{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import pandas as pd\n\ntrain = pd.read_csv(\"/kaggle/input/digit-recognizer/train.csv\")\nx = train.values[:, 1:]\ny = train.values[:, :1].reshape(42000)\n\nprint(x.shape)\nprint(y.shape)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:39:15.789713Z","iopub.execute_input":"2022-07-28T00:39:15.790193Z","iopub.status.idle":"2022-07-28T00:39:18.637750Z","shell.execute_reply.started":"2022-07-28T00:39:15.790151Z","shell.execute_reply":"2022-07-28T00:39:18.636407Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from pandas import get_dummies\nfrom matplotlib import pyplot as plt\n\nx = x.astype(\"float32\")\nx /= 225.0\nx = x.reshape(42000, 28, 28, 1)\n\nplt.imshow(x[4], cmap = \"gray\")\n\ny = get_dummies(y)\n\nprint(y.head())\n\nplt","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:39:18.641791Z","iopub.execute_input":"2022-07-28T00:39:18.642456Z","iopub.status.idle":"2022-07-28T00:39:18.913953Z","shell.execute_reply.started":"2022-07-28T00:39:18.642404Z","shell.execute_reply":"2022-07-28T00:39:18.912513Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from tensorflow.keras import Sequential\nfrom tensorflow.keras.layers import Conv2D, MaxPooling2D, Dense, Flatten\n\nmodel = Sequential()\nmodel.add(Conv2D(16,(5,5), activation = \"relu\", input_shape = (28, 28, 1), padding = \"same\"))\nmodel.add(Conv2D(16,(5,5), activation = \"relu\", padding = \"same\"))\nmodel.add(MaxPooling2D(2,2))\nmodel.add(Conv2D(16,(3,3), activation = \"relu\", padding = \"same\"))\nmodel.add(Conv2D(16,(3,3), activation = \"relu\", padding = \"same\"))\nmodel.add(MaxPooling2D(2,2))\nmodel.add(Flatten())\nmodel.add(Dense(128, activation=\"relu\"))\nmodel.add(Dense(10, activation=\"softmax\"))\n\nmodel.summary()\n\nfrom tensorflow.keras.losses import categorical_crossentropy\n\nmodel.compile(loss = categorical_crossentropy, optimizer = \"Adam\", metrics = [\"accuracy\"])","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:39:18.915308Z","iopub.execute_input":"2022-07-28T00:39:18.915674Z","iopub.status.idle":"2022-07-28T00:39:19.006692Z","shell.execute_reply.started":"2022-07-28T00:39:18.915642Z","shell.execute_reply":"2022-07-28T00:39:19.005482Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import train_test_split\nx_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.125)\n\n\nmodel_stats = model.fit(\n    x_train,\n    y_train,\n    batch_size = 196,\n    epochs = 6,\n    verbose = 1,\n    validation_data =(x_test, y_test)\n)\n","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:39:19.009406Z","iopub.execute_input":"2022-07-28T00:39:19.011690Z","iopub.status.idle":"2022-07-28T00:41:00.218137Z","shell.execute_reply.started":"2022-07-28T00:39:19.011635Z","shell.execute_reply":"2022-07-28T00:41:00.217152Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"plt.plot(model_stats.history[\"loss\"])\nplt.plot(model_stats.history[\"val_loss\"])\nplt.legend([\"train loss\", \"test loss\"], loc=\"upper left\")","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:41:00.220047Z","iopub.execute_input":"2022-07-28T00:41:00.220910Z","iopub.status.idle":"2022-07-28T00:41:00.431985Z","shell.execute_reply.started":"2022-07-28T00:41:00.220863Z","shell.execute_reply":"2022-07-28T00:41:00.430889Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import numpy as np\n\nx_pred = pd.read_csv(\"/kaggle/input/digit-recognizer/test.csv\")\n\nx_pred = np.asarray(x_pred)\nx_pred = x_pred.astype(\"float32\")\nx_pred /= 225.0\nx_pred = x_pred.reshape(x_pred.shape[0], 28, 28, 1)\n\n\n\n\ny_pred = model.predict(x_pred)\ny_pred = [np.argmax(v) for v in y_pred]\n\ndata = {\n    \"ImageId\": [x for x in range(1, 28001)],\n    \"Label\": y_pred\n}\nsubmission = pd.DataFrame(data)\n\nsubmission.to_csv(\"submission.csv\", index=False)\n","metadata":{"execution":{"iopub.status.busy":"2022-07-28T00:41:00.433394Z","iopub.execute_input":"2022-07-28T00:41:00.433827Z","iopub.status.idle":"2022-07-28T00:41:09.332571Z","shell.execute_reply.started":"2022-07-28T00:41:00.433796Z","shell.execute_reply":"2022-07-28T00:41:09.331396Z"},"trusted":true},"execution_count":null,"outputs":[]}]}