CNN_MNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "CNN MNIST",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/rachit-shah/ADBI-projects/blob/master/CNN_MNIST.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "55emSiBngazt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from mlxtend.data import loadlocal_mnist\n",
        "import pickle\n",
        "import random\n",
        "from sklearn.model_selection import train_test_split\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "\n",
        "from keras.layers import Conv2D, MaxPool2D, Dense, Dropout, Flatten\n",
        "from keras import Model,Sequential\n",
        "import keras\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\n",
        "from sklearn.metrics import confusion_matrix, precision_recall_fscore_support, accuracy_score"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UuMvl_qXgiXH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Connect to Google Drive on Colab\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iP2Y35BD6Ejb",
        "colab_type": "code",
        "outputId": "497cdbde-e652-4cc3-a2ec-330d048a2d26",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 798
        }
      },
      "source": [
        "!wget http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n",
        "!wget http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n",
        "!wget http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n",
        "!wget http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2019-10-14 18:46:10--  http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n",
            "Resolving yann.lecun.com (yann.lecun.com)... 216.165.22.6\n",
            "Connecting to yann.lecun.com (yann.lecun.com)|216.165.22.6|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 9912422 (9.5M) [application/x-gzip]\n",
            "Saving to: ‘train-images-idx3-ubyte.gz’\n",
            "\n",
            "train-images-idx3-u 100%[===================>]   9.45M  5.06MB/s    in 1.9s    \n",
            "\n",
            "2019-10-14 18:46:13 (5.06 MB/s) - ‘train-images-idx3-ubyte.gz’ saved [9912422/9912422]\n",
            "\n",
            "--2019-10-14 18:46:13--  http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n",
            "Resolving yann.lecun.com (yann.lecun.com)... 216.165.22.6\n",
            "Connecting to yann.lecun.com (yann.lecun.com)|216.165.22.6|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 28881 (28K) [application/x-gzip]\n",
            "Saving to: ‘train-labels-idx1-ubyte.gz’\n",
            "\n",
            "train-labels-idx1-u 100%[===================>]  28.20K  --.-KB/s    in 0.08s   \n",
            "\n",
            "2019-10-14 18:46:14 (336 KB/s) - ‘train-labels-idx1-ubyte.gz’ saved [28881/28881]\n",
            "\n",
            "--2019-10-14 18:46:14--  http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n",
            "Resolving yann.lecun.com (yann.lecun.com)... 216.165.22.6\n",
            "Connecting to yann.lecun.com (yann.lecun.com)|216.165.22.6|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 1648877 (1.6M) [application/x-gzip]\n",
            "Saving to: ‘t10k-images-idx3-ubyte.gz’\n",
            "\n",
            "t10k-images-idx3-ub 100%[===================>]   1.57M   109KB/s    in 14s     \n",
            "\n",
            "2019-10-14 18:46:29 (112 KB/s) - ‘t10k-images-idx3-ubyte.gz’ saved [1648877/1648877]\n",
            "\n",
            "--2019-10-14 18:46:29--  http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\n",
            "Resolving yann.lecun.com (yann.lecun.com)... 216.165.22.6\n",
            "Connecting to yann.lecun.com (yann.lecun.com)|216.165.22.6|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4542 (4.4K) [application/x-gzip]\n",
            "Saving to: ‘t10k-labels-idx1-ubyte.gz’\n",
            "\n",
            "t10k-labels-idx1-ub 100%[===================>]   4.44K  --.-KB/s    in 0s      \n",
            "\n",
            "2019-10-14 18:46:30 (332 MB/s) - ‘t10k-labels-idx1-ubyte.gz’ saved [4542/4542]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zm_2QWrW6GC4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!gunzip *.gz"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "croKZX9F-Pn5",
        "colab_type": "code",
        "outputId": "f0105e1a-9674-4b42-c067-ce9fe6a164c1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 158
        }
      },
      "source": [
        "!ls -l"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total 54796\n",
            "drwx------ 3 root root     4096 Oct 14 18:43 drive\n",
            "-rw------- 1 root root  1142313 Oct 14 18:45 grid.pkl\n",
            "drwxr-xr-x 1 root root     4096 Aug 27 16:17 sample_data\n",
            "-rw-r--r-- 1 root root  7840016 Jul 21  2000 t10k-images-idx3-ubyte\n",
            "-rw-r--r-- 1 root root    10008 Jul 21  2000 t10k-labels-idx1-ubyte\n",
            "-rw-r--r-- 1 root root 47040016 Jul 21  2000 train-images-idx3-ubyte\n",
            "-rw-r--r-- 1 root root    60008 Jul 21  2000 train-labels-idx1-ubyte\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LqDMGiEj_9Qb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X, y = loadlocal_mnist(images_path='train-images-idx3-ubyte', \n",
        "                      labels_path='train-labels-idx1-ubyte')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zRn2bPvNHT4F",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = X.reshape(len(X),28,28,1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vTB9vCzokVZM",
        "colab_type": "code",
        "outputId": "c5644352-c99c-4456-c547-2ec7b4e59e05",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X = X.astype('float32')\n",
        "X /= 255\n",
        "print('x_train shape:', X.shape)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "x_train shape: (60000, 28, 28, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S5Jr5jwUHhtc",
        "colab_type": "code",
        "outputId": "58871761-79b9-4748-a18f-10ef58a990fc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X.shape"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28, 1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uSQgBnGiAGID",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "testX, testY = loadlocal_mnist(images_path='t10k-images-idx3-ubyte', \n",
        "                              labels_path='t10k-labels-idx1-ubyte')\n",
        "testX = testX.reshape(len(testX),28,28,1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qso7ePUBAXIS",
        "colab_type": "text"
      },
      "source": [
        "#Splitting into train and validation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uGYceLvjAwBd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "VAL_SPLIT = 0.2\n",
        "num_classes=10"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yh5HgGg4Bjg4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "trainX, valX, trainY, valY = train_test_split(X,y,test_size=VAL_SPLIT)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vApbietbEnqV",
        "colab_type": "text"
      },
      "source": [
        "# Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F7QSBPRPiBwS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#parameters\n",
        "act=None #activation\n",
        "lr=0.001 #learning rate for adam optimizer\n",
        "epoch=100 #number of epochs to train\n",
        "drop=0.5 #dropout"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aviDoNhSEpuz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_model(act=None, lr=0.001, drop=0.5):\n",
        "  model = Sequential()\n",
        "  model.add(Conv2D(filters=32, kernel_size=(3,3), activation=act, input_shape=(28,28,1)))\n",
        "  model.add(MaxPool2D(pool_size=(2,2)))\n",
        "  model.add(Conv2D(filters=64, kernel_size=(3,3), activation=act))\n",
        "  model.add(MaxPool2D(pool_size=(2,2)))\n",
        "  model.add(Conv2D(filters=64, kernel_size=(3,3), activation=act))\n",
        "  model.add(Flatten())\n",
        "  model.add(Dense(units=64,activation=act))\n",
        "  if drop>0:\n",
        "    model.add(Dropout(drop))\n",
        "  model.add(Dense(units=10,activation='softmax'))\n",
        "\n",
        "  opt = keras.optimizers.Adam(lr=lr)\n",
        "  model.compile(optimizer=opt,\n",
        "                loss='sparse_categorical_crossentropy',\n",
        "                metrics=['accuracy'])\n",
        "  return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NvsEIzIEhqQJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = KerasClassifier(build_fn=create_model, epochs=epoch, act=act, lr=lr,drop=drop)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JT4DVqL5h2pF",
        "colab_type": "code",
        "outputId": "08585902-bfb9-4833-819c-eae5f911e4fd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "model.fit(trainX,trainY, verbose=True, validation_data=(valX,valY))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4267: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:148: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3622: The name tf.log is deprecated. Please use tf.math.log instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/ops/math_grad.py:1424: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1033: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1020: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3005: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n",
            "\n",
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/100\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:207: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:216: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:223: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n",
            "\n",
            "48000/48000 [==============================] - 20s 411us/step - loss: 0.2283 - acc: 0.9296 - val_loss: 0.1028 - val_acc: 0.9681\n",
            "Epoch 2/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.1113 - acc: 0.9667 - val_loss: 0.0804 - val_acc: 0.9755\n",
            "Epoch 3/100\n",
            "48000/48000 [==============================] - 13s 266us/step - loss: 0.0917 - acc: 0.9727 - val_loss: 0.0756 - val_acc: 0.9780\n",
            "Epoch 4/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0801 - acc: 0.9750 - val_loss: 0.0663 - val_acc: 0.9815\n",
            "Epoch 5/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0716 - acc: 0.9782 - val_loss: 0.0636 - val_acc: 0.9808\n",
            "Epoch 6/100\n",
            "48000/48000 [==============================] - 13s 266us/step - loss: 0.0623 - acc: 0.9808 - val_loss: 0.0631 - val_acc: 0.9819\n",
            "Epoch 7/100\n",
            "48000/48000 [==============================] - 13s 268us/step - loss: 0.0614 - acc: 0.9816 - val_loss: 0.0722 - val_acc: 0.9800\n",
            "Epoch 8/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0540 - acc: 0.9835 - val_loss: 0.0678 - val_acc: 0.9845\n",
            "Epoch 9/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0559 - acc: 0.9834 - val_loss: 0.0941 - val_acc: 0.9752\n",
            "Epoch 10/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0517 - acc: 0.9839 - val_loss: 0.0709 - val_acc: 0.9830\n",
            "Epoch 11/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0481 - acc: 0.9859 - val_loss: 0.0728 - val_acc: 0.9839\n",
            "Epoch 12/100\n",
            "48000/48000 [==============================] - 13s 262us/step - loss: 0.0493 - acc: 0.9858 - val_loss: 0.0676 - val_acc: 0.9840\n",
            "Epoch 13/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0455 - acc: 0.9865 - val_loss: 0.0613 - val_acc: 0.9856\n",
            "Epoch 14/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0424 - acc: 0.9875 - val_loss: 0.0850 - val_acc: 0.9817\n",
            "Epoch 15/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0449 - acc: 0.9868 - val_loss: 0.1153 - val_acc: 0.9735\n",
            "Epoch 16/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0424 - acc: 0.9875 - val_loss: 0.0670 - val_acc: 0.9856\n",
            "Epoch 17/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0431 - acc: 0.9876 - val_loss: 0.0689 - val_acc: 0.9862\n",
            "Epoch 18/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0399 - acc: 0.9885 - val_loss: 0.0765 - val_acc: 0.9840\n",
            "Epoch 19/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0431 - acc: 0.9876 - val_loss: 0.0734 - val_acc: 0.9828\n",
            "Epoch 20/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0387 - acc: 0.9883 - val_loss: 0.0764 - val_acc: 0.9854\n",
            "Epoch 21/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0410 - acc: 0.9883 - val_loss: 0.0980 - val_acc: 0.9796\n",
            "Epoch 22/100\n",
            "48000/48000 [==============================] - 13s 262us/step - loss: 0.0432 - acc: 0.9880 - val_loss: 0.0861 - val_acc: 0.9832\n",
            "Epoch 23/100\n",
            "48000/48000 [==============================] - 13s 272us/step - loss: 0.0360 - acc: 0.9895 - val_loss: 0.0761 - val_acc: 0.9838\n",
            "Epoch 24/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0363 - acc: 0.9894 - val_loss: 0.0910 - val_acc: 0.9818\n",
            "Epoch 25/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0397 - acc: 0.9894 - val_loss: 0.0778 - val_acc: 0.9845\n",
            "Epoch 26/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0430 - acc: 0.9887 - val_loss: 0.0937 - val_acc: 0.9843\n",
            "Epoch 27/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0378 - acc: 0.9895 - val_loss: 0.0981 - val_acc: 0.9818\n",
            "Epoch 28/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0387 - acc: 0.9898 - val_loss: 0.0865 - val_acc: 0.9845\n",
            "Epoch 29/100\n",
            "48000/48000 [==============================] - 13s 262us/step - loss: 0.0347 - acc: 0.9902 - val_loss: 0.1035 - val_acc: 0.9828\n",
            "Epoch 30/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0386 - acc: 0.9895 - val_loss: 0.0987 - val_acc: 0.9849\n",
            "Epoch 31/100\n",
            "48000/48000 [==============================] - 13s 268us/step - loss: 0.0396 - acc: 0.9903 - val_loss: 0.1154 - val_acc: 0.9828\n",
            "Epoch 32/100\n",
            "48000/48000 [==============================] - 13s 267us/step - loss: 0.0419 - acc: 0.9898 - val_loss: 0.0828 - val_acc: 0.9853\n",
            "Epoch 33/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0382 - acc: 0.9907 - val_loss: 0.1139 - val_acc: 0.9842\n",
            "Epoch 34/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0425 - acc: 0.9897 - val_loss: 0.0976 - val_acc: 0.9848\n",
            "Epoch 35/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0388 - acc: 0.9903 - val_loss: 0.0970 - val_acc: 0.9855\n",
            "Epoch 36/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0387 - acc: 0.9906 - val_loss: 0.1091 - val_acc: 0.9851\n",
            "Epoch 37/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0387 - acc: 0.9913 - val_loss: 0.0966 - val_acc: 0.9858\n",
            "Epoch 38/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0402 - acc: 0.9903 - val_loss: 0.1132 - val_acc: 0.9832\n",
            "Epoch 39/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0449 - acc: 0.9899 - val_loss: 0.1137 - val_acc: 0.9831\n",
            "Epoch 40/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0423 - acc: 0.9908 - val_loss: 0.1318 - val_acc: 0.9818\n",
            "Epoch 41/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0397 - acc: 0.9914 - val_loss: 0.1296 - val_acc: 0.9813\n",
            "Epoch 42/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0410 - acc: 0.9908 - val_loss: 0.1363 - val_acc: 0.9810\n",
            "Epoch 43/100\n",
            "48000/48000 [==============================] - 13s 267us/step - loss: 0.0392 - acc: 0.9909 - val_loss: 0.1329 - val_acc: 0.9818\n",
            "Epoch 44/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0462 - acc: 0.9910 - val_loss: 0.1441 - val_acc: 0.9822\n",
            "Epoch 45/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0504 - acc: 0.9907 - val_loss: 0.1271 - val_acc: 0.9823\n",
            "Epoch 46/100\n",
            "48000/48000 [==============================] - 13s 266us/step - loss: 0.0413 - acc: 0.9914 - val_loss: 0.1321 - val_acc: 0.9834\n",
            "Epoch 47/100\n",
            "48000/48000 [==============================] - 13s 268us/step - loss: 0.0431 - acc: 0.9911 - val_loss: 0.1547 - val_acc: 0.9800\n",
            "Epoch 48/100\n",
            "48000/48000 [==============================] - 13s 267us/step - loss: 0.0459 - acc: 0.9914 - val_loss: 0.1534 - val_acc: 0.9811\n",
            "Epoch 49/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0504 - acc: 0.9904 - val_loss: 0.1346 - val_acc: 0.9823\n",
            "Epoch 50/100\n",
            "48000/48000 [==============================] - 13s 266us/step - loss: 0.0477 - acc: 0.9910 - val_loss: 0.1599 - val_acc: 0.9786\n",
            "Epoch 51/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0492 - acc: 0.9907 - val_loss: 0.1590 - val_acc: 0.9803\n",
            "Epoch 52/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.0488 - acc: 0.9911 - val_loss: 0.1408 - val_acc: 0.9822\n",
            "Epoch 53/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0508 - acc: 0.9908 - val_loss: 0.1734 - val_acc: 0.9820\n",
            "Epoch 54/100\n",
            "48000/48000 [==============================] - 13s 263us/step - loss: 0.0533 - acc: 0.9910 - val_loss: 0.1800 - val_acc: 0.9811\n",
            "Epoch 55/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0545 - acc: 0.9907 - val_loss: 0.1682 - val_acc: 0.9814\n",
            "Epoch 56/100\n",
            "48000/48000 [==============================] - 13s 264us/step - loss: 0.0540 - acc: 0.9917 - val_loss: 0.1336 - val_acc: 0.9858\n",
            "Epoch 57/100\n",
            "48000/48000 [==============================] - 13s 260us/step - loss: 0.0550 - acc: 0.9912 - val_loss: 0.2017 - val_acc: 0.9776\n",
            "Epoch 58/100\n",
            "48000/48000 [==============================] - 13s 261us/step - loss: 0.0583 - acc: 0.9911 - val_loss: 0.1594 - val_acc: 0.9823\n",
            "Epoch 59/100\n",
            "48000/48000 [==============================] - 13s 261us/step - loss: 0.0567 - acc: 0.9915 - val_loss: 0.1509 - val_acc: 0.9842\n",
            "Epoch 60/100\n",
            "48000/48000 [==============================] - 12s 259us/step - loss: 0.0668 - acc: 0.9907 - val_loss: 0.1818 - val_acc: 0.9804\n",
            "Epoch 61/100\n",
            "48000/48000 [==============================] - 12s 260us/step - loss: 0.0698 - acc: 0.9905 - val_loss: 0.1766 - val_acc: 0.9820\n",
            "Epoch 62/100\n",
            "48000/48000 [==============================] - 12s 259us/step - loss: 0.0679 - acc: 0.9910 - val_loss: 0.1929 - val_acc: 0.9816\n",
            "Epoch 63/100\n",
            "48000/48000 [==============================] - 13s 261us/step - loss: 0.0709 - acc: 0.9909 - val_loss: 0.1612 - val_acc: 0.9842\n",
            "Epoch 64/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.0888 - acc: 0.9897 - val_loss: 0.1827 - val_acc: 0.9838\n",
            "Epoch 65/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.0624 - acc: 0.9924 - val_loss: 0.2223 - val_acc: 0.9801\n",
            "Epoch 66/100\n",
            "48000/48000 [==============================] - 12s 256us/step - loss: 0.0892 - acc: 0.9899 - val_loss: 0.2115 - val_acc: 0.9829\n",
            "Epoch 67/100\n",
            "48000/48000 [==============================] - 12s 256us/step - loss: 0.0854 - acc: 0.9904 - val_loss: 0.1980 - val_acc: 0.9831\n",
            "Epoch 68/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.0925 - acc: 0.9899 - val_loss: 0.2669 - val_acc: 0.9778\n",
            "Epoch 69/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.0962 - acc: 0.9901 - val_loss: 0.2079 - val_acc: 0.9832\n",
            "Epoch 70/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.1058 - acc: 0.9902 - val_loss: 0.2555 - val_acc: 0.9798\n",
            "Epoch 71/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.1304 - acc: 0.9884 - val_loss: 0.2724 - val_acc: 0.9792\n",
            "Epoch 72/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.1305 - acc: 0.9889 - val_loss: 0.2582 - val_acc: 0.9807\n",
            "Epoch 73/100\n",
            "48000/48000 [==============================] - 12s 260us/step - loss: 0.1460 - acc: 0.9876 - val_loss: 0.2148 - val_acc: 0.9843\n",
            "Epoch 74/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.1494 - acc: 0.9879 - val_loss: 0.2302 - val_acc: 0.9832\n",
            "Epoch 75/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.1687 - acc: 0.9870 - val_loss: 0.2389 - val_acc: 0.9827\n",
            "Epoch 76/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.1905 - acc: 0.9854 - val_loss: 0.3208 - val_acc: 0.9769\n",
            "Epoch 77/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.1957 - acc: 0.9856 - val_loss: 0.2472 - val_acc: 0.9822\n",
            "Epoch 78/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.1886 - acc: 0.9863 - val_loss: 0.2968 - val_acc: 0.9798\n",
            "Epoch 79/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.1997 - acc: 0.9856 - val_loss: 0.2468 - val_acc: 0.9833\n",
            "Epoch 80/100\n",
            "48000/48000 [==============================] - 13s 261us/step - loss: 0.2330 - acc: 0.9836 - val_loss: 0.2840 - val_acc: 0.9808\n",
            "Epoch 81/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.2808 - acc: 0.9810 - val_loss: 0.3300 - val_acc: 0.9782\n",
            "Epoch 82/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.2330 - acc: 0.9840 - val_loss: 0.2957 - val_acc: 0.9802\n",
            "Epoch 83/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.2994 - acc: 0.9799 - val_loss: 0.3480 - val_acc: 0.9772\n",
            "Epoch 84/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.2872 - acc: 0.9810 - val_loss: 0.3319 - val_acc: 0.9783\n",
            "Epoch 85/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.2905 - acc: 0.9806 - val_loss: 0.3319 - val_acc: 0.9784\n",
            "Epoch 86/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.2970 - acc: 0.9802 - val_loss: 0.2745 - val_acc: 0.9826\n",
            "Epoch 87/100\n",
            "48000/48000 [==============================] - 12s 259us/step - loss: 0.3170 - acc: 0.9791 - val_loss: 0.3300 - val_acc: 0.9788\n",
            "Epoch 88/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.3166 - acc: 0.9794 - val_loss: 0.3068 - val_acc: 0.9802\n",
            "Epoch 89/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.3555 - acc: 0.9769 - val_loss: 0.2996 - val_acc: 0.9807\n",
            "Epoch 90/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.3651 - acc: 0.9763 - val_loss: 0.4596 - val_acc: 0.9707\n",
            "Epoch 91/100\n",
            "48000/48000 [==============================] - 12s 258us/step - loss: 0.3539 - acc: 0.9772 - val_loss: 0.3390 - val_acc: 0.9779\n",
            "Epoch 92/100\n",
            "48000/48000 [==============================] - 12s 259us/step - loss: 0.3812 - acc: 0.9756 - val_loss: 0.3547 - val_acc: 0.9775\n",
            "Epoch 93/100\n",
            "48000/48000 [==============================] - 12s 257us/step - loss: 0.4157 - acc: 0.9735 - val_loss: 0.4505 - val_acc: 0.9718\n",
            "Epoch 94/100\n",
            "48000/48000 [==============================] - 15s 322us/step - loss: 0.4356 - acc: 0.9722 - val_loss: 0.3864 - val_acc: 0.9754\n",
            "Epoch 95/100\n",
            "48000/48000 [==============================] - 13s 265us/step - loss: 0.4733 - acc: 0.9699 - val_loss: 0.4243 - val_acc: 0.9733\n",
            "Epoch 96/100\n",
            "48000/48000 [==============================] - 13s 267us/step - loss: 0.4893 - acc: 0.9691 - val_loss: 0.4345 - val_acc: 0.9726\n",
            "Epoch 97/100\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.4338 - acc: 0.9725 - val_loss: 0.4179 - val_acc: 0.9733\n",
            "Epoch 98/100\n",
            "48000/48000 [==============================] - 13s 270us/step - loss: 0.5329 - acc: 0.9661 - val_loss: 0.4107 - val_acc: 0.9741\n",
            "Epoch 99/100\n",
            "48000/48000 [==============================] - 13s 269us/step - loss: 0.4869 - acc: 0.9694 - val_loss: 0.4191 - val_acc: 0.9736\n",
            "Epoch 100/100\n",
            "48000/48000 [==============================] - 13s 270us/step - loss: 0.5388 - acc: 0.9660 - val_loss: 0.5629 - val_acc: 0.9644\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f71d0d527f0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "csVLfukCpRcK",
        "colab_type": "text"
      },
      "source": [
        "# Hyperparameter Tuning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZyN6jNjXlMAa",
        "colab_type": "text"
      },
      "source": [
        "## Tune Learning Rate, Epochs and Dropout"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tEG0Wsv8lKYA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#parameters\n",
        "lr= [0.01,0.001,0.0001]\n",
        "epochs=[25,50,100]\n",
        "drop=[0,0.2,0.3,0.5]\n",
        "param_grid = dict(lr=lr,epochs=epochs,drop=drop)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EAjWfnrEputv",
        "colab_type": "code",
        "outputId": "c1e5be37-6b8f-44d0-d850-25a596682100",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "grid = GridSearchCV(estimator=model, param_grid=param_grid, n_jobs=-1, cv=2, verbose=True)  #GridSearch with 1-fold Cross Validation\n",
        "grid_result = grid.fit(X, y)\n",
        "# summarize results\n",
        "print(\"Best: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))\n",
        "means = grid_result.cv_results_['mean_test_score']\n",
        "stds = grid_result.cv_results_['std_test_score']\n",
        "params = grid_result.cv_results_['params']\n",
        "for mean, stdev, param in zip(means, stds, params):\n",
        "    print(\"%f (%f) with: %r\" % (mean, stdev, param))"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 2 folds for each of 36 candidates, totalling 72 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "/usr/local/lib/python3.6/dist-packages/joblib/externals/loky/process_executor.py:706: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
            "  \"timeout or by a memory leak.\", UserWarning\n",
            "[Parallel(n_jobs=-1)]: Done  46 tasks      | elapsed: 192.0min\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4267: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:148: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3622: The name tf.log is deprecated. Please use tf.math.log instead.\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  72 out of  72 | elapsed: 332.5min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/ops/math_grad.py:1424: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1033: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1020: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3005: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n",
            "\n",
            "Epoch 1/50\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:207: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:216: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:223: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n",
            "\n",
            "60000/60000 [==============================] - 18s 307us/step - loss: 0.4973 - acc: 0.8517\n",
            "Epoch 2/50\n",
            "60000/60000 [==============================] - 16s 264us/step - loss: 0.1772 - acc: 0.9478\n",
            "Epoch 3/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.1292 - acc: 0.9626\n",
            "Epoch 4/50\n",
            "60000/60000 [==============================] - 16s 271us/step - loss: 0.1075 - acc: 0.9680\n",
            "Epoch 5/50\n",
            "60000/60000 [==============================] - 16s 264us/step - loss: 0.0959 - acc: 0.9711\n",
            "Epoch 6/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0866 - acc: 0.9737\n",
            "Epoch 7/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0787 - acc: 0.9767\n",
            "Epoch 8/50\n",
            "60000/60000 [==============================] - 17s 279us/step - loss: 0.0743 - acc: 0.9781\n",
            "Epoch 9/50\n",
            "60000/60000 [==============================] - 16s 262us/step - loss: 0.0704 - acc: 0.9790\n",
            "Epoch 10/50\n",
            "60000/60000 [==============================] - 16s 261us/step - loss: 0.0663 - acc: 0.9803\n",
            "Epoch 11/50\n",
            "60000/60000 [==============================] - 15s 250us/step - loss: 0.0608 - acc: 0.9816\n",
            "Epoch 12/50\n",
            "60000/60000 [==============================] - 15s 246us/step - loss: 0.0581 - acc: 0.9825\n",
            "Epoch 13/50\n",
            "60000/60000 [==============================] - 15s 249us/step - loss: 0.0555 - acc: 0.9834\n",
            "Epoch 14/50\n",
            "60000/60000 [==============================] - 15s 249us/step - loss: 0.0521 - acc: 0.9838\n",
            "Epoch 15/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0498 - acc: 0.9845\n",
            "Epoch 16/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0473 - acc: 0.9856\n",
            "Epoch 17/50\n",
            "60000/60000 [==============================] - 15s 249us/step - loss: 0.0463 - acc: 0.9862\n",
            "Epoch 18/50\n",
            "60000/60000 [==============================] - 15s 248us/step - loss: 0.0432 - acc: 0.9873\n",
            "Epoch 19/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0419 - acc: 0.9868\n",
            "Epoch 20/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0412 - acc: 0.9881\n",
            "Epoch 21/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0390 - acc: 0.9879\n",
            "Epoch 22/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0370 - acc: 0.9889\n",
            "Epoch 23/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0356 - acc: 0.9890\n",
            "Epoch 24/50\n",
            "60000/60000 [==============================] - 15s 246us/step - loss: 0.0351 - acc: 0.9892\n",
            "Epoch 25/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0348 - acc: 0.9895\n",
            "Epoch 26/50\n",
            "60000/60000 [==============================] - 15s 246us/step - loss: 0.0315 - acc: 0.9901\n",
            "Epoch 27/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0313 - acc: 0.9901\n",
            "Epoch 28/50\n",
            "60000/60000 [==============================] - 15s 247us/step - loss: 0.0302 - acc: 0.9904\n",
            "Epoch 29/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0288 - acc: 0.9908\n",
            "Epoch 30/50\n",
            "60000/60000 [==============================] - 15s 248us/step - loss: 0.0288 - acc: 0.9908\n",
            "Epoch 31/50\n",
            "60000/60000 [==============================] - 15s 248us/step - loss: 0.0267 - acc: 0.9912\n",
            "Epoch 32/50\n",
            "60000/60000 [==============================] - 15s 246us/step - loss: 0.0268 - acc: 0.9913\n",
            "Epoch 33/50\n",
            "60000/60000 [==============================] - 15s 245us/step - loss: 0.0248 - acc: 0.9922\n",
            "Epoch 34/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0231 - acc: 0.9927\n",
            "Epoch 35/50\n",
            "60000/60000 [==============================] - 15s 249us/step - loss: 0.0239 - acc: 0.9922\n",
            "Epoch 36/50\n",
            "60000/60000 [==============================] - 15s 250us/step - loss: 0.0230 - acc: 0.9926\n",
            "Epoch 37/50\n",
            "60000/60000 [==============================] - 15s 250us/step - loss: 0.0231 - acc: 0.9927\n",
            "Epoch 38/50\n",
            "60000/60000 [==============================] - 15s 250us/step - loss: 0.0208 - acc: 0.9932\n",
            "Epoch 39/50\n",
            "60000/60000 [==============================] - 15s 248us/step - loss: 0.0220 - acc: 0.9927\n",
            "Epoch 40/50\n",
            "60000/60000 [==============================] - 15s 246us/step - loss: 0.0186 - acc: 0.9940\n",
            "Epoch 41/50\n",
            "60000/60000 [==============================] - 15s 250us/step - loss: 0.0201 - acc: 0.9937\n",
            "Epoch 42/50\n",
            "60000/60000 [==============================] - 15s 248us/step - loss: 0.0191 - acc: 0.9937\n",
            "Epoch 43/50\n",
            "60000/60000 [==============================] - 15s 257us/step - loss: 0.0177 - acc: 0.9944\n",
            "Epoch 44/50\n",
            "60000/60000 [==============================] - 15s 251us/step - loss: 0.0170 - acc: 0.9944\n",
            "Epoch 45/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0187 - acc: 0.9940\n",
            "Epoch 46/50\n",
            "60000/60000 [==============================] - 15s 252us/step - loss: 0.0172 - acc: 0.9945\n",
            "Epoch 47/50\n",
            "60000/60000 [==============================] - 16s 260us/step - loss: 0.0156 - acc: 0.9948\n",
            "Epoch 48/50\n",
            "60000/60000 [==============================] - 16s 259us/step - loss: 0.0156 - acc: 0.9945\n",
            "Epoch 49/50\n",
            "60000/60000 [==============================] - 16s 260us/step - loss: 0.0150 - acc: 0.9950\n",
            "Epoch 50/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0148 - acc: 0.9950\n",
            "Best: 0.985583 using {'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.102250 (0.011850) with: {'drop': 0, 'epochs': 25, 'lr': 0.01}\n",
            "0.976433 (0.001800) with: {'drop': 0, 'epochs': 25, 'lr': 0.001}\n",
            "0.981567 (0.000900) with: {'drop': 0, 'epochs': 25, 'lr': 0.0001}\n",
            "0.097517 (0.001683) with: {'drop': 0, 'epochs': 50, 'lr': 0.01}\n",
            "0.977850 (0.001450) with: {'drop': 0, 'epochs': 50, 'lr': 0.001}\n",
            "0.982933 (0.000567) with: {'drop': 0, 'epochs': 50, 'lr': 0.0001}\n",
            "0.103850 (0.001417) with: {'drop': 0, 'epochs': 100, 'lr': 0.01}\n",
            "0.972117 (0.001450) with: {'drop': 0, 'epochs': 100, 'lr': 0.001}\n",
            "0.984217 (0.001083) with: {'drop': 0, 'epochs': 100, 'lr': 0.0001}\n",
            "0.102750 (0.000817) with: {'drop': 0.2, 'epochs': 25, 'lr': 0.01}\n",
            "0.977600 (0.002533) with: {'drop': 0.2, 'epochs': 25, 'lr': 0.001}\n",
            "0.982217 (0.000783) with: {'drop': 0.2, 'epochs': 25, 'lr': 0.0001}\n",
            "0.100100 (0.001833) with: {'drop': 0.2, 'epochs': 50, 'lr': 0.01}\n",
            "0.977633 (0.000733) with: {'drop': 0.2, 'epochs': 50, 'lr': 0.001}\n",
            "0.984017 (0.001117) with: {'drop': 0.2, 'epochs': 50, 'lr': 0.0001}\n",
            "0.100217 (0.000117) with: {'drop': 0.2, 'epochs': 100, 'lr': 0.01}\n",
            "0.963833 (0.009933) with: {'drop': 0.2, 'epochs': 100, 'lr': 0.001}\n",
            "0.984033 (0.000333) with: {'drop': 0.2, 'epochs': 100, 'lr': 0.0001}\n",
            "0.101767 (0.003500) with: {'drop': 0.3, 'epochs': 25, 'lr': 0.01}\n",
            "0.976517 (0.002250) with: {'drop': 0.3, 'epochs': 25, 'lr': 0.001}\n",
            "0.981167 (0.000967) with: {'drop': 0.3, 'epochs': 25, 'lr': 0.0001}\n",
            "0.094550 (0.004150) with: {'drop': 0.3, 'epochs': 50, 'lr': 0.01}\n",
            "0.976333 (0.002033) with: {'drop': 0.3, 'epochs': 50, 'lr': 0.001}\n",
            "0.984267 (0.000533) with: {'drop': 0.3, 'epochs': 50, 'lr': 0.0001}\n",
            "0.098133 (0.000600) with: {'drop': 0.3, 'epochs': 100, 'lr': 0.01}\n",
            "0.971817 (0.003317) with: {'drop': 0.3, 'epochs': 100, 'lr': 0.001}\n",
            "0.984933 (0.001100) with: {'drop': 0.3, 'epochs': 100, 'lr': 0.0001}\n",
            "0.098017 (0.000250) with: {'drop': 0.5, 'epochs': 25, 'lr': 0.01}\n",
            "0.976833 (0.000433) with: {'drop': 0.5, 'epochs': 25, 'lr': 0.001}\n",
            "0.983567 (0.001567) with: {'drop': 0.5, 'epochs': 25, 'lr': 0.0001}\n",
            "0.101767 (0.003500) with: {'drop': 0.5, 'epochs': 50, 'lr': 0.01}\n",
            "0.980300 (0.000567) with: {'drop': 0.5, 'epochs': 50, 'lr': 0.001}\n",
            "0.985583 (0.000583) with: {'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.096117 (0.005817) with: {'drop': 0.5, 'epochs': 100, 'lr': 0.01}\n",
            "0.975900 (0.002100) with: {'drop': 0.5, 'epochs': 100, 'lr': 0.001}\n",
            "0.984000 (0.001133) with: {'drop': 0.5, 'epochs': 100, 'lr': 0.0001}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x3TpWZcjUTeE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pickle.dump(grid_result, open('grid2.pkl','wb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tr28fl_7UZPU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp grid.pkl drive/My\\ Drive"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JjCRhhcElsWr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp drive/My\\ Drive/grid.pkl ."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6M92Q8BJlwBs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "grid_results = pickle.load(open('grid.pkl','rb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9rZQrcx8qwEQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "parameters = grid_results.param_grid\n",
        "keys = list(parameters.keys())\n",
        "for x in keys:\n",
        "  grid_results.cv_results_[x] = grid_results.cv_results_.pop(\"param_\"+x)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pjTsY0ppr3St",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0435ef79-0945-4d3d-9d38-eeb2937ebffe"
      },
      "source": [
        "score"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mean_fit_time</th>\n",
              "      <th>std_fit_time</th>\n",
              "      <th>mean_score_time</th>\n",
              "      <th>std_score_time</th>\n",
              "      <th>params</th>\n",
              "      <th>split0_test_score</th>\n",
              "      <th>split1_test_score</th>\n",
              "      <th>mean_test_score</th>\n",
              "      <th>std_test_score</th>\n",
              "      <th>rank_test_score</th>\n",
              "      <th>lr</th>\n",
              "      <th>epochs</th>\n",
              "      <th>drop</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>471.794007</td>\n",
              "      <td>0.708622</td>\n",
              "      <td>3.902438</td>\n",
              "      <td>0.107943</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 50, 'lr': 0.0001}</td>\n",
              "      <td>0.986167</td>\n",
              "      <td>0.985000</td>\n",
              "      <td>0.985583</td>\n",
              "      <td>0.000583</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>965.985611</td>\n",
              "      <td>7.931139</td>\n",
              "      <td>4.182384</td>\n",
              "      <td>0.005006</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 100, 'lr': 0.0001}</td>\n",
              "      <td>0.986033</td>\n",
              "      <td>0.983833</td>\n",
              "      <td>0.984933</td>\n",
              "      <td>0.001100</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>476.990983</td>\n",
              "      <td>2.746981</td>\n",
              "      <td>3.898018</td>\n",
              "      <td>0.047087</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 50, 'lr': 0.0001}</td>\n",
              "      <td>0.984800</td>\n",
              "      <td>0.983733</td>\n",
              "      <td>0.984267</td>\n",
              "      <td>0.000533</td>\n",
              "      <td>3</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>906.421972</td>\n",
              "      <td>0.956468</td>\n",
              "      <td>3.710845</td>\n",
              "      <td>0.290406</td>\n",
              "      <td>{'drop': 0, 'epochs': 100, 'lr': 0.0001}</td>\n",
              "      <td>0.985300</td>\n",
              "      <td>0.983133</td>\n",
              "      <td>0.984217</td>\n",
              "      <td>0.001083</td>\n",
              "      <td>4</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>947.038792</td>\n",
              "      <td>8.267578</td>\n",
              "      <td>4.110316</td>\n",
              "      <td>0.110192</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 100, 'lr': 0.0001}</td>\n",
              "      <td>0.984367</td>\n",
              "      <td>0.983700</td>\n",
              "      <td>0.984033</td>\n",
              "      <td>0.000333</td>\n",
              "      <td>5</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>468.193151</td>\n",
              "      <td>3.463599</td>\n",
              "      <td>3.772971</td>\n",
              "      <td>0.039663</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 50, 'lr': 0.0001}</td>\n",
              "      <td>0.985133</td>\n",
              "      <td>0.982900</td>\n",
              "      <td>0.984017</td>\n",
              "      <td>0.001117</td>\n",
              "      <td>6</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>990.465287</td>\n",
              "      <td>1.032297</td>\n",
              "      <td>3.556588</td>\n",
              "      <td>0.775545</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 100, 'lr': 0.0001}</td>\n",
              "      <td>0.985133</td>\n",
              "      <td>0.982867</td>\n",
              "      <td>0.984000</td>\n",
              "      <td>0.001133</td>\n",
              "      <td>7</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>235.867579</td>\n",
              "      <td>0.215019</td>\n",
              "      <td>3.538659</td>\n",
              "      <td>0.029385</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 25, 'lr': 0.0001}</td>\n",
              "      <td>0.985133</td>\n",
              "      <td>0.982000</td>\n",
              "      <td>0.983567</td>\n",
              "      <td>0.001567</td>\n",
              "      <td>8</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>458.467608</td>\n",
              "      <td>0.020042</td>\n",
              "      <td>3.766435</td>\n",
              "      <td>0.019322</td>\n",
              "      <td>{'drop': 0, 'epochs': 50, 'lr': 0.0001}</td>\n",
              "      <td>0.983500</td>\n",
              "      <td>0.982367</td>\n",
              "      <td>0.982933</td>\n",
              "      <td>0.000567</td>\n",
              "      <td>9</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>230.969342</td>\n",
              "      <td>1.304813</td>\n",
              "      <td>3.373224</td>\n",
              "      <td>0.107928</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 25, 'lr': 0.0001}</td>\n",
              "      <td>0.983000</td>\n",
              "      <td>0.981433</td>\n",
              "      <td>0.982217</td>\n",
              "      <td>0.000783</td>\n",
              "      <td>10</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>231.190272</td>\n",
              "      <td>0.508442</td>\n",
              "      <td>3.632632</td>\n",
              "      <td>0.029831</td>\n",
              "      <td>{'drop': 0, 'epochs': 25, 'lr': 0.0001}</td>\n",
              "      <td>0.982467</td>\n",
              "      <td>0.980667</td>\n",
              "      <td>0.981567</td>\n",
              "      <td>0.000900</td>\n",
              "      <td>11</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>239.703217</td>\n",
              "      <td>1.666640</td>\n",
              "      <td>3.635504</td>\n",
              "      <td>0.036287</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 25, 'lr': 0.0001}</td>\n",
              "      <td>0.980200</td>\n",
              "      <td>0.982133</td>\n",
              "      <td>0.981167</td>\n",
              "      <td>0.000967</td>\n",
              "      <td>12</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>472.669152</td>\n",
              "      <td>0.692647</td>\n",
              "      <td>3.802309</td>\n",
              "      <td>0.003780</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 50, 'lr': 0.001}</td>\n",
              "      <td>0.980867</td>\n",
              "      <td>0.979733</td>\n",
              "      <td>0.980300</td>\n",
              "      <td>0.000567</td>\n",
              "      <td>13</td>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>461.567727</td>\n",
              "      <td>0.167944</td>\n",
              "      <td>3.668837</td>\n",
              "      <td>0.079127</td>\n",
              "      <td>{'drop': 0, 'epochs': 50, 'lr': 0.001}</td>\n",
              "      <td>0.976400</td>\n",
              "      <td>0.979300</td>\n",
              "      <td>0.977850</td>\n",
              "      <td>0.001450</td>\n",
              "      <td>14</td>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>467.453461</td>\n",
              "      <td>2.973203</td>\n",
              "      <td>3.786648</td>\n",
              "      <td>0.032438</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 50, 'lr': 0.001}</td>\n",
              "      <td>0.976900</td>\n",
              "      <td>0.978367</td>\n",
              "      <td>0.977633</td>\n",
              "      <td>0.000733</td>\n",
              "      <td>15</td>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>233.098912</td>\n",
              "      <td>1.971407</td>\n",
              "      <td>3.358500</td>\n",
              "      <td>0.076281</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 25, 'lr': 0.001}</td>\n",
              "      <td>0.980133</td>\n",
              "      <td>0.975067</td>\n",
              "      <td>0.977600</td>\n",
              "      <td>0.002533</td>\n",
              "      <td>16</td>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>232.572357</td>\n",
              "      <td>1.479185</td>\n",
              "      <td>3.502814</td>\n",
              "      <td>0.009773</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 25, 'lr': 0.001}</td>\n",
              "      <td>0.976400</td>\n",
              "      <td>0.977267</td>\n",
              "      <td>0.976833</td>\n",
              "      <td>0.000433</td>\n",
              "      <td>17</td>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>241.638597</td>\n",
              "      <td>1.631259</td>\n",
              "      <td>3.554324</td>\n",
              "      <td>0.087975</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 25, 'lr': 0.001}</td>\n",
              "      <td>0.974267</td>\n",
              "      <td>0.978767</td>\n",
              "      <td>0.976517</td>\n",
              "      <td>0.002250</td>\n",
              "      <td>18</td>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>227.033984</td>\n",
              "      <td>0.070495</td>\n",
              "      <td>3.628998</td>\n",
              "      <td>0.016806</td>\n",
              "      <td>{'drop': 0, 'epochs': 25, 'lr': 0.001}</td>\n",
              "      <td>0.978233</td>\n",
              "      <td>0.974633</td>\n",
              "      <td>0.976433</td>\n",
              "      <td>0.001800</td>\n",
              "      <td>19</td>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>477.804809</td>\n",
              "      <td>2.823703</td>\n",
              "      <td>3.788849</td>\n",
              "      <td>0.098861</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 50, 'lr': 0.001}</td>\n",
              "      <td>0.978367</td>\n",
              "      <td>0.974300</td>\n",
              "      <td>0.976333</td>\n",
              "      <td>0.002033</td>\n",
              "      <td>20</td>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>955.508967</td>\n",
              "      <td>2.564818</td>\n",
              "      <td>4.041946</td>\n",
              "      <td>0.063852</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 100, 'lr': 0.001}</td>\n",
              "      <td>0.973800</td>\n",
              "      <td>0.978000</td>\n",
              "      <td>0.975900</td>\n",
              "      <td>0.002100</td>\n",
              "      <td>21</td>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>905.660821</td>\n",
              "      <td>0.003485</td>\n",
              "      <td>3.856573</td>\n",
              "      <td>0.044946</td>\n",
              "      <td>{'drop': 0, 'epochs': 100, 'lr': 0.001}</td>\n",
              "      <td>0.970667</td>\n",
              "      <td>0.973567</td>\n",
              "      <td>0.972117</td>\n",
              "      <td>0.001450</td>\n",
              "      <td>22</td>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>964.259935</td>\n",
              "      <td>7.863200</td>\n",
              "      <td>3.994214</td>\n",
              "      <td>0.103397</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 100, 'lr': 0.001}</td>\n",
              "      <td>0.975133</td>\n",
              "      <td>0.968500</td>\n",
              "      <td>0.971817</td>\n",
              "      <td>0.003317</td>\n",
              "      <td>23</td>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>940.852324</td>\n",
              "      <td>7.213572</td>\n",
              "      <td>3.982933</td>\n",
              "      <td>0.024368</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 100, 'lr': 0.001}</td>\n",
              "      <td>0.973767</td>\n",
              "      <td>0.953900</td>\n",
              "      <td>0.963833</td>\n",
              "      <td>0.009933</td>\n",
              "      <td>24</td>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>916.123076</td>\n",
              "      <td>0.048734</td>\n",
              "      <td>3.883017</td>\n",
              "      <td>0.126879</td>\n",
              "      <td>{'drop': 0, 'epochs': 100, 'lr': 0.01}</td>\n",
              "      <td>0.102433</td>\n",
              "      <td>0.105267</td>\n",
              "      <td>0.103850</td>\n",
              "      <td>0.001417</td>\n",
              "      <td>25</td>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>236.467401</td>\n",
              "      <td>0.400897</td>\n",
              "      <td>3.909430</td>\n",
              "      <td>0.114049</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 25, 'lr': 0.01}</td>\n",
              "      <td>0.103567</td>\n",
              "      <td>0.101933</td>\n",
              "      <td>0.102750</td>\n",
              "      <td>0.000817</td>\n",
              "      <td>26</td>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>234.508416</td>\n",
              "      <td>0.465258</td>\n",
              "      <td>3.431666</td>\n",
              "      <td>0.036131</td>\n",
              "      <td>{'drop': 0, 'epochs': 25, 'lr': 0.01}</td>\n",
              "      <td>0.114100</td>\n",
              "      <td>0.090400</td>\n",
              "      <td>0.102250</td>\n",
              "      <td>0.011850</td>\n",
              "      <td>27</td>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>468.694160</td>\n",
              "      <td>0.174074</td>\n",
              "      <td>3.695244</td>\n",
              "      <td>0.039882</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 50, 'lr': 0.01}</td>\n",
              "      <td>0.098267</td>\n",
              "      <td>0.105267</td>\n",
              "      <td>0.101767</td>\n",
              "      <td>0.003500</td>\n",
              "      <td>28</td>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>238.271776</td>\n",
              "      <td>1.035579</td>\n",
              "      <td>4.134261</td>\n",
              "      <td>0.045043</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 25, 'lr': 0.01}</td>\n",
              "      <td>0.098267</td>\n",
              "      <td>0.105267</td>\n",
              "      <td>0.101767</td>\n",
              "      <td>0.003500</td>\n",
              "      <td>28</td>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>936.712703</td>\n",
              "      <td>4.843200</td>\n",
              "      <td>3.928072</td>\n",
              "      <td>0.110751</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 100, 'lr': 0.01}</td>\n",
              "      <td>0.100100</td>\n",
              "      <td>0.100333</td>\n",
              "      <td>0.100217</td>\n",
              "      <td>0.000117</td>\n",
              "      <td>30</td>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>461.404979</td>\n",
              "      <td>2.997088</td>\n",
              "      <td>3.651751</td>\n",
              "      <td>0.018660</td>\n",
              "      <td>{'drop': 0.2, 'epochs': 50, 'lr': 0.01}</td>\n",
              "      <td>0.098267</td>\n",
              "      <td>0.101933</td>\n",
              "      <td>0.100100</td>\n",
              "      <td>0.001833</td>\n",
              "      <td>31</td>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>967.015692</td>\n",
              "      <td>9.234196</td>\n",
              "      <td>3.936852</td>\n",
              "      <td>0.121639</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 100, 'lr': 0.01}</td>\n",
              "      <td>0.097533</td>\n",
              "      <td>0.098733</td>\n",
              "      <td>0.098133</td>\n",
              "      <td>0.000600</td>\n",
              "      <td>32</td>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>246.251275</td>\n",
              "      <td>1.888880</td>\n",
              "      <td>4.085655</td>\n",
              "      <td>0.080869</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 25, 'lr': 0.01}</td>\n",
              "      <td>0.098267</td>\n",
              "      <td>0.097767</td>\n",
              "      <td>0.098017</td>\n",
              "      <td>0.000250</td>\n",
              "      <td>33</td>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>459.089557</td>\n",
              "      <td>0.075109</td>\n",
              "      <td>3.721984</td>\n",
              "      <td>0.058817</td>\n",
              "      <td>{'drop': 0, 'epochs': 50, 'lr': 0.01}</td>\n",
              "      <td>0.095833</td>\n",
              "      <td>0.099200</td>\n",
              "      <td>0.097517</td>\n",
              "      <td>0.001683</td>\n",
              "      <td>34</td>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>944.027379</td>\n",
              "      <td>2.006245</td>\n",
              "      <td>3.923541</td>\n",
              "      <td>0.113876</td>\n",
              "      <td>{'drop': 0.5, 'epochs': 100, 'lr': 0.01}</td>\n",
              "      <td>0.090300</td>\n",
              "      <td>0.101933</td>\n",
              "      <td>0.096117</td>\n",
              "      <td>0.005817</td>\n",
              "      <td>35</td>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>476.196026</td>\n",
              "      <td>3.471605</td>\n",
              "      <td>3.618748</td>\n",
              "      <td>0.061767</td>\n",
              "      <td>{'drop': 0.3, 'epochs': 50, 'lr': 0.01}</td>\n",
              "      <td>0.098700</td>\n",
              "      <td>0.090400</td>\n",
              "      <td>0.094550</td>\n",
              "      <td>0.004150</td>\n",
              "      <td>36</td>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    mean_fit_time  std_fit_time  mean_score_time  ...      lr epochs  drop\n",
              "32     471.794007      0.708622         3.902438  ...  0.0001     50   0.5\n",
              "26     965.985611      7.931139         4.182384  ...  0.0001    100   0.3\n",
              "23     476.990983      2.746981         3.898018  ...  0.0001     50   0.3\n",
              "8      906.421972      0.956468         3.710845  ...  0.0001    100   0.0\n",
              "17     947.038792      8.267578         4.110316  ...  0.0001    100   0.2\n",
              "14     468.193151      3.463599         3.772971  ...  0.0001     50   0.2\n",
              "35     990.465287      1.032297         3.556588  ...  0.0001    100   0.5\n",
              "29     235.867579      0.215019         3.538659  ...  0.0001     25   0.5\n",
              "5      458.467608      0.020042         3.766435  ...  0.0001     50   0.0\n",
              "11     230.969342      1.304813         3.373224  ...  0.0001     25   0.2\n",
              "2      231.190272      0.508442         3.632632  ...  0.0001     25   0.0\n",
              "20     239.703217      1.666640         3.635504  ...  0.0001     25   0.3\n",
              "31     472.669152      0.692647         3.802309  ...   0.001     50   0.5\n",
              "4      461.567727      0.167944         3.668837  ...   0.001     50   0.0\n",
              "13     467.453461      2.973203         3.786648  ...   0.001     50   0.2\n",
              "10     233.098912      1.971407         3.358500  ...   0.001     25   0.2\n",
              "28     232.572357      1.479185         3.502814  ...   0.001     25   0.5\n",
              "19     241.638597      1.631259         3.554324  ...   0.001     25   0.3\n",
              "1      227.033984      0.070495         3.628998  ...   0.001     25   0.0\n",
              "22     477.804809      2.823703         3.788849  ...   0.001     50   0.3\n",
              "34     955.508967      2.564818         4.041946  ...   0.001    100   0.5\n",
              "7      905.660821      0.003485         3.856573  ...   0.001    100   0.0\n",
              "25     964.259935      7.863200         3.994214  ...   0.001    100   0.3\n",
              "16     940.852324      7.213572         3.982933  ...   0.001    100   0.2\n",
              "6      916.123076      0.048734         3.883017  ...    0.01    100   0.0\n",
              "9      236.467401      0.400897         3.909430  ...    0.01     25   0.2\n",
              "0      234.508416      0.465258         3.431666  ...    0.01     25   0.0\n",
              "30     468.694160      0.174074         3.695244  ...    0.01     50   0.5\n",
              "18     238.271776      1.035579         4.134261  ...    0.01     25   0.3\n",
              "15     936.712703      4.843200         3.928072  ...    0.01    100   0.2\n",
              "12     461.404979      2.997088         3.651751  ...    0.01     50   0.2\n",
              "24     967.015692      9.234196         3.936852  ...    0.01    100   0.3\n",
              "27     246.251275      1.888880         4.085655  ...    0.01     25   0.5\n",
              "3      459.089557      0.075109         3.721984  ...    0.01     50   0.0\n",
              "33     944.027379      2.006245         3.923541  ...    0.01    100   0.5\n",
              "21     476.196026      3.471605         3.618748  ...    0.01     50   0.3\n",
              "\n",
              "[36 rows x 13 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9DbJsYMvmPFH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 867
        },
        "outputId": "af240745-3b15-4e1c-e174-1445ccca63b2"
      },
      "source": [
        "score = pd.DataFrame(grid_results.cv_results_).sort_values(by='mean_test_score', ascending = False)\n",
        "parameters = grid_results.param_grid\n",
        "for i in parameters.keys():\n",
        "    print(i, len(parameters[i]), parameters[i])\n",
        "score[i] = score.params.apply(lambda x: x[i])\n",
        "l =['mean_test_score'] + list(parameters.keys())\n",
        "for i in list(parameters.keys()):\n",
        "    sns.swarmplot(data = score[l], x = i, y = 'mean_test_score')\n",
        "    plt.savefig(i+'.jpg', dpi = 100)\n",
        "    plt.show()"
      ],
      "execution_count": 154,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "lr 3 [0.01, 0.001, 0.0001]\n",
            "epochs 3 [25, 50, 100]\n",
            "drop 4 [0, 0.2, 0.3, 0.5]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUiiIiEjo/wNdANGU4c4s8gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbtIQWyUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBBQKIiISUCiIiEhAoSAiIoH/A4VEyj7z0xpbAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iAekTZUOtt2h",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "score[\"params\"] = score[\"params\"].map(lambda x: str(x))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rczPqLuSsWLb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 632
        },
        "outputId": "6fe180d6-88d0-406b-f08d-9a783e9d25ec"
      },
      "source": [
        "plt.figure(figsize=(15,7))\n",
        "ax = sns.barplot(x=\"params\",y=\"mean_test_score\",data=score)\n",
        "plt.xticks(rotation=90)\n",
        "plt.savefig('barplot.png',dpi=100,bbox_inches='tight')"
      ],
      "execution_count": 171,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3wAAAJnCAYAAADMclHOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4p99AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1080x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yw_cKcTYsrJs",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "d6686d9f-9cee-4530-83df-afae255bad12"
      },
      "source": [
        "score[[\"lr\",\"epochs\",\"drop\",\"mean_test_score\"]]"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>lr</th>\n",
              "      <th>epochs</th>\n",
              "      <th>drop</th>\n",
              "      <th>mean_test_score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.985583</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.984933</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.984267</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.984217</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.984033</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.984017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.984000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.983567</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.982933</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.982217</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.981567</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>0.0001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.981167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.980300</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.977850</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.977633</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.977600</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.976833</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.976517</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.001</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.976433</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>0.001</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.976333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.975900</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.972117</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.971817</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>0.001</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.963833</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.103850</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.102750</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.102250</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.101767</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.101767</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.100217</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.2</td>\n",
              "      <td>0.100100</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.098133</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>0.01</td>\n",
              "      <td>25</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.098017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.097517</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>0.01</td>\n",
              "      <td>100</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.096117</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>0.01</td>\n",
              "      <td>50</td>\n",
              "      <td>0.3</td>\n",
              "      <td>0.094550</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        lr epochs  drop  mean_test_score\n",
              "32  0.0001     50   0.5         0.985583\n",
              "26  0.0001    100   0.3         0.984933\n",
              "23  0.0001     50   0.3         0.984267\n",
              "8   0.0001    100   0.0         0.984217\n",
              "17  0.0001    100   0.2         0.984033\n",
              "14  0.0001     50   0.2         0.984017\n",
              "35  0.0001    100   0.5         0.984000\n",
              "29  0.0001     25   0.5         0.983567\n",
              "5   0.0001     50   0.0         0.982933\n",
              "11  0.0001     25   0.2         0.982217\n",
              "2   0.0001     25   0.0         0.981567\n",
              "20  0.0001     25   0.3         0.981167\n",
              "31   0.001     50   0.5         0.980300\n",
              "4    0.001     50   0.0         0.977850\n",
              "13   0.001     50   0.2         0.977633\n",
              "10   0.001     25   0.2         0.977600\n",
              "28   0.001     25   0.5         0.976833\n",
              "19   0.001     25   0.3         0.976517\n",
              "1    0.001     25   0.0         0.976433\n",
              "22   0.001     50   0.3         0.976333\n",
              "34   0.001    100   0.5         0.975900\n",
              "7    0.001    100   0.0         0.972117\n",
              "25   0.001    100   0.3         0.971817\n",
              "16   0.001    100   0.2         0.963833\n",
              "6     0.01    100   0.0         0.103850\n",
              "9     0.01     25   0.2         0.102750\n",
              "0     0.01     25   0.0         0.102250\n",
              "30    0.01     50   0.5         0.101767\n",
              "18    0.01     25   0.3         0.101767\n",
              "15    0.01    100   0.2         0.100217\n",
              "12    0.01     50   0.2         0.100100\n",
              "24    0.01    100   0.3         0.098133\n",
              "27    0.01     25   0.5         0.098017\n",
              "3     0.01     50   0.0         0.097517\n",
              "33    0.01    100   0.5         0.096117\n",
              "21    0.01     50   0.3         0.094550"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 95
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k7ie0Gaqt9FE",
        "colab_type": "text"
      },
      "source": [
        "## Tune Activation Function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WRYmDp6euBNT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#parameters\n",
        "lr= [0.0001]\n",
        "epochs=[50]\n",
        "drop=[0.5]\n",
        "act = [None,\"tanh\",\"sigmoid\",\"relu\"]\n",
        "param_grid = dict(lr=lr,epochs=epochs,drop=drop,act=act)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cbdyYgeXuXJA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "5e6d0923-508e-42a4-bcb0-114b80d8518d"
      },
      "source": [
        "grid2 = GridSearchCV(estimator=model, param_grid=param_grid, n_jobs=-1, cv=2, verbose=True)  #GridSearch with 1-fold Cross Validation\n",
        "grid_result2 = grid2.fit(X, y)\n",
        "# summarize results\n",
        "print(\"Best: %f using %s\" % (grid_result2.best_score_, grid_result2.best_params_))\n",
        "means = grid_result2.cv_results_['mean_test_score']\n",
        "stds = grid_result2.cv_results_['std_test_score']\n",
        "params = grid_result2.cv_results_['params']\n",
        "for mean, stdev, param in zip(means, stds, params):\n",
        "    print(\"%f (%f) with: %r\" % (mean, stdev, param))"
      ],
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 2 folds for each of 4 candidates, totalling 8 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done   8 out of   8 | elapsed: 33.8min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "60000/60000 [==============================] - 17s 275us/step - loss: 0.6916 - acc: 0.7877\n",
            "Epoch 2/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.2459 - acc: 0.9280\n",
            "Epoch 3/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.1762 - acc: 0.9498\n",
            "Epoch 4/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.1408 - acc: 0.9599\n",
            "Epoch 5/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.1195 - acc: 0.9661\n",
            "Epoch 6/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.1063 - acc: 0.9698\n",
            "Epoch 7/50\n",
            "60000/60000 [==============================] - 15s 258us/step - loss: 0.0942 - acc: 0.9730\n",
            "Epoch 8/50\n",
            "60000/60000 [==============================] - 15s 257us/step - loss: 0.0844 - acc: 0.9759\n",
            "Epoch 9/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0766 - acc: 0.9783\n",
            "Epoch 10/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0692 - acc: 0.9798\n",
            "Epoch 11/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0642 - acc: 0.9812\n",
            "Epoch 12/50\n",
            "60000/60000 [==============================] - 15s 257us/step - loss: 0.0608 - acc: 0.9824\n",
            "Epoch 13/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0572 - acc: 0.9831\n",
            "Epoch 14/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0522 - acc: 0.9852\n",
            "Epoch 15/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0510 - acc: 0.9852\n",
            "Epoch 16/50\n",
            "60000/60000 [==============================] - 16s 260us/step - loss: 0.0488 - acc: 0.9859\n",
            "Epoch 17/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0456 - acc: 0.9867\n",
            "Epoch 18/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0442 - acc: 0.9868\n",
            "Epoch 19/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0391 - acc: 0.9889\n",
            "Epoch 20/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0406 - acc: 0.9884\n",
            "Epoch 21/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0370 - acc: 0.9890\n",
            "Epoch 22/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0356 - acc: 0.9894\n",
            "Epoch 23/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0334 - acc: 0.9899\n",
            "Epoch 24/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0320 - acc: 0.9903\n",
            "Epoch 25/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0314 - acc: 0.9907\n",
            "Epoch 26/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0289 - acc: 0.9910\n",
            "Epoch 27/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0281 - acc: 0.9914\n",
            "Epoch 28/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0271 - acc: 0.9921\n",
            "Epoch 29/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0262 - acc: 0.9920\n",
            "Epoch 30/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0244 - acc: 0.9925\n",
            "Epoch 31/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0234 - acc: 0.9929\n",
            "Epoch 32/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0232 - acc: 0.9928\n",
            "Epoch 33/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0218 - acc: 0.9934\n",
            "Epoch 34/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0208 - acc: 0.9936\n",
            "Epoch 35/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0202 - acc: 0.9937\n",
            "Epoch 36/50\n",
            "60000/60000 [==============================] - 16s 261us/step - loss: 0.0210 - acc: 0.9935\n",
            "Epoch 37/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0188 - acc: 0.9942\n",
            "Epoch 38/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0176 - acc: 0.9946\n",
            "Epoch 39/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0179 - acc: 0.9945\n",
            "Epoch 40/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0172 - acc: 0.9946\n",
            "Epoch 41/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0171 - acc: 0.9945\n",
            "Epoch 42/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0160 - acc: 0.9949\n",
            "Epoch 43/50\n",
            "60000/60000 [==============================] - 15s 253us/step - loss: 0.0154 - acc: 0.9955\n",
            "Epoch 44/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0151 - acc: 0.9952\n",
            "Epoch 45/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0142 - acc: 0.9954\n",
            "Epoch 46/50\n",
            "60000/60000 [==============================] - 15s 255us/step - loss: 0.0142 - acc: 0.9955\n",
            "Epoch 47/50\n",
            "60000/60000 [==============================] - 15s 258us/step - loss: 0.0143 - acc: 0.9956\n",
            "Epoch 48/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0128 - acc: 0.9961\n",
            "Epoch 49/50\n",
            "60000/60000 [==============================] - 15s 254us/step - loss: 0.0131 - acc: 0.9956\n",
            "Epoch 50/50\n",
            "60000/60000 [==============================] - 15s 256us/step - loss: 0.0114 - acc: 0.9965\n",
            "Best: 0.989550 using {'act': 'relu', 'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.984033 (0.000433) with: {'act': None, 'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.988717 (0.000417) with: {'act': 'tanh', 'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.978800 (0.000933) with: {'act': 'sigmoid', 'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n",
            "0.989550 (0.000017) with: {'act': 'relu', 'drop': 0.5, 'epochs': 50, 'lr': 0.0001}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SAXwHNrSzRlJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pickle.dump(grid_result2,open('actgrid.pkl','wb'))\n",
        "!cp actgrid.pkl drive/My\\ Drive"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H2jO82cP3-RJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scores = pd.DataFrame(grid_result2.cv_results_).sort_values(by='mean_test_score', ascending = False)[[\"param_act\",\"mean_test_score\"]]\n",
        "scores[\"param_act\"] = scores[\"param_act\"].map(lambda x: str(x) if x else \"Linear\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g9fPP1Gf5gT9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "outputId": "37480bd3-7445-4755-fdc9-278992969885"
      },
      "source": [
        "sns.barplot(x=\"param_act\",y=\"mean_test_score\",data=scores)\n",
        "plt.ylim(0.97, 1.0)"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.97, 1.0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 115
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAELCAYAAADz6wBxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zeQKWKFdvKK0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "29aa4012-8c88-4c06-f3d9-52f3833dc33b"
      },
      "source": [
        "lr=0.0001 #learning rate for adam optimizer\n",
        "epoch=50 #number of epochs to train\n",
        "drop=0.5 #dropout\n",
        "for act in [None,\"relu\",\"sigmoid\",\"tanh\"]:\n",
        "  model = KerasClassifier(build_fn=create_model, epochs=epoch, act=act, lr=lr,drop=drop)\n",
        "  history = model.fit(trainX,trainY, verbose=True, validation_data=(valX,valY))\n",
        "  \n",
        "  #Print Learning Curves\n",
        "  plt.figure(figsize=(8,8))\n",
        "  plt.plot(history.history['acc'], label='accuracy')\n",
        "  plt.plot(history.history['val_acc'], label = 'val_accuracy')\n",
        "  plt.xlabel('Epoch')\n",
        "  plt.ylabel('Accuracy')\n",
        "  plt.legend(loc='lower right')\n",
        "  plt.show()\n",
        "  \n",
        "  \n",
        "  #Print Learning Curves\n",
        "  plt.figure(figsize=(8,8))\n",
        "  plt.plot(history.history['loss'], label='loss')\n",
        "  plt.plot(history.history['val_loss'], label = 'val_loss')\n",
        "  plt.xlabel('Epoch')\n",
        "  plt.ylabel('Loss')\n",
        "  plt.legend(loc='upper right')\n",
        "  plt.show()\n",
        "  "
      ],
      "execution_count": 127,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/50\n",
            "48000/48000 [==============================] - 15s 314us/step - loss: 0.5753 - acc: 0.8307 - val_loss: 0.2041 - val_acc: 0.9380\n",
            "Epoch 2/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.2147 - acc: 0.9367 - val_loss: 0.1241 - val_acc: 0.9641\n",
            "Epoch 3/50\n",
            "48000/48000 [==============================] - 13s 279us/step - loss: 0.1547 - acc: 0.9548 - val_loss: 0.0912 - val_acc: 0.9722\n",
            "Epoch 4/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.1270 - acc: 0.9629 - val_loss: 0.0796 - val_acc: 0.9750\n",
            "Epoch 5/50\n",
            "48000/48000 [==============================] - 13s 279us/step - loss: 0.1100 - acc: 0.9676 - val_loss: 0.0693 - val_acc: 0.9780\n",
            "Epoch 6/50\n",
            "48000/48000 [==============================] - 13s 279us/step - loss: 0.0963 - acc: 0.9711 - val_loss: 0.0685 - val_acc: 0.9791\n",
            "Epoch 7/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0903 - acc: 0.9731 - val_loss: 0.0632 - val_acc: 0.9806\n",
            "Epoch 8/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0827 - acc: 0.9751 - val_loss: 0.0603 - val_acc: 0.9823\n",
            "Epoch 9/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0780 - acc: 0.9760 - val_loss: 0.0568 - val_acc: 0.9823\n",
            "Epoch 10/50\n",
            "48000/48000 [==============================] - 14s 281us/step - loss: 0.0732 - acc: 0.9784 - val_loss: 0.0531 - val_acc: 0.9832\n",
            "Epoch 11/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0694 - acc: 0.9790 - val_loss: 0.0544 - val_acc: 0.9830\n",
            "Epoch 12/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0645 - acc: 0.9802 - val_loss: 0.0523 - val_acc: 0.9845\n",
            "Epoch 13/50\n",
            "48000/48000 [==============================] - 13s 281us/step - loss: 0.0605 - acc: 0.9821 - val_loss: 0.0557 - val_acc: 0.9837\n",
            "Epoch 14/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0586 - acc: 0.9821 - val_loss: 0.0510 - val_acc: 0.9851\n",
            "Epoch 15/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0579 - acc: 0.9824 - val_loss: 0.0513 - val_acc: 0.9852\n",
            "Epoch 16/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0540 - acc: 0.9840 - val_loss: 0.0503 - val_acc: 0.9851\n",
            "Epoch 17/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0512 - acc: 0.9847 - val_loss: 0.0479 - val_acc: 0.9858\n",
            "Epoch 18/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0509 - acc: 0.9847 - val_loss: 0.0503 - val_acc: 0.9847\n",
            "Epoch 19/50\n",
            "48000/48000 [==============================] - 13s 275us/step - loss: 0.0471 - acc: 0.9855 - val_loss: 0.0489 - val_acc: 0.9858\n",
            "Epoch 20/50\n",
            "48000/48000 [==============================] - 13s 276us/step - loss: 0.0443 - acc: 0.9863 - val_loss: 0.0484 - val_acc: 0.9863\n",
            "Epoch 21/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0451 - acc: 0.9859 - val_loss: 0.0513 - val_acc: 0.9852\n",
            "Epoch 22/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0422 - acc: 0.9866 - val_loss: 0.0486 - val_acc: 0.9863\n",
            "Epoch 23/50\n",
            "48000/48000 [==============================] - 14s 281us/step - loss: 0.0407 - acc: 0.9864 - val_loss: 0.0475 - val_acc: 0.9861\n",
            "Epoch 24/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0407 - acc: 0.9874 - val_loss: 0.0487 - val_acc: 0.9855\n",
            "Epoch 25/50\n",
            "48000/48000 [==============================] - 13s 281us/step - loss: 0.0385 - acc: 0.9882 - val_loss: 0.0468 - val_acc: 0.9865\n",
            "Epoch 26/50\n",
            "48000/48000 [==============================] - 14s 282us/step - loss: 0.0380 - acc: 0.9876 - val_loss: 0.0474 - val_acc: 0.9861\n",
            "Epoch 27/50\n",
            "48000/48000 [==============================] - 14s 281us/step - loss: 0.0356 - acc: 0.9890 - val_loss: 0.0456 - val_acc: 0.9863\n",
            "Epoch 28/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0344 - acc: 0.9895 - val_loss: 0.0449 - val_acc: 0.9855\n",
            "Epoch 29/50\n",
            "48000/48000 [==============================] - 14s 281us/step - loss: 0.0344 - acc: 0.9888 - val_loss: 0.0484 - val_acc: 0.9857\n",
            "Epoch 30/50\n",
            "48000/48000 [==============================] - 13s 281us/step - loss: 0.0326 - acc: 0.9897 - val_loss: 0.0475 - val_acc: 0.9862\n",
            "Epoch 31/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0310 - acc: 0.9900 - val_loss: 0.0484 - val_acc: 0.9859\n",
            "Epoch 32/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0297 - acc: 0.9906 - val_loss: 0.0459 - val_acc: 0.9873\n",
            "Epoch 33/50\n",
            "48000/48000 [==============================] - 13s 281us/step - loss: 0.0287 - acc: 0.9908 - val_loss: 0.0495 - val_acc: 0.9863\n",
            "Epoch 34/50\n",
            "48000/48000 [==============================] - 13s 281us/step - loss: 0.0271 - acc: 0.9913 - val_loss: 0.0500 - val_acc: 0.9862\n",
            "Epoch 35/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0277 - acc: 0.9909 - val_loss: 0.0465 - val_acc: 0.9868\n",
            "Epoch 36/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0261 - acc: 0.9918 - val_loss: 0.0473 - val_acc: 0.9854\n",
            "Epoch 37/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0247 - acc: 0.9920 - val_loss: 0.0486 - val_acc: 0.9862\n",
            "Epoch 38/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0241 - acc: 0.9925 - val_loss: 0.0470 - val_acc: 0.9876\n",
            "Epoch 39/50\n",
            "48000/48000 [==============================] - 13s 279us/step - loss: 0.0221 - acc: 0.9931 - val_loss: 0.0490 - val_acc: 0.9860\n",
            "Epoch 40/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0226 - acc: 0.9926 - val_loss: 0.0487 - val_acc: 0.9869\n",
            "Epoch 41/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0213 - acc: 0.9933 - val_loss: 0.0482 - val_acc: 0.9862\n",
            "Epoch 42/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0214 - acc: 0.9935 - val_loss: 0.0523 - val_acc: 0.9860\n",
            "Epoch 43/50\n",
            "48000/48000 [==============================] - 13s 280us/step - loss: 0.0205 - acc: 0.9934 - val_loss: 0.0514 - val_acc: 0.9862\n",
            "Epoch 44/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0201 - acc: 0.9933 - val_loss: 0.0474 - val_acc: 0.9880\n",
            "Epoch 45/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0203 - acc: 0.9936 - val_loss: 0.0500 - val_acc: 0.9867\n",
            "Epoch 46/50\n",
            "48000/48000 [==============================] - 13s 277us/step - loss: 0.0190 - acc: 0.9933 - val_loss: 0.0509 - val_acc: 0.9868\n",
            "Epoch 47/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0170 - acc: 0.9941 - val_loss: 0.0529 - val_acc: 0.9863\n",
            "Epoch 48/50\n",
            "48000/48000 [==============================] - 13s 276us/step - loss: 0.0178 - acc: 0.9942 - val_loss: 0.0488 - val_acc: 0.9872\n",
            "Epoch 49/50\n",
            "48000/48000 [==============================] - 13s 279us/step - loss: 0.0169 - acc: 0.9942 - val_loss: 0.0523 - val_acc: 0.9856\n",
            "Epoch 50/50\n",
            "48000/48000 [==============================] - 13s 278us/step - loss: 0.0159 - acc: 0.9943 - val_loss: 0.0522 - val_acc: 0.9869\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAHjCAYAAAAkDrYkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5P4fMIYxjQPtd2oAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHjCAYAAAA6x4aXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngFHi/wLg/b9WF6fNkAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/50\n",
            "48000/48000 [==============================] - 15s 319us/step - loss: 0.7696 - acc: 0.7592 - val_loss: 0.1841 - val_acc: 0.9460\n",
            "Epoch 2/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.2801 - acc: 0.9182 - val_loss: 0.1209 - val_acc: 0.9643\n",
            "Epoch 3/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.2075 - acc: 0.9390 - val_loss: 0.0994 - val_acc: 0.9693\n",
            "Epoch 4/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.1682 - acc: 0.9517 - val_loss: 0.0874 - val_acc: 0.9742\n",
            "Epoch 5/50\n",
            "48000/48000 [==============================] - 14s 294us/step - loss: 0.1465 - acc: 0.9570 - val_loss: 0.0736 - val_acc: 0.9772\n",
            "Epoch 6/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1275 - acc: 0.9637 - val_loss: 0.0663 - val_acc: 0.9797\n",
            "Epoch 7/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.1140 - acc: 0.9672 - val_loss: 0.0634 - val_acc: 0.9798\n",
            "Epoch 8/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.1042 - acc: 0.9698 - val_loss: 0.0570 - val_acc: 0.9815\n",
            "Epoch 9/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0970 - acc: 0.9721 - val_loss: 0.0566 - val_acc: 0.9826\n",
            "Epoch 10/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0887 - acc: 0.9749 - val_loss: 0.0533 - val_acc: 0.9828\n",
            "Epoch 11/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0815 - acc: 0.9766 - val_loss: 0.0480 - val_acc: 0.9850\n",
            "Epoch 12/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0763 - acc: 0.9780 - val_loss: 0.0486 - val_acc: 0.9857\n",
            "Epoch 13/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0699 - acc: 0.9796 - val_loss: 0.0467 - val_acc: 0.9868\n",
            "Epoch 14/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0665 - acc: 0.9806 - val_loss: 0.0432 - val_acc: 0.9879\n",
            "Epoch 15/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0606 - acc: 0.9823 - val_loss: 0.0445 - val_acc: 0.9872\n",
            "Epoch 16/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0579 - acc: 0.9829 - val_loss: 0.0423 - val_acc: 0.9871\n",
            "Epoch 17/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0531 - acc: 0.9846 - val_loss: 0.0459 - val_acc: 0.9872\n",
            "Epoch 18/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0523 - acc: 0.9856 - val_loss: 0.0405 - val_acc: 0.9887\n",
            "Epoch 19/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0504 - acc: 0.9844 - val_loss: 0.0418 - val_acc: 0.9879\n",
            "Epoch 20/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0499 - acc: 0.9852 - val_loss: 0.0419 - val_acc: 0.9876\n",
            "Epoch 21/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0441 - acc: 0.9868 - val_loss: 0.0390 - val_acc: 0.9888\n",
            "Epoch 22/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0409 - acc: 0.9879 - val_loss: 0.0369 - val_acc: 0.9895\n",
            "Epoch 23/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0427 - acc: 0.9876 - val_loss: 0.0364 - val_acc: 0.9895\n",
            "Epoch 24/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0389 - acc: 0.9886 - val_loss: 0.0424 - val_acc: 0.9891\n",
            "Epoch 25/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0387 - acc: 0.9889 - val_loss: 0.0386 - val_acc: 0.9897\n",
            "Epoch 26/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0369 - acc: 0.9893 - val_loss: 0.0415 - val_acc: 0.9895\n",
            "Epoch 27/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0356 - acc: 0.9892 - val_loss: 0.0368 - val_acc: 0.9904\n",
            "Epoch 28/50\n",
            "48000/48000 [==============================] - 14s 295us/step - loss: 0.0339 - acc: 0.9895 - val_loss: 0.0371 - val_acc: 0.9908\n",
            "Epoch 29/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0329 - acc: 0.9898 - val_loss: 0.0386 - val_acc: 0.9902\n",
            "Epoch 30/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0306 - acc: 0.9901 - val_loss: 0.0424 - val_acc: 0.9900\n",
            "Epoch 31/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0318 - acc: 0.9905 - val_loss: 0.0364 - val_acc: 0.9908\n",
            "Epoch 32/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0275 - acc: 0.9919 - val_loss: 0.0405 - val_acc: 0.9898\n",
            "Epoch 33/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0276 - acc: 0.9913 - val_loss: 0.0382 - val_acc: 0.9904\n",
            "Epoch 34/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0252 - acc: 0.9919 - val_loss: 0.0402 - val_acc: 0.9902\n",
            "Epoch 35/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0258 - acc: 0.9917 - val_loss: 0.0387 - val_acc: 0.9905\n",
            "Epoch 36/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0234 - acc: 0.9930 - val_loss: 0.0408 - val_acc: 0.9902\n",
            "Epoch 37/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0247 - acc: 0.9927 - val_loss: 0.0399 - val_acc: 0.9906\n",
            "Epoch 38/50\n",
            "48000/48000 [==============================] - 14s 292us/step - loss: 0.0238 - acc: 0.9926 - val_loss: 0.0400 - val_acc: 0.9910\n",
            "Epoch 39/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0222 - acc: 0.9927 - val_loss: 0.0403 - val_acc: 0.9903\n",
            "Epoch 40/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0221 - acc: 0.9931 - val_loss: 0.0395 - val_acc: 0.9918\n",
            "Epoch 41/50\n",
            "48000/48000 [==============================] - 14s 292us/step - loss: 0.0208 - acc: 0.9932 - val_loss: 0.0381 - val_acc: 0.9914\n",
            "Epoch 42/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0219 - acc: 0.9930 - val_loss: 0.0412 - val_acc: 0.9912\n",
            "Epoch 43/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0204 - acc: 0.9938 - val_loss: 0.0380 - val_acc: 0.9910\n",
            "Epoch 44/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0189 - acc: 0.9943 - val_loss: 0.0389 - val_acc: 0.9916\n",
            "Epoch 45/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0203 - acc: 0.9938 - val_loss: 0.0380 - val_acc: 0.9918\n",
            "Epoch 46/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0182 - acc: 0.9941 - val_loss: 0.0383 - val_acc: 0.9917\n",
            "Epoch 47/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0169 - acc: 0.9943 - val_loss: 0.0415 - val_acc: 0.9919\n",
            "Epoch 48/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0179 - acc: 0.9945 - val_loss: 0.0408 - val_acc: 0.9910\n",
            "Epoch 49/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0180 - acc: 0.9941 - val_loss: 0.0388 - val_acc: 0.9912\n",
            "Epoch 50/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0160 - acc: 0.9947 - val_loss: 0.0413 - val_acc: 0.9910\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAHjCAYAAAAkDrYkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHjCAYAAAA6x4aXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/50\n",
            "48000/48000 [==============================] - 15s 322us/step - loss: 2.3186 - acc: 0.1054 - val_loss: 2.3018 - val_acc: 0.1122\n",
            "Epoch 2/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 2.3019 - acc: 0.1113 - val_loss: 2.3021 - val_acc: 0.1122\n",
            "Epoch 3/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 2.3015 - acc: 0.1118 - val_loss: 2.3008 - val_acc: 0.1122\n",
            "Epoch 4/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 1.6107 - acc: 0.4470 - val_loss: 0.8131 - val_acc: 0.8156\n",
            "Epoch 5/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.8149 - acc: 0.7651 - val_loss: 0.5449 - val_acc: 0.8681\n",
            "Epoch 6/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.6143 - acc: 0.8301 - val_loss: 0.4044 - val_acc: 0.8979\n",
            "Epoch 7/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.4979 - acc: 0.8632 - val_loss: 0.3197 - val_acc: 0.9183\n",
            "Epoch 8/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.4166 - acc: 0.8876 - val_loss: 0.2608 - val_acc: 0.9316\n",
            "Epoch 9/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.3570 - acc: 0.9035 - val_loss: 0.2209 - val_acc: 0.9403\n",
            "Epoch 10/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.3111 - acc: 0.9168 - val_loss: 0.1893 - val_acc: 0.9479\n",
            "Epoch 11/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.2818 - acc: 0.9245 - val_loss: 0.1706 - val_acc: 0.9519\n",
            "Epoch 12/50\n",
            "48000/48000 [==============================] - 14s 282us/step - loss: 0.2543 - acc: 0.9308 - val_loss: 0.1521 - val_acc: 0.9565\n",
            "Epoch 13/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.2313 - acc: 0.9384 - val_loss: 0.1394 - val_acc: 0.9594\n",
            "Epoch 14/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.2156 - acc: 0.9405 - val_loss: 0.1300 - val_acc: 0.9610\n",
            "Epoch 15/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.2007 - acc: 0.9456 - val_loss: 0.1207 - val_acc: 0.9642\n",
            "Epoch 16/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.1882 - acc: 0.9488 - val_loss: 0.1164 - val_acc: 0.9653\n",
            "Epoch 17/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1806 - acc: 0.9503 - val_loss: 0.1067 - val_acc: 0.9683\n",
            "Epoch 18/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1669 - acc: 0.9544 - val_loss: 0.1009 - val_acc: 0.9698\n",
            "Epoch 19/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1601 - acc: 0.9554 - val_loss: 0.0966 - val_acc: 0.9713\n",
            "Epoch 20/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1516 - acc: 0.9592 - val_loss: 0.0927 - val_acc: 0.9721\n",
            "Epoch 21/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1465 - acc: 0.9595 - val_loss: 0.0902 - val_acc: 0.9734\n",
            "Epoch 22/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.1409 - acc: 0.9615 - val_loss: 0.0863 - val_acc: 0.9742\n",
            "Epoch 23/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.1349 - acc: 0.9637 - val_loss: 0.0821 - val_acc: 0.9748\n",
            "Epoch 24/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.1320 - acc: 0.9643 - val_loss: 0.0855 - val_acc: 0.9738\n",
            "Epoch 25/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1262 - acc: 0.9664 - val_loss: 0.0778 - val_acc: 0.9756\n",
            "Epoch 26/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.1193 - acc: 0.9678 - val_loss: 0.0778 - val_acc: 0.9758\n",
            "Epoch 27/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.1181 - acc: 0.9669 - val_loss: 0.0736 - val_acc: 0.9772\n",
            "Epoch 28/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.1123 - acc: 0.9678 - val_loss: 0.0723 - val_acc: 0.9774\n",
            "Epoch 29/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.1127 - acc: 0.9688 - val_loss: 0.0719 - val_acc: 0.9775\n",
            "Epoch 30/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.1093 - acc: 0.9692 - val_loss: 0.0696 - val_acc: 0.9789\n",
            "Epoch 31/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.1060 - acc: 0.9711 - val_loss: 0.0674 - val_acc: 0.9796\n",
            "Epoch 32/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.1026 - acc: 0.9718 - val_loss: 0.0687 - val_acc: 0.9788\n",
            "Epoch 33/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0988 - acc: 0.9721 - val_loss: 0.0669 - val_acc: 0.9790\n",
            "Epoch 34/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0967 - acc: 0.9731 - val_loss: 0.0651 - val_acc: 0.9790\n",
            "Epoch 35/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0944 - acc: 0.9731 - val_loss: 0.0634 - val_acc: 0.9800\n",
            "Epoch 36/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0933 - acc: 0.9738 - val_loss: 0.0623 - val_acc: 0.9799\n",
            "Epoch 37/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0911 - acc: 0.9753 - val_loss: 0.0618 - val_acc: 0.9807\n",
            "Epoch 38/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0885 - acc: 0.9751 - val_loss: 0.0614 - val_acc: 0.9809\n",
            "Epoch 39/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0894 - acc: 0.9744 - val_loss: 0.0594 - val_acc: 0.9818\n",
            "Epoch 40/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0855 - acc: 0.9768 - val_loss: 0.0575 - val_acc: 0.9821\n",
            "Epoch 41/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0816 - acc: 0.9773 - val_loss: 0.0625 - val_acc: 0.9802\n",
            "Epoch 42/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0829 - acc: 0.9765 - val_loss: 0.0585 - val_acc: 0.9814\n",
            "Epoch 43/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0797 - acc: 0.9779 - val_loss: 0.0571 - val_acc: 0.9823\n",
            "Epoch 44/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0777 - acc: 0.9781 - val_loss: 0.0564 - val_acc: 0.9826\n",
            "Epoch 45/50\n",
            "48000/48000 [==============================] - 14s 292us/step - loss: 0.0770 - acc: 0.9776 - val_loss: 0.0550 - val_acc: 0.9833\n",
            "Epoch 46/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0759 - acc: 0.9791 - val_loss: 0.0563 - val_acc: 0.9820\n",
            "Epoch 47/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0736 - acc: 0.9786 - val_loss: 0.0542 - val_acc: 0.9835\n",
            "Epoch 48/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0732 - acc: 0.9793 - val_loss: 0.0533 - val_acc: 0.9832\n",
            "Epoch 49/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0724 - acc: 0.9796 - val_loss: 0.0529 - val_acc: 0.9834\n",
            "Epoch 50/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0703 - acc: 0.9804 - val_loss: 0.0538 - val_acc: 0.9827\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHjCAYAAAA6x4aXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQh0AgEmCUAcAYJIg1AEAmCT+f4yrNUO2alyPAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHjCAYAAAA6x4aXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKBP/D7KPJvJsAzgrAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/50\n",
            "48000/48000 [==============================] - 16s 324us/step - loss: 0.6029 - acc: 0.8369 - val_loss: 0.2008 - val_acc: 0.9479\n",
            "Epoch 2/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.2115 - acc: 0.9433 - val_loss: 0.1244 - val_acc: 0.9654\n",
            "Epoch 3/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.1449 - acc: 0.9601 - val_loss: 0.0915 - val_acc: 0.9738\n",
            "Epoch 4/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.1163 - acc: 0.9679 - val_loss: 0.0755 - val_acc: 0.9775\n",
            "Epoch 5/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0980 - acc: 0.9724 - val_loss: 0.0685 - val_acc: 0.9785\n",
            "Epoch 6/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0860 - acc: 0.9762 - val_loss: 0.0606 - val_acc: 0.9820\n",
            "Epoch 7/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0755 - acc: 0.9787 - val_loss: 0.0558 - val_acc: 0.9827\n",
            "Epoch 8/50\n",
            "48000/48000 [==============================] - 14s 288us/step - loss: 0.0684 - acc: 0.9811 - val_loss: 0.0520 - val_acc: 0.9842\n",
            "Epoch 9/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0621 - acc: 0.9822 - val_loss: 0.0474 - val_acc: 0.9857\n",
            "Epoch 10/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0590 - acc: 0.9831 - val_loss: 0.0453 - val_acc: 0.9855\n",
            "Epoch 11/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0520 - acc: 0.9857 - val_loss: 0.0450 - val_acc: 0.9861\n",
            "Epoch 12/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0499 - acc: 0.9855 - val_loss: 0.0438 - val_acc: 0.9861\n",
            "Epoch 13/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0460 - acc: 0.9864 - val_loss: 0.0440 - val_acc: 0.9866\n",
            "Epoch 14/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0421 - acc: 0.9876 - val_loss: 0.0431 - val_acc: 0.9877\n",
            "Epoch 15/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0412 - acc: 0.9884 - val_loss: 0.0400 - val_acc: 0.9882\n",
            "Epoch 16/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0371 - acc: 0.9894 - val_loss: 0.0399 - val_acc: 0.9880\n",
            "Epoch 17/50\n",
            "48000/48000 [==============================] - 14s 294us/step - loss: 0.0351 - acc: 0.9906 - val_loss: 0.0388 - val_acc: 0.9881\n",
            "Epoch 18/50\n",
            "48000/48000 [==============================] - 14s 297us/step - loss: 0.0340 - acc: 0.9899 - val_loss: 0.0394 - val_acc: 0.9878\n",
            "Epoch 19/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0305 - acc: 0.9908 - val_loss: 0.0392 - val_acc: 0.9886\n",
            "Epoch 20/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0291 - acc: 0.9916 - val_loss: 0.0383 - val_acc: 0.9890\n",
            "Epoch 21/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0272 - acc: 0.9920 - val_loss: 0.0372 - val_acc: 0.9888\n",
            "Epoch 22/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0262 - acc: 0.9921 - val_loss: 0.0375 - val_acc: 0.9887\n",
            "Epoch 23/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0241 - acc: 0.9928 - val_loss: 0.0372 - val_acc: 0.9888\n",
            "Epoch 24/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0227 - acc: 0.9934 - val_loss: 0.0368 - val_acc: 0.9895\n",
            "Epoch 25/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0223 - acc: 0.9938 - val_loss: 0.0367 - val_acc: 0.9899\n",
            "Epoch 26/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0211 - acc: 0.9934 - val_loss: 0.0395 - val_acc: 0.9887\n",
            "Epoch 27/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0191 - acc: 0.9941 - val_loss: 0.0361 - val_acc: 0.9894\n",
            "Epoch 28/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0185 - acc: 0.9946 - val_loss: 0.0366 - val_acc: 0.9897\n",
            "Epoch 29/50\n",
            "48000/48000 [==============================] - 14s 292us/step - loss: 0.0175 - acc: 0.9950 - val_loss: 0.0376 - val_acc: 0.9892\n",
            "Epoch 30/50\n",
            "48000/48000 [==============================] - 14s 294us/step - loss: 0.0169 - acc: 0.9952 - val_loss: 0.0391 - val_acc: 0.9891\n",
            "Epoch 31/50\n",
            "48000/48000 [==============================] - 14s 290us/step - loss: 0.0153 - acc: 0.9959 - val_loss: 0.0360 - val_acc: 0.9899\n",
            "Epoch 32/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0149 - acc: 0.9957 - val_loss: 0.0380 - val_acc: 0.9892\n",
            "Epoch 33/50\n",
            "48000/48000 [==============================] - 14s 291us/step - loss: 0.0145 - acc: 0.9955 - val_loss: 0.0363 - val_acc: 0.9904\n",
            "Epoch 34/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0125 - acc: 0.9965 - val_loss: 0.0347 - val_acc: 0.9910\n",
            "Epoch 35/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0128 - acc: 0.9961 - val_loss: 0.0369 - val_acc: 0.9902\n",
            "Epoch 36/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0112 - acc: 0.9969 - val_loss: 0.0353 - val_acc: 0.9908\n",
            "Epoch 37/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0107 - acc: 0.9969 - val_loss: 0.0358 - val_acc: 0.9908\n",
            "Epoch 38/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0108 - acc: 0.9971 - val_loss: 0.0355 - val_acc: 0.9910\n",
            "Epoch 39/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0099 - acc: 0.9971 - val_loss: 0.0345 - val_acc: 0.9908\n",
            "Epoch 40/50\n",
            "48000/48000 [==============================] - 14s 293us/step - loss: 0.0100 - acc: 0.9971 - val_loss: 0.0362 - val_acc: 0.9907\n",
            "Epoch 41/50\n",
            "48000/48000 [==============================] - 14s 289us/step - loss: 0.0091 - acc: 0.9975 - val_loss: 0.0355 - val_acc: 0.9910\n",
            "Epoch 42/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0092 - acc: 0.9974 - val_loss: 0.0370 - val_acc: 0.9906\n",
            "Epoch 43/50\n",
            "48000/48000 [==============================] - 14s 287us/step - loss: 0.0082 - acc: 0.9976 - val_loss: 0.0349 - val_acc: 0.9900\n",
            "Epoch 44/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0075 - acc: 0.9978 - val_loss: 0.0349 - val_acc: 0.9909\n",
            "Epoch 45/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0072 - acc: 0.9981 - val_loss: 0.0359 - val_acc: 0.9904\n",
            "Epoch 46/50\n",
            "48000/48000 [==============================] - 14s 286us/step - loss: 0.0075 - acc: 0.9979 - val_loss: 0.0360 - val_acc: 0.9910\n",
            "Epoch 47/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0065 - acc: 0.9983 - val_loss: 0.0350 - val_acc: 0.9907\n",
            "Epoch 48/50\n",
            "48000/48000 [==============================] - 14s 285us/step - loss: 0.0075 - acc: 0.9978 - val_loss: 0.0351 - val_acc: 0.9911\n",
            "Epoch 49/50\n",
            "48000/48000 [==============================] - 14s 283us/step - loss: 0.0061 - acc: 0.9981 - val_loss: 0.0388 - val_acc: 0.9908\n",
            "Epoch 50/50\n",
            "48000/48000 [==============================] - 14s 284us/step - loss: 0.0061 - acc: 0.9982 - val_loss: 0.0377 - val_acc: 0.9909\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAHjCAYAAAAkDrYkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmVmZc7I3MzMrc072ZmZmZe7/I2zR+/Gb4GcAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHjCAYAAAA6x4aXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MCS_DGieRuau",
        "colab_type": "text"
      },
      "source": [
        "# Final Model on Train+Validation and Best Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3DpuomluujC1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#best parameters\n",
        "act=\"relu\" #activation\n",
        "lr=0.0001 #learning rate for adam optimizer\n",
        "epoch=50 #number of epochs to train\n",
        "drop=0.5 #dropout"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z7NMVW4Quodf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = KerasClassifier(build_fn=create_model, epochs=epoch, act=act, lr=lr,drop=drop)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "c30416c6-d885-4e94-b84d-a1eb6e87199f",
        "id": "XxpD73qLRwa8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history = model.fit(X,y, verbose=True)"
      ],
      "execution_count": 130,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "60000/60000 [==============================] - 18s 293us/step - loss: 0.6463 - acc: 0.8026\n",
            "Epoch 2/50\n",
            "60000/60000 [==============================] - 16s 264us/step - loss: 0.2216 - acc: 0.9349\n",
            "Epoch 3/50\n",
            "60000/60000 [==============================] - 16s 270us/step - loss: 0.1612 - acc: 0.9532\n",
            "Epoch 4/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.1315 - acc: 0.9624\n",
            "Epoch 5/50\n",
            "60000/60000 [==============================] - 16s 266us/step - loss: 0.1118 - acc: 0.9683\n",
            "Epoch 6/50\n",
            "60000/60000 [==============================] - 16s 266us/step - loss: 0.0957 - acc: 0.9726\n",
            "Epoch 7/50\n",
            "60000/60000 [==============================] - 16s 266us/step - loss: 0.0846 - acc: 0.9763\n",
            "Epoch 8/50\n",
            "60000/60000 [==============================] - 16s 264us/step - loss: 0.0792 - acc: 0.9777\n",
            "Epoch 9/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0718 - acc: 0.9796\n",
            "Epoch 10/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0672 - acc: 0.9811\n",
            "Epoch 11/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0618 - acc: 0.9821\n",
            "Epoch 12/50\n",
            "60000/60000 [==============================] - 16s 267us/step - loss: 0.0581 - acc: 0.9834\n",
            "Epoch 13/50\n",
            "60000/60000 [==============================] - 16s 273us/step - loss: 0.0522 - acc: 0.9858\n",
            "Epoch 14/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0501 - acc: 0.9853\n",
            "Epoch 15/50\n",
            "60000/60000 [==============================] - 16s 272us/step - loss: 0.0483 - acc: 0.9861\n",
            "Epoch 16/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0451 - acc: 0.9867\n",
            "Epoch 17/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0434 - acc: 0.9878\n",
            "Epoch 18/50\n",
            "60000/60000 [==============================] - 16s 270us/step - loss: 0.0406 - acc: 0.9882\n",
            "Epoch 19/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0396 - acc: 0.9887\n",
            "Epoch 20/50\n",
            "60000/60000 [==============================] - 16s 270us/step - loss: 0.0368 - acc: 0.9892\n",
            "Epoch 21/50\n",
            "60000/60000 [==============================] - 16s 274us/step - loss: 0.0354 - acc: 0.9897\n",
            "Epoch 22/50\n",
            "60000/60000 [==============================] - 17s 276us/step - loss: 0.0329 - acc: 0.9902\n",
            "Epoch 23/50\n",
            "60000/60000 [==============================] - 16s 267us/step - loss: 0.0324 - acc: 0.9905\n",
            "Epoch 24/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0306 - acc: 0.9912\n",
            "Epoch 25/50\n",
            "60000/60000 [==============================] - 16s 266us/step - loss: 0.0280 - acc: 0.9916\n",
            "Epoch 26/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0267 - acc: 0.9922\n",
            "Epoch 27/50\n",
            "60000/60000 [==============================] - 16s 266us/step - loss: 0.0268 - acc: 0.9923\n",
            "Epoch 28/50\n",
            "60000/60000 [==============================] - 16s 265us/step - loss: 0.0242 - acc: 0.9922\n",
            "Epoch 29/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0258 - acc: 0.9924\n",
            "Epoch 30/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0237 - acc: 0.9923\n",
            "Epoch 31/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0225 - acc: 0.9932\n",
            "Epoch 32/50\n",
            "60000/60000 [==============================] - 16s 273us/step - loss: 0.0223 - acc: 0.9933\n",
            "Epoch 33/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0201 - acc: 0.9938\n",
            "Epoch 34/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0206 - acc: 0.9938\n",
            "Epoch 35/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0185 - acc: 0.9942\n",
            "Epoch 36/50\n",
            "60000/60000 [==============================] - 16s 267us/step - loss: 0.0189 - acc: 0.9944\n",
            "Epoch 37/50\n",
            "60000/60000 [==============================] - 16s 267us/step - loss: 0.0183 - acc: 0.9942\n",
            "Epoch 38/50\n",
            "60000/60000 [==============================] - 16s 268us/step - loss: 0.0178 - acc: 0.9944\n",
            "Epoch 39/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0177 - acc: 0.9944\n",
            "Epoch 40/50\n",
            "60000/60000 [==============================] - 16s 267us/step - loss: 0.0155 - acc: 0.9953\n",
            "Epoch 41/50\n",
            "60000/60000 [==============================] - 16s 272us/step - loss: 0.0166 - acc: 0.9945\n",
            "Epoch 42/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0157 - acc: 0.9950\n",
            "Epoch 43/50\n",
            "60000/60000 [==============================] - 16s 263us/step - loss: 0.0147 - acc: 0.9955\n",
            "Epoch 44/50\n",
            "60000/60000 [==============================] - 16s 263us/step - loss: 0.0141 - acc: 0.9955\n",
            "Epoch 45/50\n",
            "60000/60000 [==============================] - 16s 269us/step - loss: 0.0136 - acc: 0.9956\n",
            "Epoch 46/50\n",
            "60000/60000 [==============================] - 16s 271us/step - loss: 0.0134 - acc: 0.9955\n",
            "Epoch 47/50\n",
            "60000/60000 [==============================] - 17s 285us/step - loss: 0.0129 - acc: 0.9956\n",
            "Epoch 48/50\n",
            "60000/60000 [==============================] - 17s 283us/step - loss: 0.0119 - acc: 0.9962\n",
            "Epoch 49/50\n",
            "60000/60000 [==============================] - 17s 286us/step - loss: 0.0127 - acc: 0.9961\n",
            "Epoch 50/50\n",
            "60000/60000 [==============================] - 17s 285us/step - loss: 0.0117 - acc: 0.9962\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8eWsP0pQOcga",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "9e1e9056-173c-4fd2-dd49-882d6fce6a9b"
      },
      "source": [
        "#Print Learning Curves\n",
        "plt.plot(history.history['acc'], label='accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend(loc='lower right')\n",
        "plt.show()"
      ],
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEKCAYAAAA4t9PUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZjYsDiRmZjYsDiRmZjYsDiRmZjYs/xcKDuVlcOoghAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "apKL0_VdOCom",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "aae0364b-2e1e-4e80-fd01-0aca54454dbd"
      },
      "source": [
        "#Print Learning Curves\n",
        "plt.plot(history.history['loss'], label='loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend(loc='upper right')\n",
        "plt.show()"
      ],
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEKCAYAAAD9xUlFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PNc9S3akOsoy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pred = pd.DataFrame(model.predict_proba(testX)).astype('int32')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V9xPZhj1Py2G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pred.to_csv('mnist.csv',index=False, header=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eRUf7P0yQB66",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp mnist.csv drive/My\\ Drive"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Pn-Wbv5heAg3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "files.download('mnist.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A1ebGHswPXM3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "preds = model.predict(testX)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UaFP_A1rInim",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        },
        "outputId": "5ed1aae0-5782-456c-b180-e453d5de8017"
      },
      "source": [
        "def plot_confusion_matrix(y_true, y_pred, classes,\n",
        "                          normalize=False,\n",
        "                          title=None,\n",
        "                          cmap=plt.cm.Blues):\n",
        "    \"\"\"\n",
        "    This function prints and plots the confusion matrix.\n",
        "    Normalization can be applied by setting `normalize=True`.\n",
        "    \"\"\"\n",
        "    if not title:\n",
        "        if normalize:\n",
        "            title = 'Normalized confusion matrix'\n",
        "        else:\n",
        "            title = 'Confusion matrix, without normalization'\n",
        "\n",
        "    # Compute confusion matrix\n",
        "    cm = confusion_matrix(y_true, y_pred)\n",
        "    # Only use the labels that appear in the data\n",
        "    #classes = classes[unique_labels(y_true, y_pred)]\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "        print(\"Normalized confusion matrix\")\n",
        "    else:\n",
        "        print('Confusion matrix, without normalization')\n",
        "\n",
        "    print(cm)\n",
        "\n",
        "    fig, ax = plt.subplots()\n",
        "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    ax.figure.colorbar(im, ax=ax)\n",
        "    # We want to show all ticks...\n",
        "    ax.set(xticks=np.arange(cm.shape[1]),\n",
        "           yticks=np.arange(cm.shape[0]),\n",
        "           # ... and label them with the respective list entries\n",
        "           xticklabels=classes, yticklabels=classes,\n",
        "           title=title,\n",
        "           ylabel='True label',\n",
        "           xlabel='Predicted label')\n",
        "\n",
        "    # Rotate the tick labels and set their alignment.\n",
        "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",\n",
        "             rotation_mode=\"anchor\")\n",
        "\n",
        "    # Loop over data dimensions and create text annotations.\n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "    for i in range(cm.shape[0]):\n",
        "        for j in range(cm.shape[1]):\n",
        "            ax.text(j, i, format(cm[i, j], fmt),\n",
        "                    ha=\"center\", va=\"center\",\n",
        "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
        "    fig.tight_layout()\n",
        "    return ax\n",
        "\n",
        "# Plot non-normalized confusion matrix\n",
        "plot_confusion_matrix(testY, preds, classes=[x for x in range(10)],\n",
        "                      title='Confusion matrix')"
      ],
      "execution_count": 150,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion matrix, without normalization\n",
            "[[ 978    0    0    0    0    0    1    1    0    0]\n",
            " [   0 1129    1    1    0    0    1    0    3    0]\n",
            " [   1    0 1021    0    4    0    0    4    2    0]\n",
            " [   0    0    0 1004    0    5    0    0    1    0]\n",
            " [   0    0    0    0  979    0    0    0    0    3]\n",
            " [   2    0    0    2    0  884    1    1    1    1]\n",
            " [   2    1    0    0    1    1  953    0    0    0]\n",
            " [   0    3    4    0    1    0    0 1014    5    1]\n",
            " [   3    0    0    0    0    0    1    1  967    2]\n",
            " [   1    0    0    0   12    1    0    0   10  985]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f016b6612e8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 150
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x864 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUMAAAEYCAYAAADGepQzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LK3dkHC6Jw7X",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e6f0ae9f-81f9-415a-f9a7-02a02252e17e"
      },
      "source": [
        "accuracy_score(testY,preds)"
      ],
      "execution_count": 152,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9914"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 152
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xLGdxxAzJWul",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 87
        },
        "outputId": "a4632588-2ad4-435f-849b-9575c12b37a2"
      },
      "source": [
        "precision_recall_fscore_support(testY,preds)"
      ],
      "execution_count": 153,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([0.99, 1.  , 1.  , 1.  , 0.98, 0.99, 1.  , 0.99, 0.98, 0.99]),\n",
              " array([1.  , 0.99, 0.99, 0.99, 1.  , 0.99, 0.99, 0.99, 0.99, 0.98]),\n",
              " array([0.99, 1.  , 0.99, 1.  , 0.99, 0.99, 1.  , 0.99, 0.99, 0.98]),\n",
              " array([ 980, 1135, 1032, 1010,  982,  892,  958, 1028,  974, 1009]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 153
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AuOHEav0KFAK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}