":13,"outputs":[]},{"cell_type":"code","source":"epochs = 200\nbatch_size = 86","metadata":{"_uuid":"b36d5d6ae8d5d4140294c595c1db8de44f28da32","execution":{"iopub.status.busy":"2023-04-15T02:17:08.297704Z","iopub.execute_input":"2023-04-15T02:17:08.298016Z","iopub.status.idle":"2023-04-15T02:17:08.301888Z","shell.execute_reply.started":"2023-04-15T02:17:08.297937Z","shell.execute_reply":"2023-04-15T02:17:08.300886Z"},"trusted":true},"execution_count":14,"outputs":[]},{"cell_type":"code","source":"history = model.fit(X_train, Y_train, batch_size = batch_size, epochs = epochs, \n          validation_data = (X_val, Y_val), verbose = 2)","metadata":{"_uuid":"09e33e2a254e416f3820f7f8db186588a5d3aad5","execution":{"iopub.status.busy":"2023-04-15T02:17:09.205643Z","iopub.execute_input":"2023-04-15T02:17:09.206074Z","iopub.status.idle":"2023-04-15T03:28:49.134743Z","shell.execute_reply.started":"2023-04-15T02:17:09.205986Z","shell.execute_reply":"2023-04-15T03:28:49.133995Z"},"trusted":true},"execution_count":15,"outputs":[{"name":"stdout","text":"Train on 188969 samples, validate on 20997 samples\nEpoch 1/200\n - 31s - loss: 0.9488 - acc: 0.7083 - val_loss: 0.1875 - val_acc: 0.9425\nEpoch 2/200\n - 21s - loss: 0.2576 - acc: 0.9243 - val_loss: 0.1781 - val_acc: 0.9478\nEpoch 3/200\n - 21s - loss: 0.1989 - acc: 0.9433 - val_loss: 0.1131 - val_acc: 0.9686\nEpoch 4/200\n - 22s - loss: 0.1856 - acc: 0.9489 - val_loss: 0.1650 - val_acc: 0.9549\nEpoch 5/200\n - 21s - loss: 0.1807 - acc: 0.9506 - val_loss: 0.1114 - val_acc: 0.9692\nEpoch 6/200\n - 21s - loss: 0.1879 - acc: 0.9497 - val_loss: 0.1166 - val_acc: 0.9709\nEpoch 7/200\n - 22s - loss: 0.1889 - acc: 0.9501 - val_loss: 0.1180 - val_acc: 0.9713\nEpoch 8/200\n - 23s - loss: 0.1979 - acc: 0.9486 - val_loss: 0.1012 - val_acc: 0.9726\nEpoch 9/200\n - 22s - loss: 0.2087 - acc: 0.9470 - val_loss: 0.1150 - val_acc: 0.9686\nEpoch 10/200\n - 23s - loss: 0.2179 - acc: 0.9443 - val_loss: 0.2257 - val_acc: 0.9448\nEpoch 11/200\n - 22s - loss: 0.2245 - acc: 0.9443 - val_loss: 0.1070 - val_acc: 0.9707\nEpoch 12/200\n - 22s - loss: 0.2367 - acc: 0.9415 - val_loss: 0.1167 - val_acc: 0.9677\nEpoch 13/200\n - 22s - loss: 0.2457 - acc: 0.9395 - val_loss: 0.1291 - val_acc: 0.9675\nEpoch 14/200\n - 22s - loss: 0.2560 - acc: 0.9379 - val_loss: 0.1625 - val_acc: 0.9563\nEpoch 15/200\n - 21s - loss: 0.2673 - acc: 0.9348 - val_loss: 0.1216 - val_acc: 0.9676\nEpoch 16/200\n - 21s - loss: 0.2789 - acc: 0.9337 - val_loss: 0.1244 - val_acc: 0.9696\nEpoch 17/200\n - 22s - loss: 0.2908 - acc: 0.9308 - val_loss: 0.1337 - val_acc: 0.9661\nEpoch 18/200\n - 21s - loss: 0.3031 - acc: 0.9281 - val_loss: 0.1411 - val_acc: 0.9646\nEpoch 19/200\n - 21s - loss: 0.3163 - acc: 0.9261 - val_loss: 0.2017 - val_acc: 0.9508\nEpoch 20/200\n - 21s - loss: 0.3315 - acc: 0.9232 - val_loss: 0.3093 - val_acc: 0.9226\nEpoch 21/200\n - 22s - loss: 0.3482 - acc: 0.9178 - val_loss: 0.3345 - val_acc: 0.9042\nEpoch 22/200\n - 21s - loss: 0.3610 - acc: 0.9171 - val_loss: 0.2556 - val_acc: 0.9478\nEpoch 23/200\n - 21s - loss: 0.3714 - acc: 0.9145 - val_loss: 0.2624 - val_acc: 0.9429\nEpoch 24/200\n - 22s - loss: 0.3853 - acc: 0.9111 - val_loss: 0.2031 - val_acc: 0.9547\nEpoch 25/200\n - 21s - loss: 0.4085 - acc: 0.9062 - val_loss: 0.3429 - val_acc: 0.8870\nEpoch 26/200\n - 22s - loss: 0.4240 - acc: 0.9050 - val_loss: 0.2122 - val_acc: 0.9473\nEpoch 27/200\n - 22s - loss: 0.4359 - acc: 0.9021 - val_loss: 0.2585 - val_acc: 0.9430\nEpoch 28/200\n - 21s - loss: 0.4550 - acc: 0.8984 - val_loss: 0.1832 - val_acc: 0.9527\nEpoch 29/200\n - 22s - loss: 0.4603 - acc: 0.8969 - val_loss: 0.2894 - val_acc: 0.9235\nEpoch 30/200\n - 21s - loss: 0.4751 - acc: 0.8931 - val_loss: 0.2096 - val_acc: 0.9465\nEpoch 31/200\n - 21s - loss: 0.4914 - acc: 0.8902 - val_loss: 0.1914 - val_acc: 0.9527\nEpoch 32/200\n - 22s - loss: 0.4980 - acc: 0.8866 - val_loss: 0.2579 - val_acc: 0.9244\nEpoch 33/200\n - 21s - loss: 0.5194 - acc: 0.8855 - val_loss: 0.1846 - val_acc: 0.9491\nEpoch 34/200\n - 21s - loss: 0.5212 - acc: 0.8826 - val_loss: 0.3291 - val_acc: 0.9241\nEpoch 35/200\n - 21s - loss: 0.5319 - acc: 0.8819 - val_loss: 0.3135 - val_acc: 0.9078\nEpoch 36/200\n - 21s - loss: 0.5356 - acc: 0.8797 - val_loss: 0.2407 - val_acc: 0.9419\nEpoch 37/200\n - 21s - loss: 0.5445 - acc: 0.8809 - val_loss: 0.3727 - val_acc: 0.8997\nEpoch 38/200\n - 22s - loss: 0.5766 - acc: 0.8736 - val_loss: 0.2242 - val_acc: 0.9367\nEpoch 39/200\n - 21s - loss: 0.5890 - acc: 0.8725 - val_loss: 0.2381 - val_acc: 0.9354\nEpoch 40/200\n - 21s - loss: 0.5947 - acc: 0.8710 - val_loss: 0.3259 - val_acc: 0.9315\nEpoch 41/200\n - 22s - loss: 0.6158 - acc: 0.8664 - val_loss: 0.2267 - val_acc: 0.9412\nEpoch 42/200\n - 22s - loss: 0.6432 - acc: 0.8603 - val_loss: 0.5498 - val_acc: 0.8822\nEpoch 43/200\n - 21s - loss: 0.6720 - acc: 0.8553 - val_loss: 0.2612 - val_acc: 0.9335\nEpoch 44/200\n - 21s - loss: 0.6927 - acc: 0.8500 - val_loss: 0.3830 - val_acc: 0.9102\nEpoch 45/200\n - 22s - loss: 0.7036 - acc: 0.8470 - val_loss: 0.3326 - val_acc: 0.9169\nEpoch 46/200\n - 21s - loss: 0.7215 - acc: 0.8438 - val_loss: 0.2977 - val_acc: 0.9225\nEpoch 47/200\n - 21s - loss: 0.7039 - acc: 0.8451 - val_loss: 0.3383 - val_acc: 0.9037\nEpoch 48/200\n - 21s - loss: 0.7191 - acc: 0.8438 - val_loss: 0.2630 - val_acc: 0.9340\nEpoch 49/200\n - 22s - loss: 0.7340 - acc: 0.8436 - val_loss: 0.2661 - val_acc: 0.9340\nEpoch 50/200\n - 21s - loss: 0.7315 - acc: 0.8418 - val_loss: 0.3120 - val_acc: 0.9170\nEpoch 51/200\n - 21s - loss: 0.7529 - acc: 0.8397 - val_loss: 0.3593 - val_acc: 0.9077\nEpoch 52/200\n - 21s - loss: 0.7585 - acc: 0.8360 - val_loss: 0.2445 - val_acc: 0.9367\nEpoch 53/200\n - 21s - loss: 0.7913 - acc: 0.8319 - val_loss: 0.6117 - val_acc: 0.8164\nEpoch 54/200\n - 21s - loss: 0.8016 - acc: 0.8312 - val_loss: 0.4846 - val_acc: 0.8819\nEpoch 55/200\n - 21s - loss: 0.8151 - acc: 0.8237 - val_loss: 0.3021 - val_acc: 0.9323\nEpoch 56/200\n - 21s - loss: 0.8493 - acc: 0.8230 - val_loss: 0.4434 - val_acc: 0.8734\nEpoch 57/200\n - 22s - loss: 0.8613 - acc: 0.8159 - val_loss: 0.3955 - val_acc: 0.8962\nEpoch 58/200\n - 21s - loss: 0.8684 - acc: 0.8152 - val_loss: 0.4005 - val_acc: 0.9235\nEpoch 59/200\n - 21s - loss: 0.9012 - acc: 0.8107 - val_loss: 0.5606 - val_acc: 0.8600\nEpoch 60/200\n - 21s - loss: 0.9522 - acc: 0.8046 - val_loss: 0.3325 - val_acc: 0.9194\nEpoch 61/200\n - 21s - loss: 0.9338 - acc: 0.8011 - val_loss: 0.3539 - val_acc: 0.9345\nEpoch 62/200\n - 21s - loss: 0.9289 - acc: 0.8025 - val_loss: 0.4119 - val_acc: 0.8859\nEpoch 63/200\n - 21s - loss: 0.9573 - acc: 0.7958 - val_loss: 0.4690 - val_acc: 0.8621\nEpoch 64/200\n - 21s - loss: 0.9526 - acc: 0.7982 - val_loss: 0.4404 - val_acc: 0.8770\nEpoch 65/200\n - 21s - loss: 0.9899 - acc: 0.7934 - val_loss: 0.3200 - val_acc: 0.9230\nEpoch 66/200\n - 21s - loss: 0.9815 - acc: 0.7926 - val_loss: 0.3708 - val_acc: 0.9028\nEpoch 67/200\n - 21s - loss: 0.9754 - acc: 0.7917 - val_loss: 0.3878 - val_acc: 0.9026\nEpoch 68/200\n - 21s - loss: 1.0235 - acc: 0.7877 - val_loss: 0.3686 - val_acc: 0.9095\nEpoch 69/200\n - 21s - loss: 1.0054 - acc: 0.7878 - val_loss: 0.6460 - val_acc: 0.8446\nEpoch 70/200\n - 21s - loss: 1.0271 - acc: 0.7869 - val_loss: 0.7932 - val_acc: 0.7628\nEpoch 71/200\n - 21s - loss: 1.0474 - acc: 0.7809 - val_loss: 0.7459 - val_acc: 0.7763\nEpoch 72/200\n - 21s - loss: 1.0688 - acc: 0.7734 - val_loss: 0.8196 - val_acc: 0.7439\nEpoch 73/200\n - 21s - loss: 1.0836 - acc: 0.7769 - val_loss: 0.5574 - val_acc: 0.8201\nEpoch 74/200\n - 21s - loss: 1.1009 - acc: 0.7703 - val_loss: 0.6629 - val_acc: 0.8079\nEpoch 75/200\n - 21s - loss: 1.1496 - acc: 0.7650 - val_loss: 1.0430 - val_acc: 0.7497\nEpoch 76/200\n - 21s - loss: 1.1287 - acc: 0.7662 - val_loss: 0.4490 - val_acc: 0.8701\nEpoch 77/200\n - 21s - loss: 1.1036 - acc: 0.7642 - val_loss: 0.5337 - val_acc: 0.8620\nEpoch 78/200\n - 21s - loss: 1.1624 - acc: 0.7563 - val_loss: 0.4510 - val_acc: 0.8963\nEpoch 79/200\n - 21s - loss: 1.1923 - acc: 0.7535 - val_loss: 0.6737 - val_acc: 0.8450\nEpoch 80/200\n - 21s - loss: 1.2195 - acc: 0.7518 - val_loss: 0.5562 - val_acc: 0.8901\nEpoch 81/200\n - 21s - loss: 1.1707 - acc: 0.7595 - val_loss: 1.2443 - val_acc: 0.6299\nEpoch 82/200\n - 21s - loss: 1.2374 - acc: 0.7462 - val_loss: 0.7248 - val_acc: 0.7896\nEpoch 83/200\n - 21s - loss: 1.2036 - acc: 0.7508 - val_loss: 0.4435 - val_acc: 0.8808\nEpoch 84/200\n - 21s - loss: 1.2289 - acc: 0.7393 - val_loss: 0.9106 - val_acc: 0.7591\nEpoch 85/200\n - 21s - loss: 1.2802 - acc: 0.7325 - val_loss: 0.9164 - val_acc: 0.7076\nEpoch 86/200\n - 21s - loss: 1.2844 - acc: 0.7305 - val_loss: 0.4189 - val_acc: 0.9070\nEpoch 87/200\n - 21s - loss: 1.2919 - acc: 0.7339 - val_loss: 0.5024 - val_acc: 0.8904\nEpoch 88/200\n - 21s - loss: 1.2742 - acc: 0.7287 - val_loss: 0.6532 - val_acc: 0.8569\nEpoch 89/200\n - 21s - loss: 1.3092 - acc: 0.7239 - val_loss: 1.6001 - val_acc: 0.7326\nEpoch 90/200\n - 21s - loss: 1.3898 - acc: 0.7176 - val_loss: 15.5080 - val_acc: 0.0377\nEpoch 91/200\n - 22s - loss: 1.6841 - acc: 0.6879 - val_loss: 0.5186 - val_acc: 0.8973\nEpoch 92/200\n - 21s - loss: 1.3931 - acc: 0.7043 - val_loss: 0.8843 - val_acc: 0.7346\nEpoch 93/200\n - 21s - loss: 1.3892 - acc: 0.7020 - val_loss: 0.6985 - val_acc: 0.8126\nEpoch 94/200\n - 21s - loss: 1.4509 - acc: 0.6920 - val_loss: 0.6147 - val_acc: 0.8364\nEpoch 95/200\n - 21s - loss: 1.6463 - acc: 0.6831 - val_loss: 14.6648 - val_acc: 0.0900\nEpoch 96/200\n - 21s - loss: 9.6397 - acc: 0.2980 - val_loss: 0.9411 - val_acc: 0.7161\nEpoch 97/200\n - 22s - loss: 1.5122 - acc: 0.6709 - val_loss: 0.8851 - val_acc: 0.7336\nEpoch 98/200\n - 21s - loss: 1.6096 - acc: 0.6411 - val_loss: 0.9083 - val_acc: 0.7940\nEpoch 99/200\n - 21s - loss: 1.4946 - acc: 0.6662 - val_loss: 0.7074 - val_acc: 0.8161\nEpoch 100/200\n - 22s - loss: 1.6756 - acc: 0.6553 - val_loss: 0.5829 - val_acc: 0.8612\nEpoch 101/200\n - 22s - loss: 1.4694 - acc: 0.6748 - val_loss: 0.6175 - val_acc: 0.8278\nEpoch 102/200\n - 21s - loss: 1.5167 - acc: 0.6651 - val_loss: 0.6941 - val_acc: 0.8151\nEpoch 103/200\n - 22s - loss: 1.5510 - acc: 0.6593 - val_loss: 0.6742 - val_acc: 0.8440\nEpoch 104/200\n - 22s - loss: 1.4622 - acc: 0.6703 - val_loss: 1.0995 - val_acc: 0.7199\nEpoch 105/200\n - 21s - loss: 1.4532 - acc: 0.6700 - val_loss: 1.0131 - val_acc: 0.7244\nEpoch 106/200\n - 22s - loss: 1.4600 - acc: 0.6729 - val_loss: 0.7927 - val_acc: 0.8139\nEpoch 107/200\n - 22s - loss: 1.4263 - acc: 0.6756 - val_loss: 0.5641 - val_acc: 0.8536\nEpoch 108/200\n - 21s - loss: 1.4754 - acc: 0.6724 - val_loss: 0.5003 - val_acc: 0.8751\nEpoch 109/200\n - 22s - loss: 1.4558 - acc: 0.6796 - val_loss: 0.6659 - val_acc: 0.8026\nEpoch 110/200\n - 22s - loss: 1.4854 - acc: 0.6603 - val_loss: 0.9029 - val_acc: 0.7697\nEpoch 111/200\n - 21s - loss: 1.4997 - acc: 0.6551 - val_loss: 1.0281 - val_acc: 0.7025\nEpoch 112/200\n - 22s - loss: 1.4223 - acc: 0.6742 - val_loss: 1.5555 - val_acc: 0.5397\nEpoch 113/200\n - 22s - loss: 1.4895 - acc: 0.6639 - val_loss: 0.6726 - val_acc: 0.8107\nEpoch 114/200\n - 22s - loss: 1.4513 - acc: 0.6639 - val_loss: 0.8070 - val_acc: 0.7878\nEpoch 115/200\n - 21s - loss: 1.5084 - acc: 0.6549 - val_loss: 0.7224 - val_acc: 0.8207\nEpoch 116/200\n - 22s - loss: 1.4432 - acc: 0.6538 - val_loss: 0.7274 - val_acc: 0.8159\nEpoch 117/200\n - 22s - loss: 1.4823 - acc: 0.6506 - val_loss: 0.8951 - val_acc: 0.7338\nEpoch 118/200\n - 21s - loss: 1.5012 - acc: 0.6479 - val_loss: 0.6993 - val_acc: 0.8186\nEpoch 119/200\n - 22s - loss: 1.4577 - acc: 0.6495 - val_loss: 0.5581 - val_acc: 0.8720\nEpoch 120/200\n - 22s - loss: 1.4499 - acc: 0.6533 - val_loss: 1.1254 - val_acc: 0.7372\nEpoch 121/200\n - 21s - loss: 1.6263 - acc: 0.6307 - val_loss: 0.8085 - val_acc: 0.7945\nEpoch 122/200\n - 22s - loss: 1.5062 - acc: 0.6343 - val_loss: 1.1545 - val_acc: 0.7232\nEpoch 123/200\n - 22s - loss: 1.5541 - acc: 0.6467 - val_loss: 0.6859 - val_acc: 0.8592\nEpoch 124/200\n - 21s - loss: 1.5530 - acc: 0.6320 - val_loss: 0.7631 - val_acc: 0.7964\nEpoch 125/200\n - 22s - loss: 1.5841 - acc: 0.6227 - val_loss: 0.6641 - val_acc: 0.8425\nEpoch 126/200\n - 22s - loss: 1.5919 - acc: 0.6249 - val_loss: 1.7317 - val_acc: 0.5571\nEpoch 127/200\n - 21s - loss: 3.3392 - acc: 0.5393 - val_loss: 0.8073 - val_acc: 0.7815\nEpoch 128/200\n - 22s - loss: 1.6255 - acc: 0.6311 - val_loss: 0.8848 - val_acc: 0.7906\nEpoch 129/200\n - 22s - loss: 1.6122 - acc: 0.6279 - val_loss: 0.9597 - val_acc: 0.7321\nEpoch 130/200\n - 21s - loss: 1.5874 - acc: 0.6248 - val_loss: 1.2385 - val_acc: 0.6481\nEpoch 131/200\n - 22s - loss: 1.6430 - acc: 0.6194 - val_loss: 0.6679 - val_acc: 0.8485\nEpoch 132/200\n - 22s - loss: 1.6806 - acc: 0.6203 - val_loss: 0.7847 - val_acc: 0.7977\nEpoch 133/200\n - 21s - loss: 1.6114 - acc: 0.6002 - val_loss: 0.7450 - val_acc: 0.8064\nEpoch 134/200\n - 22s - loss: 1.6803 - acc: 0.5985 - val_loss: 0.8146 - val_acc: 0.8289\nEpoch 135/200\n - 22s - loss: 1.7179 - acc: 0.5849 - val_loss: 0.8988 - val_acc: 0.7770\nEpoch 136/200\n - 22s - loss: 1.6885 - acc: 0.5863 - val_loss: 1.0046 - val_acc: 0.7262\nEpoch 137/200\n - 21s - loss: 1.7269 - acc: 0.5810 - val_loss: 0.6264 - val_acc: 0.8610\nEpoch 138/200\n - 22s - loss: 1.6384 - acc: 0.5963 - val_loss: 0.7749 - val_acc: 0.8134\nEpoch 139/200\n - 22s - loss: 1.7010 - acc: 0.5955 - val_loss: 0.9506 - val_acc: 0.7400\nEpoch 140/200\n - 21s - loss: 1.5824 - acc: 0.6240 - val_loss: 1.9107 - val_acc: 0.4896\nEpoch 141/200\n - 22s - loss: 1.5735 - acc: 0.6219 - val_loss: 1.5290 - val_acc: 0.5678\nEpoch 142/200\n - 22s - loss: 1.5247 - acc: 0.6230 - val_loss: 1.8216 - val_acc: 0.4491\nEpoch 143/200\n - 21s - loss: 1.5887 - acc: 0.6118 - val_loss: 1.5710 - val_acc: 0.4921\nEpoch 144/200\n - 21s - loss: 1.6026 - acc: 0.6072 - val_loss: 1.2102 - val_acc: 0.6336\nEpoch 145/200\n - 22s - loss: 1.5493 - acc: 0.6189 - val_loss: 0.6658 - val_acc: 0.8658\nEpoch 146/200\n - 21s - loss: 1.5897 - acc: 0.6116 - val_loss: 0.8239 - val_acc: 0.7907\nEpoch 147/200\n - 22s - loss: 1.6791 - acc: 0.5994 - val_loss: 0.7354 - val_acc: 0.8224\nEpoch 148/200\n - 22s - loss: 1.6691 - acc: 0.5979 - val_loss: 1.7095 - val_acc: 0.5015\nEpoch 149/200\n - 21s - loss: 1.5875 - acc: 0.6161 - val_loss: 0.7649 - val_acc: 0.7988\nEpoch 150/200\n - 22s - loss: 12.2246 - acc: 0.1736 - val_loss: 15.4280 - val_acc: 0.0428\nEpoch 151/200\n - 22s - loss: 12.4240 - acc: 0.1561 - val_loss: 0.9456 - val_acc: 0.7650\nEpoch 152/200\n - 21s - loss: 1.7926 - acc: 0.5690 - val_loss: 1.4122 - val_acc: 0.7774\nEpoch 153/200\n - 22s - loss: 1.6377 - acc: 0.6045 - val_loss: 1.0497 - val_acc: 0.6999\nEpoch 154/200\n - 22s - loss: 1.8477 - acc: 0.5757 - val_loss: 0.8732 - val_acc: 0.8165\nEpoch 155/200\n - 21s - loss: 1.7431 - acc: 0.5748 - val_loss: 1.6747 - val_acc: 0.6284\nEpoch 156/200\n - 22s - loss: 1.7116 - acc: 0.5918 - val_loss: 0.7797 - val_acc: 0.8021\nEpoch 157/200\n - 22s - loss: 1.7020 - acc: 0.5905 - val_loss: 0.8026 - val_acc: 0.8353\nEpoch 158/200\n - 22s - loss: 1.7123 - acc: 0.5904 - val_loss: 0.8695 - val_acc: 0.7731\nEpoch 159/200\n - 21s - loss: 1.6774 - acc: 0.6001 - val_loss: 0.9429 - val_acc: 0.7352\nEpoch 160/200\n - 22s - loss: 1.6868 - acc: 0.5942 - val_loss: 1.7227 - val_acc: 0.5278\nEpoch 161/200\n - 22s - loss: 1.9843 - acc: 0.5258 - val_loss: 0.9429 - val_acc: 0.7577\nEpoch 162/200\n - 21s - loss: 1.9204 - acc: 0.5386 - val_loss: 1.2188 - val_acc: 0.6685\nEpoch 163/200\n - 22s - loss: 2.2936 - acc: 0.5166 - val_loss: 1.8360 - val_acc: 0.4282\nEpoch 164/200\n - 22s - loss: 1.9833 - acc: 0.5227 - val_loss: 0.8766 - val_acc: 0.7848\nEpoch 165/200\n - 21s - loss: 1.8559 - acc: 0.5408 - val_loss: 1.2446 - val_acc: 0.7627\nEpoch 166/200\n - 21s - loss: 2.1105 - acc: 0.5430 - val_loss: 1.1094 - val_acc: 0.6991\nEpoch 167/200\n - 22s - loss: 1.8276 - acc: 0.5497 - val_loss: 1.5416 - val_acc: 0.5065\nEpoch 168/200\n - 21s - loss: 1.8253 - acc: 0.5497 - val_loss: 0.9286 - val_acc: 0.7580\nEpoch 169/200\n - 22s - loss: 1.9110 - acc: 0.5453 - val_loss: 0.9708 - val_acc: 0.7613\nEpoch 170/200\n - 22s - loss: 1.8576 - acc: 0.5441 - val_loss: 1.2591 - val_acc: 0.5968\nEpoch 171/200\n - 21s - loss: 1.8657 - acc: 0.5385 - val_loss: 1.2236 - val_acc: 0.6534\nEpoch 172/200\n - 22s - loss: 1.9041 - acc: 0.5459 - val_loss: 1.3246 - val_acc: 0.6232\nEpoch 173/200\n - 22s - loss: 1.8640 - acc: 0.5647 - val_loss: 0.7845 - val_acc: 0.8095\nEpoch 174/200\n - 21s - loss: 1.7813 - acc: 0.5595 - val_loss: 1.3580 - val_acc: 0.6211\nEpoch 175/200\n - 22s - loss: 1.9221 - acc: 0.5365 - val_loss: 0.9497 - val_acc: 0.7549\nEpoch 176/200\n - 22s - loss: 1.8375 - acc: 0.5549 - val_loss: 1.5379 - val_acc: 0.5173\nEpoch 177/200\n - 21s - loss: 1.7910 - acc: 0.5633 - val_loss: 1.3693 - val_acc: 0.5817\nEpoch 178/200\n - 22s - loss: 1.7302 - acc: 0.5788 - val_loss: 0.9303 - val_acc: 0.7375\nEpoch 179/200\n - 22s - loss: 1.7962 - acc: 0.5824 - val_loss: 0.7120 - val_acc: 0.8336\nEpoch 180/200\n - 22s - loss: 1.8224 - acc: 0.5810 - val_loss: 3.6251 - val_acc: 0.5757\nEpoch 181/200\n - 21s - loss: 1.7373 - acc: 0.5789 - val_loss: 0.6842 - val_acc: 0.8375\nEpoch 182/200\n - 22s - loss: 1.7600 - acc: 0.5822 - val_loss: 1.0437 - val_acc: 0.7427\nEpoch 183/200\n - 22s - loss: 1.7514 - acc: 0.5810 - val_loss: 0.7325 - val_acc: 0.8363\nEpoch 184/200\n - 21s - loss: 1.8301 - acc: 0.5659 - val_loss: 1.2271 - val_acc: 0.6387\nEpoch 185/200\n - 22s - loss: 1.8779 - acc: 0.5659 - val_loss: 1.7262 - val_acc: 0.4838\nEpoch 186/200\n - 22s - loss: 1.8477 - acc: 0.5573 - val_loss: 2.0767 - val_acc: 0.4653\nEpoch 187/200\n - 21s - loss: 1.8329 - acc: 0.5613 - val_loss: 0.7735 - val_acc: 0.8194\nEpoch 188/200\n - 22s - loss: 1.9816 - acc: 0.5558 - val_loss: 0.8233 - val_acc: 0.7929\nEpoch 189/200\n - 22s - loss: 1.8883 - acc: 0.5471 - val_loss: 0.7864 - val_acc: 0.8132\nEpoch 190/200\n - 21s - loss: 1.9006 - acc: 0.5366 - val_loss: 0.8787 - val_acc: 0.7937\nEpoch 191/200\n - 22s - loss: 1.9810 - acc: 0.5441 - val_loss: 0.8946 - val_acc: 0.7951\nEpoch 192/200\n - 22s - loss: 1.9958 - acc: 0.5346 - val_loss: 0.8766 - val_acc: 0.7865\nEpoch 193/200\n - 21s - loss: 2.0412 - acc: 0.5393 - val_loss: 0.9681 - val_acc: 0.7503\nEpoch 194/200\n - 22s - loss: 2.0557 - acc: 0.5244 - val_loss: 1.5807 - val_acc: 0.6528\nEpoch 195/200\n - 22s - loss: 2.0779 - acc: 0.4943 - val_loss: 1.2226 - val_acc: 0.6242\nEpoch 196/200\n - 21s - loss: 2.1057 - acc: 0.4813 - val_loss: 1.2049 - val_acc: 0.6630\nEpoch 197/200\n - 22s - loss: 2.0045 - acc: 0.5200 - val_loss: 1.6562 - val_acc: 0.5931\nEpoch 198/200\n - 22s - loss: 2.2424 - acc: 0.4520 - val_loss: 0.8964 - val_acc: 0.7977\nEpoch 199/200\n - 21s - loss: 2.6358 - acc: 0.3618 - val_loss: 2.1390 - val_acc: 0.3728\nEpoch 200/200\n - 22s - loss: 2.5462 - acc: 0.3660 - val_loss: 1.8740 - val_acc: 0.4312\n","output_type":"stream"}]},{"cell_type":"markdown","source":"### Model Evaluation","metadata":{"_uuid":"77cbbf1eccfc5da02a110be8a1dd35e6ada99ea3"}},{"cell_type":"code","source":"# Plot the loss and accuracy curves for training and validation \nfig, ax = plt.subplots(2,1)\nax[0].plot(history.history['loss'], color='b', label=\"Training loss\")\nax[0].plot(history.history['val_loss'], color='r', label=\"validation loss\",axes =ax[0])\nlegend = ax[0].legend(loc='best', shadow=True)\n\nax[1].plot(history.history['acc'], color='b', label=\"Training accuracy\")\nax[1].plot(history.history['val_acc'], color='r',label=\"Validation accuracy\")\nlegend = ax[1].legend(loc='best', shadow=True)","metadata":{"_uuid":"5c759bd14f7369d112f94639cd864b205ad1db2d","execution":{"iopub.status.busy":"2023-04-15T04:05:21.213211Z","iopub.execute_input":"2023-04-15T04:05:21.213538Z","iopub.status.idle":"2023-04-15T04:05:21.412837Z","shell.execute_reply.started":"2023-04-15T04:05:21.213482Z","shell.execute_reply":"2023-04-15T04:05:21.412102Z"},"trusted":true},"execution_count":16,"outputs":[{"output_type":"display_data","data":{"text/plain":"<matplotlib.figure.Figure at 0x7de5e01ba6a0>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSCSSFsL/B9ucvlIsBhz7AAAAAElFTkSuQmCC\n"},"metadata":{}}]},{"cell_type":"code","source":"# predict results\nresults = model.predict(test)\n\n# select the indix with the maximum probability\nresults = np.argmax(results,axis = 1)\n\nresults = pd.Series(results,name=\"Label\")","metadata":{"_uuid":"9f041b407211382c1f1620dfa4d831c2d6d96ad0","execution":{"iopub.status.busy":"2023-04-15T04:05:21.691407Z","iopub.execute_input":"2023-04-15T04:05:21.691694Z","iopub.status.idle":"2023-04-15T04:05:23.818148Z","shell.execute_reply.started":"2023-04-15T04:05:21.691643Z","shell.execute_reply":"2023-04-15T04:05:23.817575Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"submission = pd.concat([pd.Series(range(1,28001),name = \"ImageId\"),results],axis = 1)\n\nsubmission.to_csv(\"cnn_mnist_results.csv\",index=False)","metadata":{"_uuid":"f5a4f8822dd1d496c638d7398d8d3111fcb1192f","trusted":true},"execution_count":null,"outputs":[]}]}