assignment1_Fmnist_file3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Fmnist_file3.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "CmWG0V17hQZ4"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KBnxpKnUOGEu"
      },
      "source": [
        "***Problem Statement***\n",
        "The dataset is similar to MNIST, but includes images of certain clothing and accessory. The objective is to classify images into specific classes using single layer perceptron and multilayer perceptron."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5bEM1FL8O1kR"
      },
      "source": [
        "**Importing the necessary libraries**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D2-38uJZN_GJ"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from keras.layers import Dense,Flatten,Dropout\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-OlZ5iIwPHa-"
      },
      "source": [
        "**Preparing the dataset**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-f4VV17IODtJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9c296244-3a72-42e5-acb1-6f1ffc6bcbfc"
      },
      "source": [
        "(x_train,y_train),(x_test,y_test)=keras.datasets.fashion_mnist.load_data()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "40960/29515 [=========================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "26435584/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "16384/5148 [===============================================================================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n",
            "4431872/4422102 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gzzv7e0rPWVv",
        "outputId": "fc81a843-c707-4f12-ca74-598d9dbda2b3"
      },
      "source": [
        "x_train.shape,x_test.shape"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((60000, 28, 28), (10000, 28, 28))"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        },
        "id": "0PvfFDA1uENY",
        "outputId": "c0b16c91-7189-4256-d7d8-de399e7a3c8a"
      },
      "source": [
        "plt.matshow(x_train[0])"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fc7792e91d0>"
            ]
          },
          "metadata": {},
          "execution_count": 4
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2SEUt0lXPfjg",
        "outputId": "bd305adc-4426-487f-e959-48f57a9db894"
      },
      "source": [
        "y_train.shape,y_test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((60000,), (10000,))"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6iHjHpI-Pmyx"
      },
      "source": [
        "**Pre-processing the data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l2-O9_tVPlkL"
      },
      "source": [
        "x_train=x_train.reshape(x_train.shape[0],x_train.shape[1]*x_train.shape[2]).astype('float32')\n",
        "x_train/=255\n",
        "x_test=x_test.reshape(x_test.shape[0],x_test.shape[1]*x_test.shape[2]).astype('float32')\n",
        "x_test/=255"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ibgBQ3qVQl7b",
        "outputId": "fa9377aa-53fc-4fde-f940-a8e4a5da6a2a"
      },
      "source": [
        "x_train.shape,x_test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((60000, 784), (10000, 784))"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VVNglTJsTrg3"
      },
      "source": [
        "#x_train[0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r9IpxFcRRKZ-"
      },
      "source": [
        "y_train = keras.utils.to_categorical(y_train, 10) \n",
        "y_test = keras.utils.to_categorical(y_test, 10)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n-DYa7rPR0la"
      },
      "source": [
        "**Develop a Single Layer Perceptron model**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x2a8BHXqR_NN"
      },
      "source": [
        "#Building the model\n",
        "model1=keras.Sequential()\n",
        "model1.add(Dense(10,input_shape=(784,),activation='softmax'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B4KMpH9OR_cJ"
      },
      "source": [
        "#compiling the model\n",
        "model1.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vYfRi2RsR_iy",
        "outputId": "1e1b4602-9741-44ca-f589-e3f699372ae1"
      },
      "source": [
        "#Fitting the model\n",
        "history1=model1.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.6192 - accuracy: 0.7935 - val_loss: 0.4980 - val_accuracy: 0.8299\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4720 - accuracy: 0.8395 - val_loss: 0.4571 - val_accuracy: 0.8452\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4437 - accuracy: 0.8464 - val_loss: 0.4439 - val_accuracy: 0.8465\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4292 - accuracy: 0.8527 - val_loss: 0.4426 - val_accuracy: 0.8463\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4172 - accuracy: 0.8552 - val_loss: 0.4319 - val_accuracy: 0.8537\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4110 - accuracy: 0.8587 - val_loss: 0.4345 - val_accuracy: 0.8465\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4046 - accuracy: 0.8592 - val_loss: 0.4218 - val_accuracy: 0.8568\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4000 - accuracy: 0.8606 - val_loss: 0.4218 - val_accuracy: 0.8535\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3965 - accuracy: 0.8631 - val_loss: 0.4292 - val_accuracy: 0.8548\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3935 - accuracy: 0.8628 - val_loss: 0.4169 - val_accuracy: 0.8564\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BKBjKdWuUhhf"
      },
      "source": [
        "**Plotting the curves after training**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Mnmvnqt-R_pM",
        "outputId": "d51bc504-9a0f-47ad-d660-302314e1a23f"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history1.history['accuracy'])\n",
        "plt.plot(history1.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "YNK_GhYKZOeA",
        "outputId": "a87bb0b4-2ea9-4797-910d-6435d92a0310"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history1.history['loss'])\n",
        "plt.plot(history1.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7X9kUY6bZcxX"
      },
      "source": [
        "**Evaluate the model on test data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HR4DhgdUUmZO",
        "outputId": "9b49497d-3e47-4ea7-e47c-80f104a457fb"
      },
      "source": [
        "model1.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.4460 - accuracy: 0.8452\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.44601142406463623, 0.8452000021934509]"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GMOZY-skZtJH"
      },
      "source": [
        "**Analyze the model summary**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1RHBwG62Umcj",
        "outputId": "4d66fede-bd76-47a0-b1a6-49a60a015850"
      },
      "source": [
        "model1.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 10)                7850      \n",
            "=================================================================\n",
            "Total params: 7,850\n",
            "Trainable params: 7,850\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SIgXbCFFUmf9",
        "outputId": "3dfeda14-9001-4453-de30-986d1864181b"
      },
      "source": [
        "model1.get_config()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'layers': [{'class_name': 'InputLayer',\n",
              "   'config': {'batch_input_shape': (None, 784),\n",
              "    'dtype': 'float32',\n",
              "    'name': 'dense_input',\n",
              "    'ragged': False,\n",
              "    'sparse': False}},\n",
              "  {'class_name': 'Dense',\n",
              "   'config': {'activation': 'softmax',\n",
              "    'activity_regularizer': None,\n",
              "    'batch_input_shape': (None, 784),\n",
              "    'bias_constraint': None,\n",
              "    'bias_initializer': {'class_name': 'Zeros', 'config': {}},\n",
              "    'bias_regularizer': None,\n",
              "    'dtype': 'float32',\n",
              "    'kernel_constraint': None,\n",
              "    'kernel_initializer': {'class_name': 'GlorotUniform',\n",
              "     'config': {'seed': None}},\n",
              "    'kernel_regularizer': None,\n",
              "    'name': 'dense',\n",
              "    'trainable': True,\n",
              "    'units': 10,\n",
              "    'use_bias': True}}],\n",
              " 'name': 'sequential'}"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wc7W6wvZZ_GG"
      },
      "source": [
        "**Add a hidden layer to the model to make it Multilayer Perceptron**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-zwmPTojaDj-"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fBiFKCUnaEVN",
        "outputId": "62a22fc0-1955-41c2-902e-a88dacea176f"
      },
      "source": [
        "#Building the model\n",
        "model2=keras.Sequential()\n",
        "model2.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model2.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model2.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history2=model2.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.5204 - accuracy: 0.8197 - val_loss: 0.4218 - val_accuracy: 0.8538\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3868 - accuracy: 0.8611 - val_loss: 0.3956 - val_accuracy: 0.8522\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3444 - accuracy: 0.8750 - val_loss: 0.3686 - val_accuracy: 0.8711\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3198 - accuracy: 0.8830 - val_loss: 0.3741 - val_accuracy: 0.8655\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3016 - accuracy: 0.8895 - val_loss: 0.3625 - val_accuracy: 0.8733\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.2858 - accuracy: 0.8962 - val_loss: 0.3310 - val_accuracy: 0.8799\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.2737 - accuracy: 0.8997 - val_loss: 0.3346 - val_accuracy: 0.8830\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.2598 - accuracy: 0.9040 - val_loss: 0.3346 - val_accuracy: 0.8812\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.2499 - accuracy: 0.9072 - val_loss: 0.3292 - val_accuracy: 0.8839\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.2408 - accuracy: 0.9096 - val_loss: 0.3245 - val_accuracy: 0.8840\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f8jcJ3WpaEjA",
        "outputId": "1f7c9a0e-432d-4a38-d120-9404af76a785"
      },
      "source": [
        "model2.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.3478 - accuracy: 0.8799\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.34783482551574707, 0.8798999786376953]"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6-GKLy0ddjIZ"
      },
      "source": [
        "**Plotting the curves**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Fm2bBEWCcm_z",
        "outputId": "64cc48a3-c599-42ad-a08d-794479964728"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history2.history['accuracy'])\n",
        "plt.plot(history2.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W1WVo0tOcnK8"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wwj3N-Y_duhu"
      },
      "source": [
        "**Add Dropout to prevent overfitting and check its effect on accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tVo0fy1-dqrV",
        "outputId": "d0640bfa-914a-4a21-827a-40a80c562835"
      },
      "source": [
        "#Building the model\n",
        "model3=keras.Sequential()\n",
        "model3.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model3.add(Dropout(0.2))\n",
        "model3.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model3.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history3=model3.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.5578 - accuracy: 0.8032 - val_loss: 0.4103 - val_accuracy: 0.8545\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.4137 - accuracy: 0.8509 - val_loss: 0.3790 - val_accuracy: 0.8654\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3757 - accuracy: 0.8635 - val_loss: 0.3521 - val_accuracy: 0.8745\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3506 - accuracy: 0.8697 - val_loss: 0.3480 - val_accuracy: 0.8772\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3340 - accuracy: 0.8769 - val_loss: 0.3384 - val_accuracy: 0.8770\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3202 - accuracy: 0.8810 - val_loss: 0.3312 - val_accuracy: 0.8818\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3123 - accuracy: 0.8840 - val_loss: 0.3354 - val_accuracy: 0.8793\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 4s 3ms/step - loss: 0.3017 - accuracy: 0.8881 - val_loss: 0.3304 - val_accuracy: 0.8788\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.2947 - accuracy: 0.8903 - val_loss: 0.3175 - val_accuracy: 0.8877\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.2866 - accuracy: 0.8933 - val_loss: 0.3227 - val_accuracy: 0.8867\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zZb4GQhDdqva",
        "outputId": "e6bd3fa4-b653-4340-c08b-2a6925c7c894"
      },
      "source": [
        "model3.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.3463 - accuracy: 0.8785\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3463338017463684, 0.8784999847412109]"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "XY5tDbDAdqzX",
        "outputId": "89d7b781-356b-41b5-ab66-04c91ba05e25"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history3.history['accuracy'])\n",
        "plt.plot(history3.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o9TwwMVPeSIz"
      },
      "source": [
        "**Increase the number of hidden layer neuron and check its accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f8YW22M0dq5k",
        "outputId": "43278078-1970-4cdf-b4c3-a3da48684f5e"
      },
      "source": [
        "#Building the model\n",
        "model4=keras.Sequential()\n",
        "model4.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model4.add(Dropout(0.2))\n",
        "model4.add(Dense(64,activation='relu'))\n",
        "model4.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model4.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history4=model4.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 6s 3ms/step - loss: 0.5664 - accuracy: 0.7959 - val_loss: 0.4004 - val_accuracy: 0.8577\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4169 - accuracy: 0.8486 - val_loss: 0.3941 - val_accuracy: 0.8547\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3812 - accuracy: 0.8592 - val_loss: 0.3522 - val_accuracy: 0.8679\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3584 - accuracy: 0.8671 - val_loss: 0.3526 - val_accuracy: 0.8729\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3406 - accuracy: 0.8734 - val_loss: 0.3334 - val_accuracy: 0.8816\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3266 - accuracy: 0.8789 - val_loss: 0.3541 - val_accuracy: 0.8711\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3164 - accuracy: 0.8837 - val_loss: 0.3262 - val_accuracy: 0.8835\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3073 - accuracy: 0.8868 - val_loss: 0.3332 - val_accuracy: 0.8808\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3001 - accuracy: 0.8879 - val_loss: 0.3200 - val_accuracy: 0.8867\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.2934 - accuracy: 0.8907 - val_loss: 0.3174 - val_accuracy: 0.8854\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zTUm-wiRf7vl"
      },
      "source": [
        "**evaluating the model4**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UlFYSUAkdq-f",
        "outputId": "b0061f0a-316b-48ed-e480-004c065d0a4a"
      },
      "source": [
        "model4.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.3423 - accuracy: 0.8787\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.34234538674354553, 0.8787000179290771]"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U5T6QO6xf9h0"
      },
      "source": [
        "**Plotting the accuracy curves of model4**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "iAIPCftIdrBJ",
        "outputId": "e1acb882-84a7-4d42-c431-fa78bc49c5df"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history3.history['accuracy'])\n",
        "plt.plot(history3.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VilagYChdrDa"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QzwwmEWVgeZA"
      },
      "source": [
        "**Use different optimizers and check accuracy**\n",
        "\n",
        "\n",
        "*   Rmsprop\n",
        "*   Stochastic Gradient Descent\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xJ2xjwRShND8"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CmWG0V17hQZ4"
      },
      "source": [
        "## SGD"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bwDAoVS3drFi",
        "outputId": "f117bc33-f60e-425e-fb4b-3adfbb7d5198"
      },
      "source": [
        "#Building the model\n",
        "model5=keras.Sequential()\n",
        "model5.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model5.add(Dropout(0.2))\n",
        "model5.add(Dense(64,activation='relu'))\n",
        "model5.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model5.compile(optimizer='sgd',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history5=model5.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.8808 - accuracy: 0.7034 - val_loss: 0.5608 - val_accuracy: 0.8131\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.5618 - accuracy: 0.8073 - val_loss: 0.4831 - val_accuracy: 0.8317\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4971 - accuracy: 0.8259 - val_loss: 0.4521 - val_accuracy: 0.8378\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4612 - accuracy: 0.8395 - val_loss: 0.4325 - val_accuracy: 0.8474\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4369 - accuracy: 0.8464 - val_loss: 0.4146 - val_accuracy: 0.8536\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4186 - accuracy: 0.8520 - val_loss: 0.3960 - val_accuracy: 0.8570\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.4034 - accuracy: 0.8551 - val_loss: 0.3947 - val_accuracy: 0.8579\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3923 - accuracy: 0.8610 - val_loss: 0.3759 - val_accuracy: 0.8668\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3821 - accuracy: 0.8637 - val_loss: 0.3765 - val_accuracy: 0.8649\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 5s 3ms/step - loss: 0.3714 - accuracy: 0.8668 - val_loss: 0.3631 - val_accuracy: 0.8702\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8osS2DqedrH3",
        "outputId": "28ceed11-c051-4d5b-ddf4-6ae642d54b5d"
      },
      "source": [
        "model5.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.3859 - accuracy: 0.8616\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3858857750892639, 0.8615999817848206]"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "IPh9XSNWhIYq",
        "outputId": "eafd75d1-8eb3-4668-ce4e-c252196f1f0f"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history5.history['accuracy'])\n",
        "plt.plot(history5.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZYmP5HA6hUrX"
      },
      "source": [
        "## RMS Prop"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KOfQQ1IhhW5o",
        "outputId": "5227b162-53cb-43c9-cfc9-f21c4057a3d9"
      },
      "source": [
        "#Building the model\n",
        "model6=keras.Sequential()\n",
        "model6.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model6.add(Dropout(0.2))\n",
        "model6.add(Dense(64,activation='relu'))\n",
        "model6.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model6.compile(optimizer='rmsprop',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history6=model6.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.5578 - accuracy: 0.7991 - val_loss: 0.4312 - val_accuracy: 0.8421\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.4256 - accuracy: 0.8464 - val_loss: 0.3836 - val_accuracy: 0.8695\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3955 - accuracy: 0.8594 - val_loss: 0.3775 - val_accuracy: 0.8687\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3849 - accuracy: 0.8642 - val_loss: 0.3633 - val_accuracy: 0.8773\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3741 - accuracy: 0.8694 - val_loss: 0.3735 - val_accuracy: 0.8692\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3723 - accuracy: 0.8714 - val_loss: 0.4197 - val_accuracy: 0.8669\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3671 - accuracy: 0.8734 - val_loss: 0.3793 - val_accuracy: 0.8780\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3630 - accuracy: 0.8755 - val_loss: 0.3689 - val_accuracy: 0.8809\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 5s 4ms/step - loss: 0.3625 - accuracy: 0.8764 - val_loss: 0.3867 - val_accuracy: 0.8811\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3603 - accuracy: 0.8795 - val_loss: 0.3990 - val_accuracy: 0.8709\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l9w06QcChXJb",
        "outputId": "c0057328-774e-4fa5-dec0-b2a46094d2b0"
      },
      "source": [
        "model6.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.4496 - accuracy: 0.8597\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.4496222138404846, 0.8597000241279602]"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qiv1qhgkhqzS"
      },
      "source": [
        "## **Increase the hidden layers and check its effect on accuracy**\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I8HX1FDdhtJ1",
        "outputId": "7350a509-d854-4538-dd90-cf00969fc16c"
      },
      "source": [
        "#Building the model\n",
        "model7=keras.Sequential()\n",
        "model7.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model7.add(Dropout(0.2))\n",
        "model7.add(Dense(64,activation='relu'))\n",
        "model7.add(Dropout(0.2))\n",
        "model7.add(Dense(32,activation='relu'))\n",
        "model7.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model7.compile(optimizer='rmsprop',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history7=model7.fit(x_train,y_train,epochs=10,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1500/1500 [==============================] - 7s 4ms/step - loss: 0.6358 - accuracy: 0.7729 - val_loss: 0.4374 - val_accuracy: 0.8372\n",
            "Epoch 2/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.4603 - accuracy: 0.8378 - val_loss: 0.3881 - val_accuracy: 0.8612\n",
            "Epoch 3/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.4252 - accuracy: 0.8507 - val_loss: 0.3722 - val_accuracy: 0.8672\n",
            "Epoch 4/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.4123 - accuracy: 0.8561 - val_loss: 0.3860 - val_accuracy: 0.8637\n",
            "Epoch 5/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3986 - accuracy: 0.8614 - val_loss: 0.3743 - val_accuracy: 0.8688\n",
            "Epoch 6/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3930 - accuracy: 0.8641 - val_loss: 0.3742 - val_accuracy: 0.8698\n",
            "Epoch 7/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3885 - accuracy: 0.8670 - val_loss: 0.3816 - val_accuracy: 0.8691\n",
            "Epoch 8/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3889 - accuracy: 0.8695 - val_loss: 0.3685 - val_accuracy: 0.8758\n",
            "Epoch 9/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3787 - accuracy: 0.8711 - val_loss: 0.4038 - val_accuracy: 0.8698\n",
            "Epoch 10/10\n",
            "1500/1500 [==============================] - 6s 4ms/step - loss: 0.3791 - accuracy: 0.8738 - val_loss: 0.3611 - val_accuracy: 0.8806\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Zdiiq5uwhtYs",
        "outputId": "7693b32c-4a50-4a66-dcb7-155a45421141"
      },
      "source": [
        "model6.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.4496 - accuracy: 0.8597\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.4496222138404846, 0.8597000241279602]"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jFTOAw6-ko5q"
      },
      "source": [
        "\n",
        "\n",
        "\n",
        "```\n",
        "\n",
        "# **Manipulate the batch_size and epochs and check its effect on accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9ow9JrbjhtlL",
        "outputId": "060dde2b-4987-4688-f6df-d254926b39df"
      },
      "source": [
        "#Building the model\n",
        "model8=keras.Sequential()\n",
        "model8.add(Dense(128,input_shape=(784,),activation='relu'))#hidden layer\n",
        "model8.add(Dropout(0.2))\n",
        "model8.add(Dense(64,activation='relu'))\n",
        "model8.add(Dropout(0.2))\n",
        "model8.add(Dense(32,activation='relu'))\n",
        "model8.add(Dense(10,activation='softmax'))#output layer\n",
        "\n",
        "\n",
        "#compiling the model\n",
        "model8.compile(optimizer='adam',loss='categorical_crossentropy',metrics=['accuracy'])\n",
        "\n",
        "#Fitting the model\n",
        "history8=model8.fit(x_train,y_train,epochs=15,batch_size=64,validation_split=0.2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/15\n",
            "750/750 [==============================] - 4s 4ms/step - loss: 0.6686 - accuracy: 0.7611 - val_loss: 0.4310 - val_accuracy: 0.8472\n",
            "Epoch 2/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.4532 - accuracy: 0.8373 - val_loss: 0.3945 - val_accuracy: 0.8593\n",
            "Epoch 3/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.4089 - accuracy: 0.8527 - val_loss: 0.3764 - val_accuracy: 0.8652\n",
            "Epoch 4/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3819 - accuracy: 0.8616 - val_loss: 0.3594 - val_accuracy: 0.8732\n",
            "Epoch 5/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3665 - accuracy: 0.8669 - val_loss: 0.3592 - val_accuracy: 0.8702\n",
            "Epoch 6/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3563 - accuracy: 0.8699 - val_loss: 0.3535 - val_accuracy: 0.8726\n",
            "Epoch 7/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3468 - accuracy: 0.8718 - val_loss: 0.3303 - val_accuracy: 0.8812\n",
            "Epoch 8/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3314 - accuracy: 0.8795 - val_loss: 0.3417 - val_accuracy: 0.8792\n",
            "Epoch 9/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3256 - accuracy: 0.8808 - val_loss: 0.3313 - val_accuracy: 0.8805\n",
            "Epoch 10/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3160 - accuracy: 0.8840 - val_loss: 0.3370 - val_accuracy: 0.8806\n",
            "Epoch 11/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3109 - accuracy: 0.8853 - val_loss: 0.3203 - val_accuracy: 0.8837\n",
            "Epoch 12/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.3019 - accuracy: 0.8873 - val_loss: 0.3192 - val_accuracy: 0.8854\n",
            "Epoch 13/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.2992 - accuracy: 0.8895 - val_loss: 0.3249 - val_accuracy: 0.8860\n",
            "Epoch 14/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.2916 - accuracy: 0.8916 - val_loss: 0.3173 - val_accuracy: 0.8881\n",
            "Epoch 15/15\n",
            "750/750 [==============================] - 3s 4ms/step - loss: 0.2889 - accuracy: 0.8923 - val_loss: 0.3149 - val_accuracy: 0.8873\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MFD46iSFlCkc",
        "outputId": "ccb5257e-dd1f-4595-a806-a48f50e14b78"
      },
      "source": [
        "model8.evaluate(x_test,y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.3380 - accuracy: 0.8779\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3380453586578369, 0.8779000043869019]"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "E8VkWfkvlnCQ",
        "outputId": "8a337491-8209-44aa-eb72-968e3a5f47f3"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history8.history['accuracy'])\n",
        "plt.plot(history8.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxU9b3/8dcn+76HLQESENkU2d2tFWlB69rWosXW1optb63ttVrtVUv9tbe2t1XbW7XuWndLpdKKiijaekUlbEoi+5YNTALZ9+Tz++OcwBAGGCCTk5l8no/HPObMWWY+CeG855zvOd+vqCrGGGNMdxFeF2CMMaZvsoAwxhjjlwWEMcYYvywgjDHG+GUBYYwxxi8LCGOMMX5ZQBgDiMiTIvLLANfdLiLnB7smY7xmAWGMMcYvCwhjwoiIRHldgwkfFhAmZLindm4WkY9FpEFEHhORgSLymojUichSEUn3Wf9iESkUkWoReUdExvosmyQiq9ztXgTiun3Wl0Rkjbvt+yIyIcAaLxSR1SJSKyLFIjK/2/Kz3Perdpdf486PF5Hfi8gOEakRkffceeeKSImf38P57vR8EVkgIs+ISC1wjYhMF5Hl7meUi8ifRCTGZ/vxIvKmiOwRkd0i8jMRGSQijSKS6bPeZBGpEJHoQH52E34sIEyo+TIwEzgRuAh4DfgZkI3z9/xDABE5EXge+JG7bDHwDxGJcXeWfweeBjKAv7rvi7vtJOBx4HogE3gIWCQisQHU1wB8A0gDLgS+JyKXuu873K33f92aJgJr3O1+B0wBznBrugXoDPB3cgmwwP3MZ4EO4MdAFnA6MAP4vltDMrAUeB0YApwAvKWqu4B3gCt83vdq4AVVbQuwDhNmLCBMqPlfVd2tqqXAv4EPVXW1qjYDC4FJ7npfA15V1TfdHdzvgHicHfBpQDRwn6q2qeoCYIXPZ8wDHlLVD1W1Q1WfAlrc7Q5LVd9R1U9UtVNVP8YJqc+5i68Clqrq8+7nVqnqGhGJAL4N3Kiqpe5nvq+qLQH+Tpar6t/dz2xS1ZWq+oGqtqvqdpyA66rhS8AuVf29qjarap2qfuguewqYCyAikcCVOCFq+ikLCBNqdvtMN/l5neRODwF2dC1Q1U6gGMhxl5XqgT1V7vCZHg7c5J6iqRaRamCou91hicipIrLMPTVTA3wX55s87nts8bNZFs4pLn/LAlHcrYYTReSfIrLLPe303wHUAPAKME5E8nGO0mpU9aNjrMmEAQsIE67KcHb0AIiI4OwcS4FyIMed12WYz3Qx8CtVTfN5JKjq8wF87nPAImCoqqYCfwa6PqcYGOlnm0qg+RDLGoAEn58jEuf0lK/uXTI/CKwHRqlqCs4pON8aRvgr3D0KewnnKOJq7Oih37OAMOHqJeBCEZnhNrLehHOa6H1gOdAO/FBEokXkcmC6z7aPAN91jwZERBLdxufkAD43Gdijqs0iMh3ntFKXZ4HzReQKEYkSkUwRmege3TwO3CMiQ0QkUkROd9s8NgJx7udHA7cDR2oLSQZqgXoRGQN8z2fZP4HBIvIjEYkVkWQROdVn+V+Aa4CLsYDo9ywgTFhS1Q0434T/F+cb+kXARaraqqqtwOU4O8I9OO0VL/tsWwBcB/wJ2AtsdtcNxPeBu0SkDrgTJ6i63ncncAFOWO3BaaA+xV38E+ATnLaQPcBvgAhVrXHf81Gco58G4ICrmvz4CU4w1eGE3Ys+NdThnD66CNgFbAI+77P8/3Aax1epqu9pN9MPiQ0YZIzxJSJvA8+p6qNe12K8ZQFhjNlHRKYBb+K0odR5XY/xlp1iMsYAICJP4dwj8SMLBwN2BGGMMeYQ7AjCGGOMX2HTsVdWVpbm5eV5XYYxxoSUlStXVqpq93trgDAKiLy8PAoKCrwuwxhjQoqIHPJyZjvFZIwxxi8LCGOMMX5ZQBhjjPErbNog/Glra6OkpITm5mavSwm6uLg4cnNziY62sV2MMT0jrAOipKSE5ORk8vLyOLDjzvCiqlRVVVFSUkJ+fr7X5RhjwkRYn2Jqbm4mMzMzrMMBQETIzMzsF0dKxpjeE9YBAYR9OHTpLz+nMab3hPUpJmOMCVUdnUpjazsNLR00tLbT2PXc2k59SweNLe00tDrPmUmxXHXqsCO/6VGygAiy6upqnnvuOb7//e8f1XYXXHABzz33HGlpaUGqzBgTbKrKnoZWtlc1sK2ykV01Tc7O3d3xOzv7dhpbO2jweW5obae5rTPgz5k8LM0CIhRVV1fzwAMPHBQQ7e3tREUd+te/ePHiYJdmjOkh1Y2tbKts2BcE2/dNN1DX3H7AujGRESTGRpIQE3XAc2ZiAomxUSTERO57ToqNOnA9d9n+11HEx0QSExWc1gILiCC79dZb2bJlCxMnTiQ6Opq4uDjS09NZv349Gzdu5NJLL6W4uJjm5mZuvPFG5s2bB+zvOqS+vp7Zs2dz1lln8f7775OTk8Mrr7xCfHy8xz+ZMf1LbXMb2yudnf72ysZ9AbC9qoHqxrZ964nAkNR48rMSuWTiEPIyE8nPSiQvK5Hc9HhioyI9/CmOTr8JiF/8o5Cistoefc9xQ1L4+UXjD7vO3Xffzbp161izZg3vvPMOF154IevWrdt3Oerjjz9ORkYGTU1NTJs2jS9/+ctkZmYe8B6bNm3i+eef55FHHuGKK67gb3/7G3Pnzu3Rn8UYA02tHWyrbGBrZb0bBk4QbK9soKqh9YB1h6TGkZeVyAUnDyY/0wmAvMwEhmYkEBcdOiFwOP0mIPqK6dOnH3Cvwh//+EcWLlwIQHFxMZs2bTooIPLz85k4cSIAU6ZMYfv27b1WrzHhRlUpr2lma4UTBFs+q2drZQNbKxoorW46YN2BKbHkZSYyc9xANwCco4HhmeETAofTbwLiSN/0e0tiYuK+6XfeeYelS5eyfPlyEhISOPfcc/3eyxAbG7tvOjIykqampoPWMcYcqKGlnW2VDWypqGdrxf7nbZUNNLV17FsvMSaSEdlJTM1L52vZQxmR7YRAflYiCTE9uItsbYSGCohJgtgkiIo98jYe6zcB4ZXk5GTq6vyP3lhTU0N6ejoJCQmsX7+eDz74oJerMya0dXYqpdVN7hFA/b4Q2FrRwK7a/V+2RCA3PZ4RWUmcNiKTEdmJjMhOZGR2EgOSY3vuPqK2Zti7Daq2wJ4t7vNW57mu7MB1I6IhNtkJi5jk/dOxyW6IJPtMJ0Fsis90t3Wi4pwfsodZQARZZmYmZ555JieddBLx8fEMHDhw37JZs2bx5z//mbFjxzJ69GhOO+00Dys1pm9q7+ikrLqZ7VUN7KhqYHtVIzuqGthR1ciOPY20tu+/HDQ5LooR2UmccUImI7OTGJGVyIjspJ49JdTeCnu3+wSATxDUlAA+wzgnZELGSBjxOec5eSC0NUFLLbTUQ2s9tNS503XQWAV7d7jz3XmByJkC173dMz+fj7AZk3rq1KnafcCgTz/9lLFjx3pUUe/rbz+vCR8t7R2U7G1yAqCycV8Q7NzTSPGeRto79++n4qIjyMt02gGGu20CXUGQlRTTM0cDHW3OjtpvCBSD+tyjEJcGmSOdANj3PAIyRkB8+vHV0dkJbQ1OWLTUOYFxQLDUOdPxGTDlm8f0ESKyUlWn+ltmRxDGmF7R1NrBzj2NBxwJ7KxyXpdVN+GTASTFRjE8M4Fxg1OYfdKgfYGQl5W4/5SQKjRUQnMNtO2FvU2wu8H5ht7W5OxY25qgrdE5/981ve/R5M7vtqy1Edq7tfPFpjg7/NypMOFrBwZCQkbwfmkREftPIzE4eJ9zCBYQxphj1tTaQWV9C5X1LVTVtzrPDa1U1DnPlXUtVDW0UFnfyp5ul4mmJUQzPDORKcPTuXxyLnnuEUFeZgIZiTFIZ4dz3r66GGrWQHExrNvpvi52Tue0B9hBpURAdIL7iIeYROc5OgGSBu6fjnGXRydC2rD9QZCYFZRz/H2dBYQx5gA1jW1U1DdT2bXDd5/3v3amq+pbaGjt8PseSbFRZCXFkJkUS35WIlPzMhiUErfvXoHhGYmkRnc4O/manVC9FvYWw/bi/QFQWwba7f0TsyF1KAwcDyfOcnbi8RnuTt8nAKK7AsANg8iYfrmDP14WEMb0Yy3tHRSW1bJqx15W76xm1c69lNcc/K1cBDITY8hMjCUrOYaJQ9PISoolMymGbPc5y+d5X4NwZydUboDSlVCxHja4O//qYmj4rNuHREBKjhMAw89wntOGus/DIDXX2eGbXmMBYUw/Ul7TxKodThCs3rmXdaW1tHY4Da45afFMGZ7OhNxUBqbEHbDDT0+IITIigG/gdbtgcwGUFjihULp6/5U4kbH7d/ijZ0HqMJ8AGArJQyDSdkl9if1rGBOmDnd0EBMVwYScVK45M4/Jw9KYNCydgSlxR/kB9VC+xgmCEjcQakudZRFRMPAkOOVrkDPVuQwz8wSn0dWEDAuIIDvW7r4B7rvvPubNm0dCQkIQKjPhZldNM6t27mXVjr2s2rmXdWW1++4R6Do6mDwsncnD0xk3OOXoegDt7HBOEZV0HR2sgs+K9l/umZ4Hw05zwiB3KgyaANFHGTimzwlqQIjILOAPQCTwqKre3W35MOApIM1d51ZVXSwi0cCjwGS3xr+o6q+DWWuwHKq770Dcd999zJ071wKiN7Q1wwcPON+I04Y5O7z0PEjPd06BRMV4XeE+zW3OlUO7appZU1zN6p3VrN65l7LuRwdnHMfRQU3p/tNEJSuhbLVz2Sg41/3nTIExFzrPOVOcq3xM2AlaQIhIJHA/MBMoAVaIyCJVLfJZ7XbgJVV9UETGAYuBPOCrQKyqniwiCUCRiDyvqtuDVW+w+Hb3PXPmTAYMGMBLL71ES0sLl112Gb/4xS9oaGjgiiuuoKSkhI6ODu644w52795NWVkZn//858nKymLZsmVe/yjha+Mb8NpPnS4S0obBhteho2X/8q7G0/Q8SB++Pzi6QiQh87ivkOna6VfWO5eIVta3sKemnsaa3bTV7EbrK4horCSquYrkjmoyqSFd6hhFBxOihRviokgZHElyXCSJ0RFECLCrE8oVPugE1Pm2r+7zAa+7zWuugfrdTmER0TB4Akz6+v6jg4wRdkVQPxHMI4jpwGZV3QogIi8AlwC+AaFAijudCpT5zE8UkSggHmgFjq+v7tduhV2fHNdbHGTQyTD77sOu4tvd95IlS1iwYAEfffQRqsrFF1/Mv/71LyoqKhgyZAivvvoq4PTRlJqayj333MOyZcvIyrJvZ0GxZyu8fhtsfB0yR8HVC2Hkec6VN/W7nO4Uuj82vbl/59klJsnniKPbI20YO2s62LC7jsq6Zur2VtJcs4u2us+gvpLIpkpimiudnb7Ukim1jKCGaVJLmjQcXLNAe3QMLbEZdMRnEhcXR0xkpLPDlghnBXEfEZHu64huy7u/lgNfRyfA4FPcU0Unh0SnciY4ghkQOUCxz+sS4NRu68wHlojIDUAicL47fwFOmJQDCcCPVXVP9w8QkXnAPIBhw3p+uL2etmTJEpYsWcKkSZMAqK+vZ9OmTZx99tncdNNN/PSnP+VLX/oSZ599tseVhrnWRnjvXvi/PziNqTPvglO/t/80UkQEpAxxHsPP8LN9A1TvPDg8qjbD5qUH3LzViRCt6ZyMkkEtMeLnvgGBprhUWuMy6YzPQpJOICplAO1pA4lKHuBc+584wDmNk5hNVGwyUfYN3vQCrxuprwSeVNXfi8jpwNMichLO0UcHMARIB/4tIku7jka6qOrDwMPg9MV02E86wjf93qCq3HbbbVx//fUHLVu1ahWLFy/m9ttvZ8aMGdx5550eVBjmVGH9P+H1nzk3Z538VSccUoYc3fvEJMKAsc6ju85OaipLeGXZ+3yybi1D+YxzBzSQk5FES+ogItIHEpU8cN/OnsQBkJBJfGQUdoW/6WuCGRClwFCf17nuPF/XArMAVHW5iMQBWcBVwOuq2gZ8JiL/B0wFthJifLv7/uIXv8gdd9zB17/+dZKSkigtLSU6Opr29nYyMjKYO3cuaWlpPProowdsa6eYekDlJnjtFtjyNgwYB9e8Cnln9ehHNLV28MT723jwnS3UtyRz+aQ5XD5zFLnpdpGBCU3BDIgVwCgRyccJhjk4O35fO4EZwJMiMhaIAyrc+efhHFEkAqcB9wWx1qDx7e579uzZXHXVVZx++ukAJCUl8cwzz7B582ZuvvlmIiIiiI6O5sEHHwRg3rx5zJo1iyFDhlgj9bFqqYd//RaWP+DchTvrNzDtOz16Q1Z7Ryd/XVnCfUs3sru2hRljBnDzrNGMGZRy5I2N6cOC2t23iFyAs2OPBB5X1V+JyF1Agaoucq9cegRIwmmYvkVVl4hIEvAEMA4Q4AlV/Z/DfZZ1993/ft7DUoV1f4Mlt0NdOUycC+f/HJIG9OBHKG8U7uK3b2xga0UDk4elcevssUzPD2Lvnsb0MM+6+1bVxTiXrvrOu9Nnugg408929TiXuhpz9HYXOaeTtv/buRrniqdh6LQe/YjlW6r4zevrWVNczQkDknj46inMHDew50YmM6YP8LqR2pie01wD79wNHz4EcSnwpXth8jfdyz17RlFZLb99Yz3vbKhgcGocv/3yBC6fnENUpHUhYcJP2AeEqvaLb3XhMjLgMenshI9fgDfvdAaQmXINzLizRwdyKd7TyD1vbuTva0pJjo3ittlj+OYZeT03jKUxfVBYB0RcXBxVVVVkZmaGdUioKlVVVcTF9cO+b8rXwuKbofhDyJ0GX/8rDJnUY29fVd/Cn5Zt5tkPdiIC158zku99biSpCdE99hnG9FVhHRC5ubmUlJRQUVHhdSlBFxcXR25u7rFt3LgHlt/v3AA2ZJLz6Os9bzbugbd/CSufcAaMueQBOOXKHqu5oaWdx97bxsP/2kpjaztXTB3KjeePYnCq3a1g+o+wDojo6Gjy8/O9LqPvamuCD/8M/77X6bM/MhY+dC6xJSbJaeDtCowhk5z+h7wIjfYW5y7lig1QudHpVXTru06bw/Tr4dxbIT6tRz6qtb2TF1bs5I9vbaayvoUvjh/IzV8czQkDknvk/Y0JJWEdEOYQOjvg45ecb+C1JXDibDh/vnPUULnR6bmzfI3zvOLR/V1HxKb4hMbE/aHRU6fvWurdANjgjEJW4YbB3m37u5VGnA7z8s+Gc26BQScd88epKhV1Lezc07jvsXB1KTuqGpmen8HD35jC5GHpPfOzGROCgnofRG/ydx+E8WPL27DkTtj9ibODn/n/nJ3toXS0OTvpMjcwylbD7nXQ4Q5AH5e2PywGu89pww4fGo17fELA51Fbsn+diCgnsLJOhOwxkD3amc4adVTDTta3tFPs7vyL3cfOPY0U722ieE8jLe2dB6x/Uk4KN80czbmjs8O63cqYLoe7D8ICor/Y9Ylzlc+WtyFtuHPT2LjLju2UUXsrVHy6PzDKVjv3HnS2OcvjM/aHxsDxzpVFFeudI4LKDdDg0yYUFe/s9LPHQLYbBlmjISMfIo/cENze0Ul5TfO+ANjZLQD2NLQesH5ybBRDMxIYlpHA0Ix4hmUkkOu+zkmLt6uSTL9jAdGf1ZTA27+Ctc875+nPuQWmXdvzXTi3t8Duwv2BUb7GCQ11ey+NS3V2/L4hkD3aGYwngJBSVUr2NrG2pJq1xdUUldeyc08jZdXNdHTu/xuOihBy0t0df/qBQTAsI4HU+Gg7MjDGh2d3UhsPNdc4XVp/8KDT7cSZP4Sz/rPHGnMPEhULOZOdR5e2JqeTvKQBkDTwqNoqqupb+LikhjXF1XxcUs3akpp9RwMxURGMGZTMpKHpXHKKEwBdRwWDUuLspjVjeogFRLhpb4WCx+Hd30DTXpjwNTjvv5x2gd4WHe+MRnYEDS3trCutcY4OSmpYW1xNyd4mwMmUEwckM2PMAE4ZmsYpuWmMHpR8dOMpG2OOiQVEuFCFwoXw1i+cwWtGnOuMdTD4FI8LO1BbRycbdtXtPzIormHTZ3V0nSXKTY/nlNw0vnH6cCbkpnFSTipJsfZnaowX7H9eONjxvtNraelKGDAe5v4NRs7oE+MG76hqYNXOvawtdo4QCstqaXWvHMpIjGFCbiqzThrEKUNTmZCbRlaSDW9pTF9hARHKKjbC0vmw4VVIHuLeTTynRzunO6ay6lpYtLaMhatLWFfqDCUeHx3JyTmpfNM9Mpg4NI3c9HhrMDamD7OACEV1u+Hdu2HlU84A8zPudMZUjvFu5LLmtg6WFO3m5VUl/HtTJR2dykk5Kdx+4VjOGpXFCdlJ1nhsTIixgAgle7ZCwROw4jHoaIHp18E5NzvjG3ugs1P5YFsVC1eV8tq6XdS3tDM4NY5554zg8kk5jBpo3VMYE8osIPq6jnbY9IYTClveAomEcZfAebdD5khPStq0u46XV5fyyupSymqaSYyJZPbJg7l8Ug6njcgkIsJOGxkTDiwg+qraclj9NKx8EmpLnTaGc2+Dyd+AlCG9Xk5lfQuL1pSxcHUpn5TWEBkhnD0qi5/OHsMXxg0iPsbuQDYm3FhA9CWqsO1d52hhw2LobIeR58Hs3zgd6kX27j9XV7vCwlUl/MunXeGOL43j4lOGkJ1sVxwZE84sIPqCpr2w5jnnBreqzRCfDqd9D6Z8q9dPI3V2Kh9u28PC1SW89sku6tx2hevOHsHlk3M40doVjOk3LCC8ogqlq6DgMVj3N6dL7dzpcNlDMO5SiO7d0eG2VzbwUkExr6wpo7S66YB2hVNHZBJp7QrG9DtBDQgRmQX8AYgEHlXVu7stHwY8BaS569yqqovdZROAh4AUoBOYpqrNway3V7Q2wCcLnGAoXwvRic5IaNOuhUEn93o5ja3t/PGtzTz67610qnL2qGxumTXa2hWMMcELCBGJBO4HZgIlwAoRWaSqRT6r3Q68pKoPisg4YDGQJyJRwDPA1aq6VkQygbZg1dorPlvvnEJa+wK01MCAcXDB75y+kuJSPCnpzaLdzF9USGl1E1+ZksvNXxzNwJR+OK61McavYB5BTAc2q+pWABF5AbgE8A0IxTlCAEgFytzpLwAfq+paAFWtCmKdwdPeCuv/ASsehx3vQWSMc4nq1Gth2GmedYVRWt3E/EWFvFm0m1EDknhx3mmcOiLTk1qMMX1XMAMiByj2eV0CnNptnfnAEhG5AUgEznfnnwioiLwBZAMvqOpvu3+AiMwD5gEMG+ZBb6WHs34x/ONGaPjMHaBnPky62rOb2sDpKO/x97Zx39JNKMpPZ43h2rPyrWdUY4xfXjdSXwk8qaq/F5HTgadF5CS3rrOAaUAj8JY7qMVbvhur6sPAw+AMGNS7pR+CKrz/v87obYNPgUsfcDrOO5aR23pQwfY9/NfCdWzYXcf5Ywfw84vGMzTDu645jDF9XzADohQY6vM6153n61pgFoCqLheROCAL52jjX6paCSAii4HJwFv0Ze2t8Op/Oje4jb8MLn3wqMZPDoY9Da3c/dqnvFRQwpDUOB6+egpfGD/I05qMMaEhmAGxAhglIvk4wTAHuKrbOjuBGcCTIjIWiAMqgDeAW0QkAWgFPgfcG8Raj1/jHnjxaqet4ZxbnLuePTxq6OxUFqws4devfUpdczvXnzOCH84YRaKNrWCMCVDQ9haq2i4iP8DZ2UcCj6tqoYjcBRSo6iLgJuAREfkxToP1NeoMkr1XRO7BCRkFFqvqq8Gq9bhVboLnrnDGf778EZhwhaflbNhVx+1//4QV2/cydXg6v7zsJMYM8uZKKWNM6BJnfxz6pk6dqgUFBb3/wVvfgZe+ARHRMOc5GNa9Hb73NLa284e3NvHYv7eRHBfFbbPH8pUpudZ5njHmkNz23an+ltn5huNR8AQs/glkjoKrXoT04Z6VsqRwF7/4RxGl1U1cMTWXW2ePJSMxxrN6jDGhzwLiWHR2wJI74IP74YSZ8JXHPbvZrWRvI/MXFbH0092MHpjMX797OtPyMjypxRgTXiwgjlZLHfztO7DxdTj1u/CFX/V6L6vg3NPw2Hvb+MPSTQDcNnsM3z4rn2gbtc0Y00MsII5G9U54bg5UrIcLfw/TvuNJGR9t28Ptf/+EjbvrmTluIPMvHk9OmreX0xpjwo8FRKCKV8ALVzr3Osxd4IzT0MtUld8v2ciflm0mJy2eR78xlfPHDez1Oowx/YMFRCA+WQB//z6kDIZrXoXs0Z6U8cA7W/jTss18dUouv7hkPAkx9s9njAke28Mcjiq8+xt459cw7Az42jOQ6E2ndk9/sIP/eWMDl04cwm++PMEuXTXGBJ0FxKG0NcErP4B1C+CUq+Ci+yDKmyE2/766lDtfWcf5YwfwP189xcLBGNMrLCD8qf8MXrgKSlY4vbCe+SPPuuZeWrSbm/66llPzM/jTVZPtKiVjTK+xgOhudyE89zVoqIQrnoZxF3tWyvtbKvn+c6s4aUgKj35zGnHRNsKbMab3WED42vgGLPg2xCbDt1+DIZM8K2VtcTXXPVXA8IwEnvzWdJKskz1jTC+z8xXgNEYvfwCenwOZI+G6tz0Nh4276/jmEx+RkRTDM985lXTrMsMY4wH7WtrRBotvhpVPwNiL4LKHICbRs3J2VjUy99EPiYmM4NlrT7Mxoo0xnrGAqCmGdS/DWf8J593h6RgOu2ubmfvYh7R2dPLivNMZlmkjvhljvGMBkTECfvARJHs7ytrehlaufuxDqupbePa60xg9KNnTeowxxgICPA+H+pZ2rnniI7ZXNfLkt6YxcWiap/UYYwxYI7Xnmts6uO6pAtaV1XL/VZM5Y2SW1yUZYwxgAeGpto5OfvDcapZvreJ3X53ATOt4zxjTh1hAeKSzU7llwccs/XQ3d10ynssm5XpdkjHGHMACwgOqyvx/FLJwdSk3f3E03zg9z+uSjDHmIBYQHvj9ko38ZfkO5p0zgu+fO9Lrcowxxq+gBoSIzBKRDSKyWURu9bN8mIgsE5HVIvKxiFzgZ3m9iPwkmHX2pofedcZ0mDNtKLfNHoN41AmgMcYcSdACQkQigfuB2cA44EoRGddttduBl1R1EjAHeKDb8nuA14JVY297/qOd/Pq19Vw4YTC/uuxkC9g083wAABY1SURBVAdjTJ8WzCOI6cBmVd2qqq3AC8Al3dZRIMWdTgXKuhaIyKXANqAwiDX2mn+sLeNnCz/h3NHZ3HvFRCJtTAdjTB8XUECIyMsicqGIHE2g5ADFPq9L3Hm+5gNzRaQEWAzc4H5eEvBT4BdHqGueiBSISEFFRcVRlNa7lq3/jB+/uIapw9N58OtTiImyph9jTN8X6J7qAeAqYJOI3C0iPTUo85XAk6qaC1wAPO2G0HzgXlWtP9zGqvqwqk5V1anZ2dk9VFLP+mjbHr77zEpGD0rmsWumER9jYzoYY0JDQF1tqOpSYKmIpOLs1JeKSDHwCPCMqrb52awUGOrzOted5+taYJb7GctFJA7IAk4FviIivwXSgE4RaVbVPwX+o3lvXWkN1z65gpz0eJ769nRS4qK9LskYYwIW8LkOEckErgG+A6wG/gBMBt48xCYrgFEiki8iMTiN0Iu6rbMTmOG+/1ggDqhQ1bNVNU9V84D7gP8OtXAorW7iG49/REp8NM9ceypZSd6MZ22MMccqoCMIEVkIjAaeBi5S1XJ30YsiUuBvG1VtF5EfAG8AkcDjqlooIncBBaq6CLgJeEREfozTYH2Nqurx/Uh9wxvrdrGnoZU3fnQOQ9LivS7HGGOOWqC9uf5RVZf5W6CqUw+1kaouxml89p13p890EXDm4T5YVecHWGOfUlhWS1ZSrHXbbYwJWYGeYhonIvv6oBaRdBH5fpBqCgtF5bWMH5Jy5BWNMaaPCjQgrlPV6q4XqroXuC44JYW+lvYONu2uY5wFhDEmhAUaEJHic9uve5d0THBKCn2bdtfT3ql2BGGMCWmBtkG8jtMg/ZD7+np3nvGjqKwWgHGDLSCMMaEr0ID4KU4ofM99/SbwaFAqCgNF5bUkxESSl5nodSnGGHPMAr1RrhN40H2YIygsq2Hs4BQirL8lY0wIC7QvplEiskBEikRka9cj2MWFos5O5dPyOmt/MMaEvEAbqZ/AOXpoBz4P/AV4JlhFhbLivY3Ut7Rb+4MxJuQFGhDxqvoWIKq6w7157cLglRW6Ct0G6vFDUj2uxBhjjk+gjdQtbi+rm9zuM0qBpOCVFbqKymqJjBBGDbRfjzEmtAV6BHEjkAD8EJgCzAW+GayiQllhWQ2jBiQRF23dehtjQtsRjyDcm+K+pqo/AeqBbwW9qhBWVF7LmSOzvC7DGGOO2xGPIFS1AzirF2oJeZX1LeyubbEuNowxYSHQNojVIrII+CvQ0DVTVV8OSlUhat8d1BYQxpgwEGhAxAFVwHk+8xSwgPBRaF1sGGPCSKB3Ulu7QwCKymvJSYsnLcH6MTTGhL5AR5R7AueI4QCq+u0eryiEFZbV2OklY0zYCPQU0z99puOAy4Cyni8ndDW2trOtsoGLTxnidSnGGNMjAj3F9Dff1yLyPPBeUCoKUZ+W16Fq7Q/GmPAR6I1y3Y0CBvRkIaGuqNztYiPHutgwxoSHQNsg6jiwDWIXzhgRxlVUVkNqfDRDUuO8LsUYY3pEQEcQqpqsqik+jxO7n3byR0RmicgGEdksIrf6WT5MRJaJyGoR+VhELnDnzxSRlSLyift83sHv3rcUldUyfkgKPiOzGmNMSAt0PIjLRCTV53WaiFx6hG0igfuB2cA44EoRGddttduBl1R1EjAHeMCdXwlcpKon4/T59HQgdXqlvaOT9bvqrP3BGBNWAm2D+Lmq1nS9UNVq4OdH2GY6sFlVt6pqK/ACcEm3dRTo2qum4l4ZpaqrVbXrKqlCIF5EYgOstddtrWygpb2T8TkWEMaY8BFoQPhb70jtFzlAsc/rEneer/nAXBEpARYDN/h5ny8Dq1S1pfsCEZknIgUiUlBRUXGEcoJnXxcbg62B2hgTPgINiAIRuUdERrqPe4CVPfD5VwJPqmoucAHwtDvuBAAiMh74DXC9v41V9WFVnaqqU7Ozs3ugnGNTWFZDTFQEI7MTPavBGGN6WqABcQPQCryIc6qoGfiPI2xTCgz1eZ3rzvN1LfASgKoux7kJLwtARHKBhcA3VHVLgHV6oqi8ljGDkomKPNarho0xpu8J9Ea5BuCgq5COYAUwSkTycYJhDnBVt3V2AjOAJ0VkLE5AVIhIGvAqcKuq/t9Rfm6vUlUKy2qZfdIgr0sxxpgeFehVTG+6O+2u1+ki8sbhtlHVduAHwBvApzhXKxWKyF0icrG72k3AdSKyFngeuEZV1d3uBOBOEVnjPvrkjXnlNc1UN7bZFUzGmLATaF9MWe6VSwCo6t5Adtiquhin8dl33p0+00XAmX62+yXwywBr89S+Lr6HWAO1MSa8BHrSvFNEhnW9EJE8/PTu2h8VldUiAmMGJXtdijHG9KhAjyD+C3hPRN4FBDgbmBe0qkJIYVkN+ZmJJMYG+qs0xpjQEGhXG68DU4ENOG0FNwFNQawrZBSV19oYEMaYsBRoZ33fAW7EuVR1DXAasJwDhyDtd2oa2yjZ28RVpw478srGGBNiAm2DuBGYBuxQ1c8Dk4Dqw28S/vZ18W0N1MaYMBRoQDSrajOAiMSq6npgdPDKCg2FZU73VHaJqzEmHAXaslri3gfxd+BNEdkL7AheWaGhqLyWAcmxZCf32X4EjTHmmAV6J/Vl7uR8EVmG0/Pq60GrKkQUlVkDtTEmfB31tZmq+m4wCgk1Le0dbP6snhlj++QN3sYYc9ysd7ljtGl3Pe2dal18G2PClgXEMepqoB5vp5iMMWHKAuIYFZXVkhQbxbCMBK9LMcaYoLCAOEaFZbWMHZxMRIR4XYoxxgSFBcQx6OxUPi2vtfsfjDFhzQLiGOzY00hDa4fdQW2MCWsWEMegaN8YEHYEYYwJXxYQx6CwrIaoCGHUwCSvSzHGmKCxgDgGReW1nDAgidioSK9LMcaYoLGAOAaF1sWGMaYfsIA4Sp/VNVNR12IN1MaYsGcBcZT2NVDbJa7GmDAX1IAQkVkiskFENovIrX6WDxORZSKyWkQ+FpELfJbd5m63QUS+GMw6j0bXIEF2iskYE+6OujfXQIlIJHA/MBMoAVaIyCJVLfJZ7XbgJVV9UETGAYuBPHd6DjAeGAIsFZETVbUjWPUGqqisltz0eFLjo70uxRhjgiqYRxDTgc2qulVVW4EXgEu6raNA11fxVKDMnb4EeEFVW1R1G7DZfT/PFZXVWgd9xph+IZgBkQMU+7wucef5mg/MFZESnKOHG45iW0RknogUiEhBRUVFT9V9SA0t7WyrarAuvo0x/YLXjdRXAk+qai5wAfC0iARck6o+rKpTVXVqdnZ20Irssn5XLarWxbcxpn8IWhsEUAoM9Xmd687zdS0wC0BVl4tIHJAV4La9zrrYMMb0J8E8glgBjBKRfBGJwWl0XtRtnZ3ADAARGQvEARXuenNEJFZE8oFRwEdBrDUghWW1pCdEMzg1zutSjDEm6IJ2BKGq7SLyA+ANIBJ4XFULReQuoEBVFwE3AY+IyI9xGqyvUVUFCkXkJaAIaAf+o09cwVTu3EEtYmNAGGPCXzBPMaGqi3Ean33n3ekzXQSceYhtfwX8Kpj1HY22jk7W76rjmjPyvC7FGGN6hdeN1CFja0UDre2ddge1MabfsIAIUGFZDWBXMBlj+g8LiAAVldUSGxVBflai16UYY0yvsIAIUGFZLWMGpxAVab8yY0z/YHu7AKiqcwWTtT8YY/oRC4gAlFY3UdPUZjfIGWP6FQuIAHTdQW0N1MaY/sQCIgBF5bWIwJhByV6XYowxvcYCIgCFZbWMyEokISao9xUaY0yfYgERgKKyWsbZGNTGmH7GAuIIqhtbKa1usvYHY0y/YwFxBPvGoLZLXI0x/YwFxBHYGBDGmP7KAuIIispqGZgSS1ZSrNelGGNMr7KAOILCslrGWwO1MaYfsoA4jOa2DjZX1Fv7gzGmX7KAOIyNu+vo6FS7gskY0y9ZQByGNVAbY/ozC4jDKCyrJTk2iqHpCV6XYowxvc4C4jCKymsZOziFiAjxuhRjjOl1FhCH0NGpfFpea6eXjDH9VlADQkRmicgGEdksIrf6WX6viKxxHxtFpNpn2W9FpFBEPhWRP4pIr36N31HVQGNrhwWEMabfClr3pCISCdwPzARKgBUiskhVi7rWUdUf+6x/AzDJnT4DOBOY4C5+D/gc8E6w6u3OutgwxvR3wTyCmA5sVtWtqtoKvABccpj1rwSed6cViANigFggGtgdxFoPUlhWS3SkcOJAGwPCGNM/BTMgcoBin9cl7ryDiMhwIB94G0BVlwPLgHL38Yaqfupnu3kiUiAiBRUVFT1afFFZLScMSCYmypppjDH9U1/Z+80BFqhqB4CInACMBXJxQuU8ETm7+0aq+rCqTlXVqdnZ2T1akNPFhp1eMsb0X8EMiFJgqM/rXHeeP3PYf3oJ4DLgA1WtV9V64DXg9KBU6cdndc1U1rdY+4Mxpl8LZkCsAEaJSL6IxOCEwKLuK4nIGCAdWO4zeyfwORGJEpFonAbqg04xBUuhewe1HUEYY/qzoAWEqrYDPwDewNm5v6SqhSJyl4hc7LPqHOAFVVWfeQuALcAnwFpgrar+I1i1dtfVxcZYCwhjTD8WtMtcAVR1MbC427w7u72e72e7DuD6YNZ2OEVltQzLSCAlLtqrEowxxnN9pZG6Tykqr7X2B2NMv2cB0U19SzvbKhus/cEY0+9ZQHSzvty6+DbGGLCAOMj+K5hsmFFjTP9mAdFNUVktGYkxDEyJ9boUY4zxlAVEN4XlNYwfkkIvdx5rjDF9jgWEj7aOTjbuqrcrmIwxBguIA2ypqKe1o9MaqI0xBguIAxSWWhcbxhjTxQLCR1F5LXHREeRnJXldijHGeM4CwkdhWQ1jBqUQGWEN1MYYYwHhUlWKymqt/cEYY1wWEK6SvU3UNrdb+4MxxrgsIFxFXV1s2CWuxhgDWEDsU1hWS4TAmEEWEMYYAxYQ+xSV1TIiO4n4mEivSzHGmD7BAsJVVFZj7Q/GGOPDAgLY29BKWU2ztT8YY4wPCwj2N1BbF9/GGLOfBQRO+wPYIEHGGOPLAgLnDurBqXFkJMZ4XYoxxvQZQQ0IEZklIhtEZLOI3Opn+b0issZ9bBSRap9lw0RkiYh8KiJFIpIXrDqLymut/cEYY7qJCtYbi0gkcD8wEygBVojIIlUt6lpHVX/ss/4NwCSft/gL8CtVfVNEkoDOYNTZ3NbBlooGZo0fFIy3N8aYkBXMI4jpwGZV3aqqrcALwCWHWf9K4HkAERkHRKnqmwCqWq+qjcEosq65nQtPHsz0/MxgvL0xxoSsoB1BADlAsc/rEuBUfyuKyHAgH3jbnXUiUC0iL7vzlwK3qmpHTxeZnRzLH6+cdOQVjTGmn+krjdRzgAU+ARAFnA38BJgGjACu6b6RiMwTkQIRKaioqOitWo0xpl8IZkCUAkN9Xue68/yZg3t6yVUCrHFPT7UDfwcmd99IVR9W1amqOjU7O7uHyjbGGAPBDYgVwCgRyReRGJwQWNR9JREZA6QDy7ttmyYiXXv984Ci7tsaY4wJnqAFhPvN/wfAG8CnwEuqWigid4nIxT6rzgFeUFX12bYD5/TSWyLyCSDAI8Gq1RhjzMHEZ78c0qZOnaoFBQVel2GMMSFFRFaq6lR/y/pKI7Uxxpg+xgLCGGOMXxYQxhhj/AqbNggRqQB2HMdbZAGVPVROsIVSrRBa9YZSrRBa9YZSrRBa9R5PrcNV1e99AmETEMdLRAoO1VDT14RSrRBa9YZSrRBa9YZSrRBa9QarVjvFZIwxxi8LCGOMMX5ZQOz3sNcFHIVQqhVCq95QqhVCq95QqhVCq96g1GptEMYYY/yyIwhjjDF+WUAYY4zxq98HxJHGze5LRGSoiCxzx+guFJEbva7pSEQkUkRWi8g/va7lSEQkTUQWiMh6dyz0072u6VBE5Mfu38A6EXleROK8rsmXiDwuIp+JyDqfeRki8qaIbHKf072sscshav0f9+/gYxFZKCJpXtboy1+9PstuEhEVkaye+Kx+HRA+42bPBsYBV7rDnfZV7cBNqjoOOA34jz5eL8CNOL35hoI/AK+r6hjgFPpo3SKSA/wQmKqqJwGROL0i9yVPArO6zbsVeEtVRwFvua/7gic5uNY3gZNUdQKwEbitt4s6jCc5uF5EZCjwBWBnT31Qvw4Ijn7cbE+parmqrnKn63B2YDneVnVoIpILXAg86nUtRyIiqcA5wGMAqtqqqtXeVnVYUUC8iEQBCUCZx/UcQFX/BezpNvsS4Cl3+ing0l4t6hD81aqqS9whCwA+wBnwrE84xO8W4F7gFqDHrjzq7wHhb9zsPrvD9SUiecAk4ENvKzms+3D+YDu9LiQA+UAF8IR7SuxREUn0uih/VLUU+B3ON8VyoEZVl3hbVUAGqmq5O70LGOhlMUfh28BrXhdxOCJyCVCqqmt78n37e0CEJBFJAv4G/EhVa72uxx8R+RLwmaqu9LqWAEXhDGv7oKpOAhroO6dADuCeu78EJ9SGAIkiMtfbqo6OO0BYn7/GXkT+C+fU7rNe13IoIpIA/Ay4s6ffu78HxNGMm90niEg0Tjg8q6ove13PYZwJXCwi23FO3Z0nIs94W9JhlQAlqtp1RLYAP+Og9xHnA9tUtUJV24CXgTM8rikQu0VkMID7/JnH9RyWiFwDfAn4uu+Il33QSJwvC2vd/2+5wCoRGXS8b9zfAyKgcbP7ChERnHPkn6rqPV7Xcziqepuq5qpqHs7v9W1V7bPfclV1F1AsIqPdWTPou+Og7wROE5EE929iBn20Qb2bRcA33elvAq94WMthicgsnNOjF6tqo9f1HI6qfqKqA1Q1z/3/VgJMdv+mj0u/DohDjZvtbVWHdSZwNc638TXu4wKviwojNwDPisjHwETgvz2uxy/3KGcBsAr4BOf/cZ/qFkJEngeWA6NFpERErgXuBmaKyCaco6C7vayxyyFq/ROQDLzp/j/7s6dF+jhEvcH5rL595GSMMcYr/foIwhhjzKFZQBhjjPHLAsIYY4xfFhDGGGP8soAwxhjjlwWEMX2AiJwbCj3emv7FAsIYY4xfFhDGHAURmSsiH7k3Tz3kjndRLyL3uuMzvCUi2e66E0XkA58xBdLd+SeIyFIRWSsiq0RkpPv2ST7jUTzr3iVtjGcsIIwJkIiMBb4GnKmqE4EO4OtAIlCgquOBd4Gfu5v8BfipO6bAJz7znwXuV9VTcPpQ6urhdBLwI5yxSUbg3DlvjGeivC7AmBAyA5gCrHC/3MfjdDjXCbzorvMM8LI7vkSaqr7rzn8K+KuIJAM5qroQQFWbAdz3+0hVS9zXa4A84L3g/1jG+GcBYUzgBHhKVQ8YXUxE7ui23rH2X9PiM92B/f80HrNTTMYE7i3gKyIyAPaNsTwc5//RV9x1rgLeU9UaYK+InO3Ovxp41x0JsERELnXfI9btz9+YPse+oRgTIFUtEpHbgSUiEgG0Af+BM7jQdHfZZzjtFOB0af1nNwC2At9y518NPCQid7nv8dVe/DGMCZj15mrMcRKRelVN8roOY3qanWIyxhjjlx1BGGOM8cuOIIwxxvhlAWGMMcYvCwhjjDF+WUAYY4zxywLCGGOMX/8fSVtbuXml1jcAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OWYVfgn_hXMp"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K7cd8qzNhXP7"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bg1X_WFClGa0"
      },
      "source": [
        "**What parameters should be chosen to classify the images into various categories?**\n",
        "\n",
        "\n",
        "\n",
        "*   Number of hidden layers\n",
        "*   Number of units in hidden layers\n",
        "*   activation function in hidden layer\n",
        "*   activation function in output layer\n",
        "*   epochs\n",
        "*   optimizer\n",
        "*   batch_size\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    }
  ]
}