digit-recognizer-using-anns-cnns.ipynb

{"cells":[{"source":"<a href=\"https://www.kaggle.com/code/faressayah/digit-recognizer-using-anns-cnns?scriptVersionId=118474101\" target=\"_blank\"><img align=\"left\" alt=\"Kaggle\" title=\"Open in Kaggle\" src=\"https://kaggle.com/static/images/open-in-kaggle.svg\"></a>","metadata":{},"cell_type":"markdown","outputs":[],"execution_count":0},{"cell_type":"markdown","id":"01ccd22e","metadata":{"papermill":{"duration":0.009873,"end_time":"2023-02-07T13:44:00.206617","exception":false,"start_time":"2023-02-07T13:44:00.196744","status":"completed"},"tags":[]},"source":["# 1. Artificial Neural Network"]},{"cell_type":"code","execution_count":1,"id":"94d473d5","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:00.226104Z","iopub.status.busy":"2023-02-07T13:44:00.225305Z","iopub.status.idle":"2023-02-07T13:44:06.370249Z","shell.execute_reply":"2023-02-07T13:44:06.369259Z"},"papermill":{"duration":6.157304,"end_time":"2023-02-07T13:44:06.372643","exception":false,"start_time":"2023-02-07T13:44:00.215339","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["(42000, 785)\n","(28000, 784)\n"]}],"source":["import numpy as np\n","import pandas as pd\n","\n","import matplotlib.pyplot as plt\n","%matplotlib inline\n","\n","from sklearn.model_selection import train_test_split\n","\n","\n","train = pd.read_csv(\"/kaggle/input/digit-recognizer/train.csv\")\n","test = pd.read_csv(\"/kaggle/input/digit-recognizer/test.csv\")\n","\n","print(train.shape)\n","print(test.shape)\n","\n","X = train.drop('label', axis=1)\n","y = train.label\n","\n","X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)"]},{"cell_type":"markdown","id":"f1137c75","metadata":{"papermill":{"duration":0.009974,"end_time":"2023-02-07T13:44:06.391521","exception":false,"start_time":"2023-02-07T13:44:06.381547","status":"completed"},"tags":[]},"source":["## Explore the data\n","\n","Let's explore the format of the dataset before training the model. The following shows there are 60,000 images in the training set, with each image represented as 28 x 28 pixels:"]},{"cell_type":"code","execution_count":2,"id":"018d857b","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:06.409951Z","iopub.status.busy":"2023-02-07T13:44:06.40965Z","iopub.status.idle":"2023-02-07T13:44:06.421746Z","shell.execute_reply":"2023-02-07T13:44:06.420122Z"},"papermill":{"duration":0.024048,"end_time":"2023-02-07T13:44:06.424309","exception":false,"start_time":"2023-02-07T13:44:06.400261","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["Train shape:\n","(37800, 28, 28)\n","(37800,)\n","Test shape:\n","(4200, 28, 28)\n","(4200,)\n"]}],"source":["X_train = X_train.values.reshape(-1,28, 28)\n","X_test = X_test.values.reshape(-1, 28, 28)\n","\n","X_test = np.array(X_test)\n","y_test = np.array(y_test)\n","\n","print(\"Train shape:\")\n","print(X_train.shape)\n","print(y_train.shape)\n","\n","print(\"Test shape:\")\n","print(X_test.shape)\n","print(y_test.shape)"]},{"cell_type":"markdown","id":"c5ec953f","metadata":{"papermill":{"duration":0.008348,"end_time":"2023-02-07T13:44:06.44168","exception":false,"start_time":"2023-02-07T13:44:06.433332","status":"completed"},"tags":[]},"source":["## Preprocess the data\n","\n","The data must be preprocessed before training the network. If you inspect the first image in the training set, you will see that the pixel values fall in the range of 0 to 255:"]},{"cell_type":"code","execution_count":3,"id":"7f5f5931","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:06.459777Z","iopub.status.busy":"2023-02-07T13:44:06.459502Z","iopub.status.idle":"2023-02-07T13:44:06.714947Z","shell.execute_reply":"2023-02-07T13:44:06.714039Z"},"papermill":{"duration":0.267068,"end_time":"2023-02-07T13:44:06.717232","exception":false,"start_time":"2023-02-07T13:44:06.450164","status":"completed"},"tags":[]},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["plt.figure()\n","plt.imshow(X_train[0])\n","plt.colorbar()\n","plt.grid(False)\n","plt.show()"]},{"cell_type":"markdown","id":"05ba7532","metadata":{"papermill":{"duration":0.008588,"end_time":"2023-02-07T13:44:06.734999","exception":false,"start_time":"2023-02-07T13:44:06.726411","status":"completed"},"tags":[]},"source":["Scale these values to a range of 0 to 1 before feeding them to the neural network model. To do so, divide the values by 255. It's important that the training set and the testing set be preprocessed in the same way:"]},{"cell_type":"code","execution_count":4,"id":"757f46f8","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:06.753682Z","iopub.status.busy":"2023-02-07T13:44:06.753381Z","iopub.status.idle":"2023-02-07T13:44:06.855547Z","shell.execute_reply":"2023-02-07T13:44:06.854455Z"},"papermill":{"duration":0.114596,"end_time":"2023-02-07T13:44:06.858276","exception":false,"start_time":"2023-02-07T13:44:06.74368","status":"completed"},"tags":[]},"outputs":[],"source":["X_train = X_train / 255.0\n","X_test = X_test / 255.0"]},{"cell_type":"markdown","id":"431d937b","metadata":{"papermill":{"duration":0.008808,"end_time":"2023-02-07T13:44:06.877469","exception":false,"start_time":"2023-02-07T13:44:06.868661","status":"completed"},"tags":[]},"source":["To verify that the data is in the correct format and that you're ready to build and train the network, let's display the first 25 images from the training set and display the class name below each image."]},{"cell_type":"code","execution_count":5,"id":"0f44b3b1","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:06.896694Z","iopub.status.busy":"2023-02-07T13:44:06.896363Z","iopub.status.idle":"2023-02-07T13:44:06.903826Z","shell.execute_reply":"2023-02-07T13:44:06.902906Z"},"papermill":{"duration":0.019668,"end_time":"2023-02-07T13:44:06.905836","exception":false,"start_time":"2023-02-07T13:44:06.886168","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["5"]},"execution_count":5,"metadata":{},"output_type":"execute_result"}],"source":["y_train.iloc[1]"]},{"cell_type":"code","execution_count":6,"id":"9c31b1f6","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:06.92572Z","iopub.status.busy":"2023-02-07T13:44:06.924883Z","iopub.status.idle":"2023-02-07T13:44:08.867526Z","shell.execute_reply":"2023-02-07T13:44:08.866579Z"},"papermill":{"duration":1.955212,"end_time":"2023-02-07T13:44:08.869748","exception":false,"start_time":"2023-02-07T13:44:06.914536","status":"completed"},"tags":[]},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 720x1080 with 25 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["plt.figure(figsize=(10, 15))\n","for i in range(25):\n","    plt.subplot(5, 5, i + 1)\n","    plt.grid(False)\n","    plt.imshow(X_train[i], cmap=plt.cm.binary)\n","    plt.xlabel(y_train.iloc[i])\n","\n","plt.show()"]},{"cell_type":"markdown","id":"6a75f891","metadata":{"papermill":{"duration":0.009506,"end_time":"2023-02-07T13:44:08.888919","exception":false,"start_time":"2023-02-07T13:44:08.879413","status":"completed"},"tags":[]},"source":["## Build the model\n","Building the neural network requires configuring the layers of the model, then compiling the model.\n","\n","## Set up the layers\n","The basic building block of a neural network is the *layer*. Layers extract representations from the data fed into them. Hopefully, these representations are meaningful for the problem at hand.\n","\n","\n","Most of deep learning consists of chaining together simple layers. Most layers, such as `tf.keras.layers.Dense`, have parameters that are learned during training."]},{"cell_type":"code","execution_count":7,"id":"ad24b3db","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:08.909146Z","iopub.status.busy":"2023-02-07T13:44:08.908826Z","iopub.status.idle":"2023-02-07T13:44:16.435606Z","shell.execute_reply":"2023-02-07T13:44:16.434597Z"},"papermill":{"duration":7.540346,"end_time":"2023-02-07T13:44:16.438669","exception":false,"start_time":"2023-02-07T13:44:08.898323","status":"completed"},"tags":[]},"outputs":[{"name":"stderr","output_type":"stream","text":["2023-02-07 13:44:13.691526: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:13.780304: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:13.781224: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:13.783580: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\n","To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n","2023-02-07 13:44:13.783918: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:13.784970: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:13.785867: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:16.016362: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:16.017271: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:16.017927: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n","2023-02-07 13:44:16.018529: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15401 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n"]}],"source":["from tensorflow.keras.models import Sequential\n","from tensorflow.keras.layers import Dense, Dropout, Flatten\n","from tensorflow.keras.optimizers import Adam\n","\n","model = Sequential(\n","    [\n","        Flatten(input_shape=(28, 28)), \n","        Dense(128, activation='relu'), \n","        Dropout(0.2),\n","        Dense(10, activation='softmax')\n","    ]\n",")"]},{"cell_type":"markdown","id":"6d2adeca","metadata":{"papermill":{"duration":0.009476,"end_time":"2023-02-07T13:44:16.458938","exception":false,"start_time":"2023-02-07T13:44:16.449462","status":"completed"},"tags":[]},"source":["The first layer in this network, `tf.keras.layers.Flatten`, transforms the format of the images from a two-dimensional array (of 28 by 28 pixels) to a one-dimensional array (of 28 * 28 = 784 pixels). Think of this layer as unstacking rows of pixels in the image and lining them up. This layer has no parameters to learn; it only reformats the data.\n","\n","After the pixels are flattened, the network consists of a sequence of two `tf.keras.layers.Dense` layers. These are densely connected, or fully connected, neural layers. The first Dense layer has 128 nodes (or neurons). The second (and last) layer is a 10-node softmax layer that returns an array of 10 probability scores that sum to 1. Each node contains a score that indicates the probability that the current image belongs to one of the 10 classes.\n","****"]},{"cell_type":"markdown","id":"77c08158","metadata":{"papermill":{"duration":0.0093,"end_time":"2023-02-07T13:44:16.477942","exception":false,"start_time":"2023-02-07T13:44:16.468642","status":"completed"},"tags":[]},"source":["## Compile the model\n","Before the model is ready for training, it needs a few more settings. These are added during the model's compile step:\n","\n","- *Loss function* —This measures how accurate the model is during training. You want to minimize this function to \"steer\" the model in the right direction.\n","- *Optimizer* —This is how the model is updated based on the data it sees and its loss function.\n","- *Metrics* —Used to monitor the training and testing steps. The following example uses *accuracy*, the fraction of the images that are correctly classified."]},{"cell_type":"code","execution_count":8,"id":"22d79de1","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:44:16.499372Z","iopub.status.busy":"2023-02-07T13:44:16.498315Z","iopub.status.idle":"2023-02-07T13:44:16.511918Z","shell.execute_reply":"2023-02-07T13:44:16.511018Z"},"papermill":{"duration":0.026279,"end_time":"2023-02-07T13:44:16.513889","exception":false,"start_time":"2023-02-07T13:44:16.48761","status":"completed"},"tags":[]},"outputs":[],"source":["model.compile(\n","    optimizer=Adam(0.001), \n","    loss='sparse_categorical_crossentropy', \n","    metrics=['accuracy']\n",")"]},{"cell_type":"markdown","id":"f3f4535a","metadata":{"papermill":{"duration":0.009246,"end_time":"2023-02-07T13:44:16.533001","exception":false,"start_time":"2023-02-07T13:44:16.523755","status":"completed"},"tags":[]},"source":["## Train the model\n","\n","Training the neural network model requires the following steps:\n","\n","1. Feed the training data to the model. In this example, the training data is in the `X_train` and `y_train` arrays.\n","2. The model learns to associate images and labels.\n","3. You ask the model to make predictions about a test set—in this example, the `X_test` array. Verify that the predictions match the labels from the `y_test` array.\n","\n","To start training, call the `model.fit` method—so called because it \"fits\" the model to the training data:"]},{"cell_type":"code","execution_count":9,"id":"bc084317","metadata":{"_kg_hide-output":true,"execution":{"iopub.execute_input":"2023-02-07T13:44:16.55346Z","iopub.status.busy":"2023-02-07T13:44:16.553088Z","iopub.status.idle":"2023-02-07T13:45:39.342339Z","shell.execute_reply":"2023-02-07T13:45:39.341269Z"},"papermill":{"duration":82.802151,"end_time":"2023-02-07T13:45:39.344826","exception":false,"start_time":"2023-02-07T13:44:16.542675","status":"completed"},"scrolled":true,"tags":[]},"outputs":[{"name":"stderr","output_type":"stream","text":["2023-02-07 13:44:17.076846: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n"]},{"name":"stdout","output_type":"stream","text":["Epoch 1/20\n","1182/1182 [==============================] - 4s 2ms/step - loss: 0.3498 - accuracy: 0.8989 - val_loss: 0.1950 - val_accuracy: 0.9429\n","Epoch 2/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.1679 - accuracy: 0.9507 - val_loss: 0.1444 - val_accuracy: 0.9564\n","Epoch 3/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.1265 - accuracy: 0.9621 - val_loss: 0.1098 - val_accuracy: 0.9686\n","Epoch 4/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0983 - accuracy: 0.9704 - val_loss: 0.0985 - val_accuracy: 0.9719\n","Epoch 5/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0842 - accuracy: 0.9734 - val_loss: 0.0936 - val_accuracy: 0.9700\n","Epoch 6/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0715 - accuracy: 0.9771 - val_loss: 0.0874 - val_accuracy: 0.9705\n","Epoch 7/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0632 - accuracy: 0.9788 - val_loss: 0.0875 - val_accuracy: 0.9733\n","Epoch 8/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0572 - accuracy: 0.9819 - val_loss: 0.0787 - val_accuracy: 0.9755\n","Epoch 9/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0537 - accuracy: 0.9824 - val_loss: 0.0790 - val_accuracy: 0.9774\n","Epoch 10/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0444 - accuracy: 0.9858 - val_loss: 0.0823 - val_accuracy: 0.9764\n","Epoch 11/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0417 - accuracy: 0.9861 - val_loss: 0.0799 - val_accuracy: 0.9750\n","Epoch 12/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0396 - accuracy: 0.9867 - val_loss: 0.0867 - val_accuracy: 0.9745\n","Epoch 13/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0339 - accuracy: 0.9887 - val_loss: 0.0893 - val_accuracy: 0.9762\n","Epoch 14/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0347 - accuracy: 0.9889 - val_loss: 0.0908 - val_accuracy: 0.9762\n","Epoch 15/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0305 - accuracy: 0.9898 - val_loss: 0.0870 - val_accuracy: 0.9767\n","Epoch 16/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0296 - accuracy: 0.9900 - val_loss: 0.0893 - val_accuracy: 0.9760\n","Epoch 17/20\n","1182/1182 [==============================] - 3s 2ms/step - loss: 0.0281 - accuracy: 0.9910 - val_loss: 0.0915 - val_accuracy: 0.9776\n","Epoch 18/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0265 - accuracy: 0.9912 - val_loss: 0.0910 - val_accuracy: 0.9769\n","Epoch 19/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0271 - accuracy: 0.9909 - val_loss: 0.0927 - val_accuracy: 0.9776\n","Epoch 20/20\n","1182/1182 [==============================] - 2s 2ms/step - loss: 0.0238 - accuracy: 0.9915 - val_loss: 0.0919 - val_accuracy: 0.9769\n"]}],"source":["r = model.fit(\n","    X_train, y_train, \n","    validation_data=(X_test, y_test),\n","    epochs=20\n",")"]},{"cell_type":"markdown","id":"52049bc8","metadata":{"papermill":{"duration":0.055933,"end_time":"2023-02-07T13:45:39.45839","exception":false,"start_time":"2023-02-07T13:45:39.402457","status":"completed"},"tags":[]},"source":["As the model trains, the loss and accuracy metrics are displayed. This model reaches an accuracy of about 0.99 (or 99%) on the training data."]},{"cell_type":"markdown","id":"de8e1d2b","metadata":{"papermill":{"duration":0.055352,"end_time":"2023-02-07T13:45:39.569173","exception":false,"start_time":"2023-02-07T13:45:39.513821","status":"completed"},"tags":[]},"source":["## Evaluate accuracy\n","Next, compare how the model performs on the test dataset:"]},{"cell_type":"code","execution_count":10,"id":"aa4c1b27","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:39.684978Z","iopub.status.busy":"2023-02-07T13:45:39.683956Z","iopub.status.idle":"2023-02-07T13:45:42.824842Z","shell.execute_reply":"2023-02-07T13:45:42.823416Z"},"papermill":{"duration":3.202337,"end_time":"2023-02-07T13:45:42.827431","exception":false,"start_time":"2023-02-07T13:45:39.625094","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["1182/1182 - 1s - loss: 0.0061 - accuracy: 0.9985\n","Train accuracy: 0.9985449910163879\n","\n","132/132 - 0s - loss: 0.0919 - accuracy: 0.9769\n","Test accuracy: 0.9769047498703003\n","\n"]}],"source":["train_loss, train_acc = model.evaluate(X_train, y_train, verbose=2)\n","print(f'Train accuracy: {train_acc}\\n')\n","\n","test_loss, test_acc = model.evaluate(X_test, y_test, verbose=2)\n","print(f'Test accuracy: {test_acc}\\n')"]},{"cell_type":"code","execution_count":11,"id":"d176cc44","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:42.976576Z","iopub.status.busy":"2023-02-07T13:45:42.976017Z","iopub.status.idle":"2023-02-07T13:45:43.387882Z","shell.execute_reply":"2023-02-07T13:45:43.386912Z"},"papermill":{"duration":0.506639,"end_time":"2023-02-07T13:45:43.390247","exception":false,"start_time":"2023-02-07T13:45:42.883608","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["<matplotlib.legend.Legend at 0x7fb49c55fd90>"]},"execution_count":11,"metadata":{},"output_type":"execute_result"},{"data":{"image/png":"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\n","text/plain":["<Figure size 864x576 with 2 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["plt.figure(figsize=(12, 8))\n","\n","plt.subplot(2, 2, 1)\n","plt.plot(r.history['loss'], label='loss')\n","plt.plot(r.history['val_loss'], label='val_loss')\n","plt.xlabel(\"Epochs\")\n","plt.ylabel(\"Loss function\")\n","plt.legend()\n","\n","plt.subplot(2, 2, 2)\n","plt.plot(r.history['accuracy'], label='accuracy')\n","plt.plot(r.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel(\"Epochs\")\n","plt.ylabel(\"Accuracy\")\n","plt.legend()"]},{"cell_type":"markdown","id":"a5582edb","metadata":{"papermill":{"duration":0.083734,"end_time":"2023-02-07T13:45:43.556856","exception":false,"start_time":"2023-02-07T13:45:43.473122","status":"completed"},"tags":[]},"source":["It turns out that the accuracy on the test dataset is a little less than the accuracy on the training dataset. This gap between training accuracy and test accuracy represents overfitting. Overfitting is when a machine learning model performs worse on new, previously unseen inputs than on the training data."]},{"cell_type":"markdown","id":"a2e3ee40","metadata":{"papermill":{"duration":0.05544,"end_time":"2023-02-07T13:45:43.667745","exception":false,"start_time":"2023-02-07T13:45:43.612305","status":"completed"},"tags":[]},"source":["## Makes predictions\n","\n","With the model trained, you can use it to make predictions about some images."]},{"cell_type":"code","execution_count":12,"id":"3a053898","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:43.780077Z","iopub.status.busy":"2023-02-07T13:45:43.779713Z","iopub.status.idle":"2023-02-07T13:45:44.040543Z","shell.execute_reply":"2023-02-07T13:45:44.039625Z"},"papermill":{"duration":0.320038,"end_time":"2023-02-07T13:45:44.042915","exception":false,"start_time":"2023-02-07T13:45:43.722877","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["array([2.1531051e-07, 2.1905835e-05, 2.2849565e-05, 3.7150204e-07,\n","       1.1132324e-07, 1.2086915e-09, 1.6568263e-06, 3.0215983e-07,\n","       9.9995255e-01, 1.3705894e-08], dtype=float32)"]},"execution_count":12,"metadata":{},"output_type":"execute_result"}],"source":["predictions = model.predict(X_test)\n","predictions[0]"]},{"cell_type":"markdown","id":"3a8e1f9e","metadata":{"papermill":{"duration":0.056371,"end_time":"2023-02-07T13:45:44.155861","exception":false,"start_time":"2023-02-07T13:45:44.09949","status":"completed"},"tags":[]},"source":["Here, the model has predicted the label for each image in the testing set. Let's take a look at the first prediction:\n","\n","A prediction is an array of 10 numbers. They represent the model's \"confidence\" that the image corresponds to each of the 10 different articles of clothing. You can see which label has the highest confidence value:"]},{"cell_type":"code","execution_count":13,"id":"9838dbd5","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:44.270261Z","iopub.status.busy":"2023-02-07T13:45:44.269587Z","iopub.status.idle":"2023-02-07T13:45:44.276445Z","shell.execute_reply":"2023-02-07T13:45:44.275459Z"},"papermill":{"duration":0.066913,"end_time":"2023-02-07T13:45:44.278536","exception":false,"start_time":"2023-02-07T13:45:44.211623","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["8"]},"execution_count":13,"metadata":{},"output_type":"execute_result"}],"source":["np.argmax(predictions[0])"]},{"cell_type":"markdown","id":"0e5d6366","metadata":{"papermill":{"duration":0.056718,"end_time":"2023-02-07T13:45:44.390995","exception":false,"start_time":"2023-02-07T13:45:44.334277","status":"completed"},"tags":[]},"source":["So, the model is most confident that the number on this image is a 7. Examining the test label shows that this classification is correct:"]},{"cell_type":"code","execution_count":14,"id":"8e3d361f","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:44.559703Z","iopub.status.busy":"2023-02-07T13:45:44.559352Z","iopub.status.idle":"2023-02-07T13:45:44.565914Z","shell.execute_reply":"2023-02-07T13:45:44.565016Z"},"papermill":{"duration":0.068825,"end_time":"2023-02-07T13:45:44.567848","exception":false,"start_time":"2023-02-07T13:45:44.499023","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["8"]},"execution_count":14,"metadata":{},"output_type":"execute_result"}],"source":["y_test[0]"]},{"cell_type":"markdown","id":"88d689cf","metadata":{"papermill":{"duration":0.056117,"end_time":"2023-02-07T13:45:44.679577","exception":false,"start_time":"2023-02-07T13:45:44.62346","status":"completed"},"tags":[]},"source":["Graph this to look at the full set of 10 class predictions."]},{"cell_type":"code","execution_count":15,"id":"976b737a","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:44.792879Z","iopub.status.busy":"2023-02-07T13:45:44.792535Z","iopub.status.idle":"2023-02-07T13:45:44.800796Z","shell.execute_reply":"2023-02-07T13:45:44.799688Z"},"papermill":{"duration":0.067927,"end_time":"2023-02-07T13:45:44.803207","exception":false,"start_time":"2023-02-07T13:45:44.73528","status":"completed"},"tags":[]},"outputs":[],"source":["def plot_image(i, predictions_array, true_label, img):\n","    predictions_array, true_label, img = predictions_array, true_label[i], img[i]\n","    plt.grid(False)\n","    plt.xticks([])\n","    plt.yticks([])\n","\n","    plt.imshow(img, cmap=plt.cm.binary)\n","\n","    predicted_label = np.argmax(predictions_array)\n","    if predicted_label == true_label:\n","        color = 'blue'\n","    else:\n","        color = 'red'\n","\n","    plt.xlabel(f\"{predicted_label} {100*np.max(predictions_array):2.0f}% ({true_label})\", \n","               color=color)\n","\n","def plot_value_array(i, predictions_array, true_label):\n","    predictions_array, true_label = predictions_array, true_label[i]\n","    plt.grid(False)\n","    plt.xticks(range(10))\n","    plt.yticks([])\n","    thisplot = plt.bar(range(10), predictions_array, color=\"#777777\")\n","    plt.ylim([0, 1])\n","    predicted_label = np.argmax(predictions_array)\n","\n","    thisplot[predicted_label].set_color('red')\n","    thisplot[true_label].set_color('blue')"]},{"cell_type":"markdown","id":"a7ab6402","metadata":{"papermill":{"duration":0.056569,"end_time":"2023-02-07T13:45:44.916097","exception":false,"start_time":"2023-02-07T13:45:44.859528","status":"completed"},"tags":[]},"source":["Let's look at the 0th image, predictions, and prediction array. Correct prediction labels are blue and incorrect prediction labels are red. The number gives the percentage (out of 100) for the predicted label."]},{"cell_type":"code","execution_count":16,"id":"d01b58d6","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:45.030315Z","iopub.status.busy":"2023-02-07T13:45:45.029268Z","iopub.status.idle":"2023-02-07T13:45:45.174166Z","shell.execute_reply":"2023-02-07T13:45:45.173074Z"},"papermill":{"duration":0.205419,"end_time":"2023-02-07T13:45:45.177578","exception":false,"start_time":"2023-02-07T13:45:44.972159","status":"completed"},"tags":[]},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 432x216 with 2 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["i = 0\n","plt.figure(figsize=(6,3))\n","plt.subplot(1,2,1)\n","plot_image(i, predictions[i], y_test, X_test)\n","plt.subplot(1,2,2)\n","plot_value_array(i, predictions[i],  y_test)\n","plt.show()"]},{"cell_type":"code","execution_count":17,"id":"48f432ad","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:45.341266Z","iopub.status.busy":"2023-02-07T13:45:45.340906Z","iopub.status.idle":"2023-02-07T13:45:45.488014Z","shell.execute_reply":"2023-02-07T13:45:45.486234Z"},"papermill":{"duration":0.208938,"end_time":"2023-02-07T13:45:45.491238","exception":false,"start_time":"2023-02-07T13:45:45.2823","status":"completed"},"tags":[]},"outputs":[{"data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAWAAAADCCAYAAAB3whgdAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAALt0lEQVR4nO3dcazd5V3H8feXFihMGljL1hZwd1PANSTFtjLqkgZhTOYqytRkdW1QFzVkKrolCkZT50zApDphcXEbIBoHc95uaHAZI2E4TRza264ZlZXhbLdaJpTuusno1su+/vH7tes45/b+zu25/d72vl/Jyb3nuc/3PM+5bT/3Oc/zO72RmUiSTrzTqicgSXOVASxJRQxgSSpiAEtSEQNYkooYwJJUZH71BKRqixcvzpGRkaE+5o4dMDHRvf/8+bBixVCnoFlibGxsf2ae3+9rBrDmvJGREbZu3TrUx4wYrP/EBAx5CpolImLPZF9zC0KSihjAklTEAJakIgPtAc/EYYV02O7du9m/f/+Au6fSyWugAJ6JwwrpsNWrV1dPQTqh3IKQpCIGsCQVMYAlqYgBLElFDGBJKmIAS1IRA1iSihjAklTEAJakIgawJBUxgCWpiAEsSUUMYEkqYgBLUhEDWJKKGMCSVMQAlqQiBrAkFTGAJanIQL8TTifekiVLetoWL17ct++6det62m6//fahz0nScLgClqQiBrAkFTGAJamIASxJRTyEm+U2bdrU0/biiy/27btx48aZno6kIXIFLElFDGBJKmIAS1IRA1iSihjAklTEqyCG5NChQz1to6OjffuuX7++8+PedNNN056TpNnNFbAkFTGAJamIASxJRQxgSSriIdyQ3HnnnT1tmzdv7tt3kEM4SacuV8CSVMQAlqQiBrAkFTGAJamIASxJRU7pqyD27dvX07Zs2bIZGWv79u09bQcPHuzb97nnnutpW7Ro0dDnJGl2cwUsSUUMYEkqYgBLUhEDWJKKnNKHcOPj4z1twziE27VrV0/bAw880NPW7/8IBnjyySd72tasWXPc85J0cnEFLElFDGBJKmIAS1IRA1iSihjAklTklL4KYvny5TPyuHv27Olpe/7553valixZ0rfeKx4kgStgSSpjAEtSEQNYkooYwJJUxACWpCIGsCQVMYAlqYgBLElFDGBJKmIAS1IRA1iSihjAklTEAJakIgawJBUxgCWpiAEsSUUMYEkqYgBLUhEDWJKKGMCSVMQAlqQiBrAkFTGAJamIASxJRQxgSSpiAEtSkfnVEziVjY+P920fGxvraVu1atUMz0bSbOMKWJKKGMCSVMQAlqQiBrAkFfEQbhqWLl3a07ZgwYKetoMHD/atP3DgwNDnJOnk4wpYkooYwJJUxACWpCIGsCQVMYAlqYhXQUzDJZdc0tO2cOHCnrbJroKQJHAFLEllDGBJKmIAS1IRA1iSingINw1nnnlmT9tpp3X/WfbCCy8MczqSTlKugCWpiAEsSUUMYEkqYgBLUhEDWJKKeBVEgS1btvS0XX/99QUzkVTJFbAkFTGAJamIASxJRQxgSSriIdyQbNiwoadt8+bNnftKmntcAUtSEQNYkooYwJJUxACWpCIGsCQV8SqIAqOjoz1t1157bcFMJFVyBSxJRQxgSSpiAEtSEQNYkop4CDck69at62mb7K3I27Ztm+npSDoJuAKWpCIGsCQVMYAlqYgBLElFDGBJKuJVEEOycuXKnrbbbrutb99bb711pqcj6STgCliSihjAklTEAJakIgawJBXxEG5IzjnnnJ62W265pW/fydolzS2ugCWpiAEsSUUMYEkqYgBLUhEDWJKKGMCSVMQAlqQiBrAkFTGAJamIASxJRQxgSSpiAEtSEQNYkooYwJJUxACWpCIGsCQVMYAlqYgBLElFDGBJKmIAS1IRA1iSigz0W5HHxsb2R8SemZqM5rxXVU9AOpEGCuDMPH+mJiJJc41bEJJUxACWpCIGsCQVMYAlqUhkZvUcpFIR8Sww6NU9i4H90xjOutkz5omqe9VkFzAYwNI0RMTWzFxt3XDqKsaseI4v5RaEJBWZcwEcwT0RPBPB48foszaCbRFMRPCzL/najRF8sb3deFT7qyN4rG3/2wjOaNt/JoKdEfxzBIvath+I4CPHGD8ieCSChceacwSbI7h6et8JSdXmXAAD9wLXTdHny8AvAPcd3RjBy4FNwOuAK4BNEZzXfvmPgfdmcjHwNeDtbfu7gCuBvwZ+vm37I+D3jzH+TwA7Mvn6FHN+H3DLFM9FM+OD1g21rmLMiuf4PebkHnAEI8CDmVw2Rb97236j7f31wFWZ/Gp7/wPAo8BHgGeBJZlMRLAG+INMfjyCfwF+Gngb8C1gJ3BDJu88xrj3AR/M5NGp5hzBGPDmTL7a9flLmh3m4gr4eFwAfOWo+3vbtkXAeCYTL2kHeDfwEPAG4H7g94D3TDHO64GxjnPa1vaXdJIxgAcTfdryGO1k8nAmqzL5SZqV8CeASyMYjeBDEZzdp/blmXyj45yeAZZ17KvjFBHXRcSuiHgqIjpv/0TEPRHxTERMevYwSd1FEfHpiHgiInZGxM0d6xZExL9FxI627t0DjjsvIrZHxIMD1OyOiM9HxOciYusAdedGxGhEfKF9nms61FzajnP49vWI+M2O4/1W+z15PCLuj4gFHetubmt2dh1rKgbwYPYCFx11/0JgH801gedGHPnPjQ63H9EG7Y3A+4HbgF+iWeW+rc84ExGd/2wWAC90fQKavoiYB/w58CZgObA+IpZ3LL+Xqc8e+pkA3pWZr6U5S3hHxzG/BVydmSuAy4HrIuLKAca9GXhi0MkCP5aZlw94mdYdwCcz84eAFV3Gzcxd7TiXA6uAbwIfn6ouIi4AfgNYnZmXAfOAt3aouwz4ZZqznxXAuoi4eKq6qRjAg3kIeGME57WHb28EHsokgU/DkSsmbgT+/iW1vw3ckckh4CyaFfJ3oO8KeBfwmo5zugQmv6JDQ3UF8FRmfikzv02z9/9TXQoz8zPAgUEHzMynM3Nb+/k3aMLpgmNXQTb+r717envrdOATERcCbwbuGnS+g4qIhcBa4G6AzPx2Zo4P+DDXAP+ZmV3fTDMfOCsi5tP8+9s3RX+A1wKfzcxvZuYE8E/ADQPOs8ecC+AI7gf+lWYbYG/EkasVju7zIxHsBX4O+EAEOwEyOUCzf/vv7e0P2zaA3wHeGcFTNHvCdx/1eMuA1ZlHQvlPgM/SBPX3XGnR+kfgqqnmHMHpwA8CnV/u6bhMdgZwQkTECPDDwGMd+8+LiM/RbFM9nJmd6oA/o1kwfGfAKSbwqYgYi4hf6VjzGpoD7L9stzzuioiXDTjuW2nOV6aeYOZ/A5tprnR6GvjfzPxUh9LHgbURsSgizqa5UumiKWo6TcjbLLtBLoV8uEO/GyDfUz3fuXKj+YF811H3NwLvG6B+BHh8mmN/H82W1VumUXsuzSu0yzr0XQe8v/38KuDBAcZZ1n58BbADWNuhZjXNNsvr2vt3AJ3/TgNn0GwBvrJj//OAR4DzaV4VPABs6Fj7dppD788AfwG893j/Ts25FfDJIJOngQ8dfiPGMcynWU3rxJjsDGBGRcTpwBbgw5n5sUHrs3lJ/yjd9qBfD1wfEbtptliujoi/6TjOvvbjMzT7sVd0KNsL7M3vrs5HgZVdxmu9CdiWmf/Tsf8bgP/KzGcz8xDwMeBHuxRm5t2ZuTIz19JsJ31xgHn2ZQDPUpl8NL/7RozJ+vxdJuMnaEpqtp0ujohXR8QZNC99/2EmB4yIoNnOeiIz/3SAuvMj4tz287NogucLU9Vl5q2ZeWFmjtA8v0cyc0OH8V4WEecc/pzmfGTKs4nM/CrwlYi4tG26BviPqeqOsp6O2w+tLwNXRsTZ7ff2GjoeNkbEK9qP3w+8ZcBx+xroVxJJc1lmTkTEr9Ecxs4D7snMnV1qI+J+mpf0iyNiL7ApM+8+dhXQrEg3Ap9v93MBfjczPzFF3VLgr9orN04DPpqZnS8pm4ZXAh9vMo35wH2Z+cmOtb8OfLj9ofYl4Be7FLV7sddC88aoLjLzsYgYpdlKmAC20/2dbVsiYhFwCHhHZn6t67iTmZPvhJOk2cAtCEkqYgBLUhEDWJKKGMCSVMQAlqQiBrAkFTGAJamIASxJRf4fQ25WO5WAWUgAAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x216 with 2 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["i = 12\n","plt.figure(figsize=(6,3))\n","plt.subplot(1,2,1)\n","plot_image(i, predictions[i], y_test, X_test)\n","plt.subplot(1,2,2)\n","plot_value_array(i, predictions[i],  y_test)\n","plt.show()"]},{"cell_type":"markdown","id":"73798de7","metadata":{"papermill":{"duration":0.056147,"end_time":"2023-02-07T13:45:45.651109","exception":false,"start_time":"2023-02-07T13:45:45.594962","status":"completed"},"tags":[]},"source":["Let's plot several images with their predictions. Note that the model can be wrong even when very confident."]},{"cell_type":"code","execution_count":18,"id":"eaa9b550","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:45.764723Z","iopub.status.busy":"2023-02-07T13:45:45.764389Z","iopub.status.idle":"2023-02-07T13:45:47.752878Z","shell.execute_reply":"2023-02-07T13:45:47.751996Z"},"papermill":{"duration":2.047991,"end_time":"2023-02-07T13:45:47.754957","exception":false,"start_time":"2023-02-07T13:45:45.706966","status":"completed"},"tags":[]},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 864x720 with 30 Axes>"]},"metadata":{"needs_background":"light"},"output_type":"display_data"}],"source":["# Plot the first X test images, their predicted labels, and the true labels.\n","# Color correct predictions in blue and incorrect predictions in red.\n","num_rows = 5\n","num_cols = 3\n","num_images = num_rows * num_cols\n","plt.figure(figsize=(2 * 2 * num_cols, 2 * num_rows))\n","for i in range(num_images):\n","    plt.subplot(num_rows, 2 * num_cols, 2 * i + 1)\n","    plot_image(i, predictions[i], y_test, X_test)\n","    plt.subplot(num_rows, 2*num_cols, 2*i+2)\n","    plot_value_array(i, predictions[i], y_test)\n","plt.tight_layout()\n","plt.show()"]},{"cell_type":"markdown","id":"2a23e6e4","metadata":{"papermill":{"duration":0.056757,"end_time":"2023-02-07T13:45:47.869987","exception":false,"start_time":"2023-02-07T13:45:47.81323","status":"completed"},"tags":[]},"source":["# 2. Convolutional Neural Networks (CNNs)"]},{"cell_type":"markdown","id":"a95be80b","metadata":{"papermill":{"duration":0.057641,"end_time":"2023-02-07T13:45:47.984478","exception":false,"start_time":"2023-02-07T13:45:47.926837","status":"completed"},"tags":[]},"source":["## Reshaping the data"]},{"cell_type":"code","execution_count":19,"id":"b3494340","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:48.099049Z","iopub.status.busy":"2023-02-07T13:45:48.098691Z","iopub.status.idle":"2023-02-07T13:45:48.104826Z","shell.execute_reply":"2023-02-07T13:45:48.10391Z"},"papermill":{"duration":0.065754,"end_time":"2023-02-07T13:45:48.10687","exception":false,"start_time":"2023-02-07T13:45:48.041116","status":"completed"},"tags":[]},"outputs":[],"source":["X_train = np.expand_dims(X_train, -1)\n","X_test = np.expand_dims(X_test, -1)"]},{"cell_type":"code","execution_count":20,"id":"8f51d334","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:48.22197Z","iopub.status.busy":"2023-02-07T13:45:48.221369Z","iopub.status.idle":"2023-02-07T13:45:48.225869Z","shell.execute_reply":"2023-02-07T13:45:48.224935Z"},"papermill":{"duration":0.063924,"end_time":"2023-02-07T13:45:48.227873","exception":false,"start_time":"2023-02-07T13:45:48.163949","status":"completed"},"tags":[]},"outputs":[],"source":["from keras.layers import Conv2D, MaxPool2D\n","from keras.callbacks import EarlyStopping"]},{"cell_type":"code","execution_count":21,"id":"8f714e65","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:48.344686Z","iopub.status.busy":"2023-02-07T13:45:48.343811Z","iopub.status.idle":"2023-02-07T13:45:48.435479Z","shell.execute_reply":"2023-02-07T13:45:48.434279Z"},"papermill":{"duration":0.153097,"end_time":"2023-02-07T13:45:48.43849","exception":false,"start_time":"2023-02-07T13:45:48.285393","status":"completed"},"tags":[]},"outputs":[],"source":["model = Sequential()\n","\n","model.add(Conv2D(filters=32, kernel_size=(5, 5), activation='relu', padding='same', \n","                 input_shape=X_train[0].shape))\n","model.add(Conv2D(filters=32, kernel_size=(5, 5), activation='relu', padding='same'))\n","model.add(MaxPool2D(pool_size=(2, 2)))\n","model.add(Dropout(0.25))\n","\n","model.add(Conv2D(filters=32, kernel_size=(3, 3), activation='relu', padding='same', \n","                 input_shape=X_train[0].shape))\n","model.add(Conv2D(filters=32, kernel_size=(3, 3), activation='relu', padding='same'))\n","model.add(MaxPool2D(pool_size=(2, 2), strides=(2, 2)))\n","model.add(Dropout(0.25))\n","\n","model.add(Flatten())\n","model.add(Dense(256, activation='relu'))\n","model.add(Dropout(0.5))\n","\n","model.add(Dense(10, activation='softmax'))\n","\n","model.compile(\n","    loss='sparse_categorical_crossentropy', \n","    optimizer=Adam(0.0001), \n","    metrics=['accuracy']\n",")"]},{"cell_type":"code","execution_count":22,"id":"1abf16ce","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:45:48.633083Z","iopub.status.busy":"2023-02-07T13:45:48.632668Z","iopub.status.idle":"2023-02-07T13:45:48.637487Z","shell.execute_reply":"2023-02-07T13:45:48.636569Z"},"papermill":{"duration":0.098826,"end_time":"2023-02-07T13:45:48.643004","exception":false,"start_time":"2023-02-07T13:45:48.544178","status":"completed"},"tags":[]},"outputs":[],"source":["early_stop = EarlyStopping(monitor='val_loss', patience=2)"]},{"cell_type":"code","execution_count":23,"id":"02da983a","metadata":{"_kg_hide-output":true,"execution":{"iopub.execute_input":"2023-02-07T13:45:48.816297Z","iopub.status.busy":"2023-02-07T13:45:48.815852Z","iopub.status.idle":"2023-02-07T13:46:40.297309Z","shell.execute_reply":"2023-02-07T13:46:40.296348Z"},"papermill":{"duration":51.57121,"end_time":"2023-02-07T13:46:40.299453","exception":false,"start_time":"2023-02-07T13:45:48.728243","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["Epoch 1/10\n"]},{"name":"stderr","output_type":"stream","text":["2023-02-07 13:45:50.507388: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n"]},{"name":"stdout","output_type":"stream","text":["1182/1182 [==============================] - 11s 4ms/step - loss: 0.5853 - accuracy: 0.8056 - val_loss: 0.1395 - val_accuracy: 0.9595\n","Epoch 2/10\n","1182/1182 [==============================] - 4s 4ms/step - loss: 0.1735 - accuracy: 0.9459 - val_loss: 0.0901 - val_accuracy: 0.9733\n","Epoch 3/10\n","1182/1182 [==============================] - 5s 4ms/step - loss: 0.1272 - accuracy: 0.9617 - val_loss: 0.0705 - val_accuracy: 0.9790\n","Epoch 4/10\n","1182/1182 [==============================] - 4s 4ms/step - loss: 0.1038 - accuracy: 0.9680 - val_loss: 0.0615 - val_accuracy: 0.9812\n","Epoch 5/10\n","1182/1182 [==============================] - 5s 4ms/step - loss: 0.0877 - accuracy: 0.9738 - val_loss: 0.0507 - val_accuracy: 0.9845\n","Epoch 6/10\n","1182/1182 [==============================] - 5s 4ms/step - loss: 0.0761 - accuracy: 0.9770 - val_loss: 0.0508 - val_accuracy: 0.9843\n","Epoch 7/10\n","1182/1182 [==============================] - 4s 4ms/step - loss: 0.0710 - accuracy: 0.9781 - val_loss: 0.0420 - val_accuracy: 0.9881\n","Epoch 8/10\n","1182/1182 [==============================] - 5s 4ms/step - loss: 0.0618 - accuracy: 0.9810 - val_loss: 0.0400 - val_accuracy: 0.9874\n","Epoch 9/10\n","1182/1182 [==============================] - 4s 4ms/step - loss: 0.0592 - accuracy: 0.9818 - val_loss: 0.0371 - val_accuracy: 0.9902\n","Epoch 10/10\n","1182/1182 [==============================] - 5s 4ms/step - loss: 0.0529 - accuracy: 0.9838 - val_loss: 0.0382 - val_accuracy: 0.9879\n"]}],"source":["r = model.fit(\n","    X_train, y_train, epochs=10, \n","    validation_data=(X_test, y_test), \n","    callbacks=[early_stop]\n",")"]},{"cell_type":"code","execution_count":24,"id":"b4de9207","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:40.507507Z","iopub.status.busy":"2023-02-07T13:46:40.506856Z","iopub.status.idle":"2023-02-07T13:46:43.696109Z","shell.execute_reply":"2023-02-07T13:46:43.694124Z"},"papermill":{"duration":3.295538,"end_time":"2023-02-07T13:46:43.699096","exception":false,"start_time":"2023-02-07T13:46:40.403558","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["1182/1182 - 2s - loss: 0.0236 - accuracy: 0.9923\n","\n","Train accuracy: 0.9923015832901001\n","132/132 - 0s - loss: 0.0382 - accuracy: 0.9879\n","\n","Test accuracy: 0.9878571629524231\n"]}],"source":["train_loss, train_acc = model.evaluate(X_train, y_train, verbose=2)\n","print(f'\\nTrain accuracy: {train_acc}')\n","\n","test_loss, test_acc = model.evaluate(X_test, y_test, verbose=2)\n","print(f'\\nTest accuracy: {test_acc}')"]},{"cell_type":"code","execution_count":25,"id":"d208f6a1","metadata":{"_kg_hide-output":true,"execution":{"iopub.execute_input":"2023-02-07T13:46:43.900649Z","iopub.status.busy":"2023-02-07T13:46:43.899725Z","iopub.status.idle":"2023-02-07T13:46:48.568721Z","shell.execute_reply":"2023-02-07T13:46:48.567578Z"},"papermill":{"duration":4.772448,"end_time":"2023-02-07T13:46:48.571318","exception":false,"start_time":"2023-02-07T13:46:43.79887","status":"completed"},"tags":[]},"outputs":[{"name":"stdout","output_type":"stream","text":["Epoch 1/10\n","132/132 [==============================] - 0s 4ms/step - loss: 0.0666 - accuracy: 0.9802\n","Epoch 2/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0565 - accuracy: 0.9821\n","Epoch 3/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0541 - accuracy: 0.9838\n","Epoch 4/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0478 - accuracy: 0.9838\n","Epoch 5/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0412 - accuracy: 0.9862\n","Epoch 6/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0329 - accuracy: 0.9905\n","Epoch 7/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0379 - accuracy: 0.9869\n","Epoch 8/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0294 - accuracy: 0.9898\n","Epoch 9/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0335 - accuracy: 0.9890\n","Epoch 10/10\n","132/132 [==============================] - 0s 3ms/step - loss: 0.0288 - accuracy: 0.9905\n"]}],"source":["r = model.fit(X_test, y_test, epochs=10)"]},{"cell_type":"markdown","id":"7b40abd8","metadata":{"papermill":{"duration":0.105089,"end_time":"2023-02-07T13:46:48.780972","exception":false,"start_time":"2023-02-07T13:46:48.675883","status":"completed"},"tags":[]},"source":["# 3. Predict the test images"]},{"cell_type":"code","execution_count":26,"id":"c5884df5","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:49.125636Z","iopub.status.busy":"2023-02-07T13:46:49.125195Z","iopub.status.idle":"2023-02-07T13:46:49.245032Z","shell.execute_reply":"2023-02-07T13:46:49.243738Z"},"papermill":{"duration":0.362459,"end_time":"2023-02-07T13:46:49.248033","exception":false,"start_time":"2023-02-07T13:46:48.885574","status":"completed"},"tags":[]},"outputs":[],"source":["test = test / 255"]},{"cell_type":"code","execution_count":27,"id":"ea9da0e7","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:49.545362Z","iopub.status.busy":"2023-02-07T13:46:49.544975Z","iopub.status.idle":"2023-02-07T13:46:49.55179Z","shell.execute_reply":"2023-02-07T13:46:49.550878Z"},"papermill":{"duration":0.149312,"end_time":"2023-02-07T13:46:49.554685","exception":false,"start_time":"2023-02-07T13:46:49.405373","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["(28000, 28, 28, 1)"]},"execution_count":27,"metadata":{},"output_type":"execute_result"}],"source":["test = test.values.reshape(-1, 28, 28, 1)\n","test.shape"]},{"cell_type":"code","execution_count":28,"id":"47a64e96","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:49.769625Z","iopub.status.busy":"2023-02-07T13:46:49.769296Z","iopub.status.idle":"2023-02-07T13:46:51.341415Z","shell.execute_reply":"2023-02-07T13:46:51.340265Z"},"papermill":{"duration":1.679358,"end_time":"2023-02-07T13:46:51.344052","exception":false,"start_time":"2023-02-07T13:46:49.664694","status":"completed"},"tags":[]},"outputs":[],"source":["predictions = model.predict(test)"]},{"cell_type":"code","execution_count":29,"id":"563cf9f8","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:51.584349Z","iopub.status.busy":"2023-02-07T13:46:51.583585Z","iopub.status.idle":"2023-02-07T13:46:51.596901Z","shell.execute_reply":"2023-02-07T13:46:51.595237Z"},"papermill":{"duration":0.150222,"end_time":"2023-02-07T13:46:51.599963","exception":false,"start_time":"2023-02-07T13:46:51.449741","status":"completed"},"tags":[]},"outputs":[{"data":{"text/plain":["array([2, 0, 9, ..., 3, 9, 2])"]},"execution_count":29,"metadata":{},"output_type":"execute_result"}],"source":["predictions = np.argmax(predictions, axis=1)\n","predictions"]},{"cell_type":"code","execution_count":30,"id":"fad56904","metadata":{"execution":{"iopub.execute_input":"2023-02-07T13:46:51.932663Z","iopub.status.busy":"2023-02-07T13:46:51.931808Z","iopub.status.idle":"2023-02-07T13:46:52.024236Z","shell.execute_reply":"2023-02-07T13:46:52.023268Z"},"papermill":{"duration":0.265042,"end_time":"2023-02-07T13:46:52.028857","exception":false,"start_time":"2023-02-07T13:46:51.763815","status":"completed"},"tags":[]},"outputs":[],"source":["submission = pd.read_csv('/kaggle/input/digit-recognizer/sample_submission.csv')\n","submission['Label'] = predictions\n","submission.to_csv('submission.csv', index=False)"]},{"cell_type":"code","execution_count":null,"id":"3c747976","metadata":{"papermill":{"duration":0.162581,"end_time":"2023-02-07T13:46:52.355573","exception":false,"start_time":"2023-02-07T13:46:52.192992","status":"completed"},"tags":[]},"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":"3.7.12"},"papermill":{"default_parameters":{},"duration":183.439173,"end_time":"2023-02-07T13:46:56.303854","environment_variables":{},"exception":null,"input_path":"__notebook__.ipynb","output_path":"__notebook__.ipynb","parameters":{},"start_time":"2023-02-07T13:43:52.864681","version":"2.3.4"}},"nbformat":4,"nbformat_minor":5}