data/MNIST_CNN.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ki3N30vdGe11"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "from keras.datasets import mnist\n",
        "from tensorflow.keras import regularizers\n",
        "import numpy as np\n",
        "import tensorflow as tf"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QkmtW_UqGe18",
        "scrolled": true
      },
      "outputs": [],
      "source": [
        "# (X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
        "# npzfile = np.load(\"/content/neg_with_noise.npz\")\n",
        "# #print(type(mnist.load_data()))\n",
        "# npfile=sorted(npzfile.files)\n",
        "# X_N_train=npzfile['arr_0']\n",
        "# y_N_train=npzfile['arr_1']\n",
        "# X_N_test=npzfile['arr_2']\n",
        "# y_N_test=npzfile['arr_3']\n",
        "# #print(X_N_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DJqZQNRzVk0c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "27ed387c-8cbe-40e5-a96e-5d698b4868a8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(10000, 28, 28) 10000\n"
          ]
        }
      ],
      "source": [
        "# print(X_N_train.shape,len(X_N_train))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8vOVX9JhVko1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ab4a9cd1-4ed9-4949-cbfd-9c948d2080bb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(10000,) 10000\n"
          ]
        }
      ],
      "source": [
        "# print(y_N_train.shape, len(y_N_train))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-tUwrYFIVkOP",
        "outputId": "cd8784a4-81e1-48e5-ef78-d4240e7afa76"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1000, 28, 28) 1000\n"
          ]
        }
      ],
      "source": [
        "# print(X_N_test.shape, len(X_N_test))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xH_UNBnXV4q3",
        "outputId": "fb507dcb-11fb-42be-b66b-75853a7f1daa"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1000,) 1000\n"
          ]
        }
      ],
      "source": [
        "# print(y_N_test.shape , len(y_N_test))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lc9OImyeWCy3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b58ae9d5-5744-4d35-8d62-daba83291cb6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'tuple'>\n"
          ]
        }
      ],
      "source": [
        "# X_train=np.append(X_train, X_N_train,0)\n",
        "# y_train=np.append(y_train, y_N_train,0)\n",
        "# X_test=np.append(X_test, X_N_test,0)\n",
        "# y_test=np.append(y_test, y_N_test,0)\n",
        "npzfile = np.load(\"/content/minst_with_P&N&noise.npz\")\n",
        "print(type(mnist.load_data()))\n",
        "npfile=sorted(npzfile.files)\n",
        "X_train=npzfile['arr_0']\n",
        "y_train=npzfile['arr_1']\n",
        "X_test=npzfile['arr_2']\n",
        "y_test=npzfile['arr_3']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iI1Oh5jvbIvk"
      },
      "outputs": [],
      "source": [
        "from sklearn.utils import shuffle\n",
        "for i in range(130):\n",
        "  X_train, y_train = shuffle(X_train, y_train)\n",
        "  X_test, y_test = shuffle(X_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tct7KKKPmFtz",
        "outputId": "9efd8945-ce63-4c64-fb25-e1da44d796ca"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(11000,)"
            ]
          },
          "metadata": {},
          "execution_count": 121
        }
      ],
      "source": [
        "y_test.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ewEB9WtDGe19",
        "outputId": "7766abe2-9496-4335-9cf3-95494c4786f3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "70000"
            ]
          },
          "metadata": {},
          "execution_count": 122
        }
      ],
      "source": [
        "len(X_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IqMTkhYOGe2A",
        "outputId": "7148df5e-8876-4faa-be17-891b3d43b601"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "11000"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ],
      "source": [
        "len(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v5eKUOG4Ge2B",
        "outputId": "9059bcae-9d0d-4604-c32b-853d91669890"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(70000, 28, 28) (70000,) (11000, 28, 28) (11000,)\n"
          ]
        }
      ],
      "source": [
        "# from google.colab import drive\n",
        "#drive.mount('/content/drive')\n",
        "print(X_train.shape,y_train.shape,X_test.shape,y_test.shape)\n",
        "# np.savez(\"/content/minst_with_P&N\", X_train,y_train,X_test,y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "Po_s8HBkGe2C",
        "outputId": "a8eb12ac-43a0-4214-d191-4b5d6a11ad42"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(28, 28)\n",
            "Digit class: 4\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.imshow(X_train[1], 'gray')\n",
        "print(X_train[1].shape)\n",
        "print(\"Digit class:\", y_train[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "L4_zmwOcGe2D",
        "scrolled": true
      },
      "outputs": [],
      "source": [
        "X_train = X_train.astype('float32') / 255\n",
        "X_test = X_test.astype('float32') / 255\n",
        "#X_train"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H6pXpvz3Ge2E",
        "outputId": "8d339ed5-dd94-487e-a9b5-de9768c9c41a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 1., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., ..., 0., 0., 1.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 1., 0.]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 127
        }
      ],
      "source": [
        "from keras.utils import np_utils\n",
        "\n",
        "Y_train = np_utils.to_categorical(y_train, 11) # 10 classes to codify\n",
        "Y_test = np_utils.to_categorical(y_test, 11)\n",
        "Y_train"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JqJhld0cGe2G",
        "outputId": "d72e1b2a-15a2-4173-98d4-cad2a0f4043c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(70000, 28, 28, 1)\n",
            "24500.0\n",
            "(45500, 28, 28, 1)\n",
            "70000\n"
          ]
        }
      ],
      "source": [
        "import keras\n",
        "from keras.models import Sequential\n",
        "from keras.layers.core import Dense\n",
        "from keras.layers.core import Activation\n",
        "from keras.layers.core import Dropout\n",
        "from keras.layers.convolutional import Convolution2D, MaxPooling2D, AveragePooling2D\n",
        "from keras.layers.core import Flatten\n",
        "\n",
        "train_T= X_train.reshape(70000, 28, 28, 1)\n",
        "val_data=len(train_T)*35/100\n",
        "X_valtensor = train_T[:int(val_data)]\n",
        "Y_valtensor = Y_train[:int(val_data)]\n",
        "X_traintensor = train_T[int(val_data):]\n",
        "Y_traintensor = Y_train[int(val_data):]\n",
        "testtensor = X_test.reshape(11000, 28, 28, 1)\n",
        "print(train_T.shape)\n",
        "print(len(train_T)*35/100)\n",
        "print(X_traintensor.shape)\n",
        "print(len(X_valtensor)+len(X_traintensor))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q9yAT6rhGe2H",
        "outputId": "2d06a785-7245-4cdd-d59a-49a5811621e7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_11\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_31 (Conv2D)          (None, 26, 26, 16)        160       \n",
            "                                                                 \n",
            " activation_42 (Activation)  (None, 26, 26, 16)        0         \n",
            "                                                                 \n",
            " conv2d_32 (Conv2D)          (None, 24, 24, 32)        4640      \n",
            "                                                                 \n",
            " activation_43 (Activation)  (None, 24, 24, 32)        0         \n",
            "                                                                 \n",
            " max_pooling2d_11 (MaxPoolin  (None, 12, 12, 32)       0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_22 (Dropout)        (None, 12, 12, 32)        0         \n",
            "                                                                 \n",
            " conv2d_33 (Conv2D)          (None, 10, 10, 64)        18496     \n",
            "                                                                 \n",
            " activation_44 (Activation)  (None, 10, 10, 64)        0         \n",
            "                                                                 \n",
            " flatten_11 (Flatten)        (None, 6400)              0         \n",
            "                                                                 \n",
            " dense_22 (Dense)            (None, 128)               819328    \n",
            "                                                                 \n",
            " activation_45 (Activation)  (None, 128)               0         \n",
            "                                                                 \n",
            " dropout_23 (Dropout)        (None, 128)               0         \n",
            "                                                                 \n",
            " dense_23 (Dense)            (None, 11)                1419      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 844,043\n",
            "Trainable params: 844,043\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        }
      ],
      "source": [
        "img_rows = 28\n",
        "img_cols = 28\n",
        "kernel_size = 3\n",
        "pool_size = 2\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Convolution2D(16, kernel_size=(3, 3),input_shape=(img_rows, img_cols, 1)))\n",
        "model.add(Activation('relu'))\n",
        "\n",
        "model.add(Convolution2D(32, (3, 3)))\n",
        "model.add(Activation('relu')),\n",
        "\n",
        "model.add(MaxPooling2D(pool_size=(pool_size, pool_size)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Convolution2D(64, (3, 3)))\n",
        "model.add(Activation('relu')),\n",
        "\n",
        "model.add(Flatten())\n",
        "\n",
        "model.add(Dense(128)),\n",
        "model.add(Activation('relu'))\n",
        "model.add(Dropout(0.5))\n",
        "#model.add(Dense(80, activation = 'relu'))\n",
        "model.add(Dense(11, activation = 'softmax'))\n",
        "print(model.summary())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0PWLCmV0Ge2J",
        "scrolled": true,
        "outputId": "67d20b24-dec7-4538-f90a-600dfcf1e5a9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0270 - accuracy: 0.9915 - val_loss: 0.0332 - val_accuracy: 0.9903\n",
            "Epoch 2/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0220 - accuracy: 0.9933 - val_loss: 0.0316 - val_accuracy: 0.9912\n",
            "Epoch 3/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0195 - accuracy: 0.9934 - val_loss: 0.0365 - val_accuracy: 0.9911\n",
            "Epoch 4/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0182 - accuracy: 0.9941 - val_loss: 0.0319 - val_accuracy: 0.9921\n",
            "Epoch 5/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0182 - accuracy: 0.9940 - val_loss: 0.0325 - val_accuracy: 0.9922\n",
            "Epoch 6/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0155 - accuracy: 0.9949 - val_loss: 0.0402 - val_accuracy: 0.9917\n",
            "Epoch 7/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0148 - accuracy: 0.9951 - val_loss: 0.0323 - val_accuracy: 0.9922\n",
            "Epoch 8/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0144 - accuracy: 0.9949 - val_loss: 0.0404 - val_accuracy: 0.9917\n",
            "Epoch 9/9\n",
            "356/356 [==============================] - 59s 166ms/step - loss: 0.0137 - accuracy: 0.9955 - val_loss: 0.0393 - val_accuracy: 0.9912\n"
          ]
        }
      ],
      "source": [
        "model.compile(loss='categorical_crossentropy', optimizer=tf.keras.optimizers.Adam(learning_rate=1e-3), metrics=[\"accuracy\"])\n",
        "\n",
        "model_history = model.fit(\n",
        "    X_traintensor,\n",
        "    Y_traintensor,\n",
        "    batch_size=128,\n",
        "    epochs=9,\n",
        "    verbose=1,\n",
        "    validation_data=(X_valtensor, Y_valtensor)\n",
        ")#24"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 463
        },
        "id": "cWGlmYbpGe2K",
        "outputId": "83f29b44-0b9f-470f-deda-0db6beac5e9a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "344/344 [==============================] - 4s 11ms/step - loss: 0.0312 - accuracy: 0.9923\n",
            "Test loss 0.031244434416294098\n",
            "Test accuracy 0.9922727346420288\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1296x432 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "preds = model.predict(testtensor)\n",
        "\n",
        "score = model.evaluate(testtensor, Y_test)\n",
        "print(\"Test loss\", score[0])\n",
        "print(\"Test accuracy\", score[1])\n",
        "\n",
        "plt.figure(figsize=(18,6))\n",
        "\n",
        "# Loss Curves\n",
        "plt.subplot(1,2,1)\n",
        "plt.plot(model_history.history['loss'], linewidth=3.0)\n",
        "plt.plot(model_history.history['val_loss'], linewidth=3.0)\n",
        "plt.legend(['Training loss', 'Validation Loss'], fontsize=18)\n",
        "plt.xlabel('Epochs', fontsize=16)\n",
        "plt.ylabel('Loss', fontsize=16)\n",
        "plt.title('Loss Curves', fontsize=16)\n",
        "\n",
        "# Accuracy Curves\n",
        "plt.subplot(1,2,2)\n",
        "plt.plot(model_history.history[\"accuracy\"], linewidth=3.0)\n",
        "plt.plot(model_history.history['val_accuracy'], linewidth=3.0)\n",
        "plt.legend(['Training Accuracy', 'Validation Accuracy'], fontsize=18)\n",
        "plt.xlabel('Epochs', fontsize=16)\n",
        "plt.ylabel('Accuracy', fontsize=16)\n",
        "plt.title('Accuracy Curves', fontsize=16)\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DAKa2arYGe2L"
      },
      "outputs": [],
      "source": [
        "from keras.models import load_model\n",
        "\n",
        "model.save('MNIST_model_L6N.h5')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 885
        },
        "id": "XaMTTaVv6IB-",
        "outputId": "960e37f6-cfaa-4aa8-8495-c2ac4897f8f4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(11000,)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "from sklearn.metrics import ConfusionMatrixDisplay\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "y_pred = model.predict(testtensor)\n",
        "print(y_test.shape)\n",
        "y_pred_plot=np.argmax(y_pred, axis=1)\n",
        "y_test_plot=np.argmax(Y_test, axis=1)\n",
        "\n",
        "cm = confusion_matrix(y_test_plot, y_pred_plot)\n",
        "\n",
        "disp = ConfusionMatrixDisplay(confusion_matrix=cm)\n",
        "#disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=labels)\n",
        "\n",
        "disp.plot(cmap=plt.cm.Blues)\n",
        "plt.show()\n",
        "\n",
        "import seaborn as sns\n",
        "cmn = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "fig, ax = plt.subplots(figsize=(10,10))\n",
        "sns.heatmap(cmn, annot=True, fmt='.2f')\n",
        "plt.ylabel('Actual')\n",
        "plt.xlabel('Predicted')\n",
        "plt.show(block=False)"
      ]
    }
  ],
  "metadata": {
    "anaconda-cloud": {},
    "colab": {
      "collapsed_sections": [],
      "name": "MNIST_CNN.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.5"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}