own_dataset_cnn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "own_dataset_cnn.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Rojan116/NepaliHandWrittenAlphabetRecognizer/blob/master/own_dataset_cnn.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Fa0w7bIx81Sq",
        "colab_type": "code",
        "outputId": "70901e95-af64-4968-cd0d-db3a5cb25da3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        }
      },
      "source": [
        "!git clone https://github.com/Rojan116/NepaliHandWrittenAlphabetRecognizer.git"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'NepaliHandWrittenAlphabetRecognizer'...\n",
            "remote: Enumerating objects: 1827, done.\u001b[K\n",
            "remote: Counting objects: 100% (1827/1827), done.\u001b[K\n",
            "remote: Compressing objects: 100% (1825/1825), done.\u001b[K\n",
            "remote: Total 99358 (delta 2), reused 1826 (delta 2), pack-reused 97531\u001b[K\n",
            "Receiving objects: 100% (99358/99358), 137.61 MiB | 8.46 MiB/s, done.\n",
            "Resolving deltas: 100% (1552/1552), done.\n",
            "Checking out files: 100% (93803/93803), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jHtN-fzK9KOB",
        "colab_type": "code",
        "outputId": "32456252-f17c-4ca3-9b53-0203a3f35517",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        }
      },
      "source": [
        "ls"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "augmented_cnn_36_class.ipynb   \u001b[0m\u001b[01;34mnepalibarna\u001b[0m/\n",
            "augmented_cnn.ipynb            \u001b[01;34mNepaliHandWrittenAlphabetRecognizer\u001b[0m/\n",
            "augmented_cnn_predicted.ipynb  new_cnn_9766_acc.ipynb\n",
            "cnn_98_.ipynb                  new_cnn_98%.ipynb\n",
            "CNN.ipynb                      new_cnn.ipynb\n",
            "custom_model.ipynb             new_cnn_may14.ipynb\n",
            "dev.py                         new_cnn_may15.ipynb\n",
            "\u001b[01;34mdhrcdata\u001b[0m/                      new_cnn_may16.ipynb\n",
            "image_augmentation.ipynb       Untitled.ipynb\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6-Y4SAWO9Q1G",
        "colab_type": "code",
        "outputId": "f81bc4f6-f968-41df-dd40-cec383f7f96f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "cd NepaliHandWrittenAlphabetRecognizer"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/NepaliHandWrittenAlphabetRecognizer/NepaliHandWrittenAlphabetRecognizer\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v_xvSHzeG_RV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "3f19b9b4-5533-4816-fff1-cee3ed3c19ba"
      },
      "source": [
        "ls"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "augmented_cnn_36_class.ipynb   dev.py                    new_cnn.ipynb\n",
            "augmented_cnn.ipynb            \u001b[0m\u001b[01;34mdhrcdata\u001b[0m/                 new_cnn_may14.ipynb\n",
            "augmented_cnn_predicted.ipynb  image_augmentation.ipynb  new_cnn_may15.ipynb\n",
            "cnn_98_.ipynb                  \u001b[01;34mnepalibarna\u001b[0m/              new_cnn_may16.ipynb\n",
            "CNN.ipynb                      new_cnn_9766_acc.ipynb    Untitled.ipynb\n",
            "custom_model.ipynb             new_cnn_98%.ipynb\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AK0RS0g19VaY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import os\n",
        "import tensorflow as tf\n",
        "from tqdm import tqdm\n",
        "import random\n",
        "from keras.utils import to_categorical\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.callbacks import ReduceLROnPlateau\n",
        "\n",
        "from keras.models import Sequential\n",
        "from keras.layers import *\n",
        "from keras.optimizers import *"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XUB5C0nB9YOg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "PATH = os.getcwd()\n",
        "DATADIR = PATH + '/nepalibarna'\t\t\n",
        "train_data_path = PATH + '/nepalibarna'\n",
        "test_data_path = PATH + '/nepalibarna'\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MOK18pvNNPTZ",
        "colab_type": "code",
        "outputId": "05b8b220-c64b-42b7-98cc-2f83d68b920c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "source": [
        "#counting categories\n",
        "new_cat = []\n",
        "for x in os.listdir(train_data_path):\n",
        "  new_cat.append(x)\n",
        "\n",
        "print(new_cat)  \n",
        "classes = np.unique(new_cat)\n",
        "print(len(classes))\n"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['character_22_pha', 'character_14_dha', 'character_33_ha', 'character_31_sa', 'character_17_tha', 'character_15_na', 'character_28_la', 'character_01_ka', 'character_05_n..a', 'character_03_ga', 'character_21_pa', 'character_19_dha', 'character_13_da', 'character_34_chyan', 'character_09_jha', 'character_04_gha', 'character_30_sa', 'character_36_gyan', 'character_07_chha', 'character_20_na', 'character_12_tha', 'character_26_ya', 'character_08_ja', 'character_16_ta', 'character_27_ra', 'character_25_ma', 'character_29_wa', 'character_11_ta', 'character_18_da', 'character_24_bha', 'character_23_ba', 'character_02_kha', 'character_32_sa', 'character_06_cha', 'character_35_tra', 'character_10_yaa']\n",
            "36\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kTFBK9wD9aqu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "CATEGORIES = [\n",
        "\n",
        "\"character_01_ka\",\"character_02_kha\",\"character_03_ga\",\"character_04_gha\",\"character_05_n..a\",\n",
        "\"character_06_cha\",\"character_07_chha\",\"character_08_ja\",\"character_09_jha\",\"character_10_yaa\",\n",
        "\"character_11_ta\",\"character_12_tha\",\"character_13_da\",\"character_14_dha\",\"character_15_na\",\n",
        "\"character_16_ta\",\"character_17_tha\",\"character_18_da\",\"character_19_dha\",\"character_20_na\",\"character_21_pa\",\n",
        "\"character_22_pha\",\"character_23_ba\",\"character_24_bha\",\"character_25_ma\",\"character_26_ya\",\"character_27_ra\",\n",
        "\"character_28_la\",\"character_29_wa\",\"character_30_sa\",\"character_31_sa\",\"character_32_sa\",\n",
        "\"character_33_ha\",\"character_34_chyan\",\"character_35_tra\",\"character_36_gyan\"\n",
        "\n",
        "]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "as1xbqAW9d5P",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "IMG_SIZE = 32\n",
        "training_data = []\n",
        "testing_data = []\n",
        "\n",
        "def create_training_data():\n",
        "  for catagory in CATEGORIES:\n",
        "    path = os.path.join(train_data_path, catagory)\n",
        "    class_num = CATEGORIES.index(catagory)\n",
        "    for img in os.listdir(path):\n",
        "        try:\n",
        "          img_array = cv2.imread(os.path.join(path,img),cv2.IMREAD_GRAYSCALE)\n",
        "          new_array = cv2.resize(img_array,(IMG_SIZE,IMG_SIZE))\n",
        "          training_data.append([new_array,class_num])\n",
        "        except Exception as e:\n",
        "          pass\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "create_training_data()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_7a9Hftp9kSg",
        "colab_type": "code",
        "outputId": "6df558a9-3392-4838-f5a5-9827e968b263",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "#shuffle training data for trainning diff image set in every batch\n",
        "random.shuffle(training_data)\n",
        "\n",
        "#hot encode image dataset\n",
        "\n",
        "train_x = []\n",
        "train_y = []\n",
        "\n",
        "\n",
        "#seperating features and labels from list\n",
        "for features, labels in training_data:\n",
        "\ttrain_x.append(features)\n",
        "\ttrain_y.append(labels)\n",
        "\n",
        "\n",
        "#converting list to array\n",
        "train_y_array = np.array(train_y)\n",
        "\n",
        "#converting image dataset to input 32*32 dataset \n",
        "\n",
        "train_image_array = np.array(train_x).reshape(-1,32,32,1)\n",
        "\n",
        "#counting total number of class in dataset\n",
        "classes = np.unique(train_y_array)\n",
        "print(classes)\n",
        "print(\"total number of class={} \".format(len(classes)))\n",
        "\n"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23\n",
            " 24 25 26 27 28 29 30 31 32 33 34 35]\n",
            "total number of class=36 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DJyBFLIrFs_G",
        "colab_type": "code",
        "outputId": "5fbd5172-df52-41c7-d5e4-21b0126aec9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "# convert from int to float\n",
        "x_train = train_image_array.astype('float32')\n",
        "# define data preparation\n",
        "datagen = ImageDataGenerator(featurewise_center=True,\n",
        "                             featurewise_std_normalization=True,\n",
        "                             rotation_range=15,\n",
        "                             width_shift_range=0.1,\n",
        "                             height_shift_range=0.1,\n",
        "                              zoom_range=0.2\n",
        "                            )\n",
        "# fit parameters from data\n",
        "datagen.fit(x_train)\n",
        "# configure batch size and retrieve one batch of images\n",
        "for X_batch, y_batch in datagen.flow(x_train, train_y, batch_size=9):\n",
        "\t# create a grid of 3x3 images\n",
        "\tfor i in range(0, 9):\n",
        "\t\tplt.subplot(330 + 1 + i)\n",
        "\t\tplt.imshow(X_batch[i].reshape(32, 32), cmap=plt.get_cmap('gray'))\n",
        "\t# show the plot\n",
        "\tplt.show()\n",
        "\tbreak"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OcD8KB1rGHpt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "#one hot enoded class \n",
        "train_y = to_categorical(train_y_array)\n",
        "\n",
        "\n",
        "train_y = np.array(train_y)\n",
        "\n",
        "#all train x, y and test x y are in array form now\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HTyQPCKc9AuW",
        "colab_type": "code",
        "outputId": "9f5a788b-4c9d-4ac2-e945-ce5182f30000",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 887
        }
      },
      "source": [
        "#new model\n",
        "\n",
        "# Set the CNN model \n",
        "# my CNN architechture is In -> [[Conv2D->relu]*2 -> MaxPool2D -> Dropout]*2 -> Flatten -> Dense -> Dropout -> Out\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
        "                 activation ='relu', input_shape = (32,32,1)))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(MaxPool2D(pool_size=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "\n",
        "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(256, activation = \"relu\"))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(36, activation = \"softmax\"))\n",
        "\n",
        "# Define the optimizer\n",
        "#optimizer = RMSprop(lr=0.001, rho=0.9, epsilon=1e-08, decay=0.0)\n",
        "optimizer = Adam(lr=0.0001)\n",
        "\n",
        "# Compile the model\n",
        "model.compile(optimizer = optimizer , loss = \"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "\n",
        "#model.summary()\n",
        "\n",
        "history = model.fit(x=train_image_array,y=train_y,epochs=20,batch_size=100, validation_split=0.3)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Colocations handled automatically by placer.\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.cast instead.\n",
            "Train on 1251 samples, validate on 537 samples\n",
            "Epoch 1/20\n",
            "1251/1251 [==============================] - 6s 5ms/step - loss: 15.1417 - acc: 0.0304 - val_loss: 14.5671 - val_acc: 0.0242\n",
            "Epoch 2/20\n",
            "1251/1251 [==============================] - 0s 164us/step - loss: 14.9515 - acc: 0.0344 - val_loss: 12.4449 - val_acc: 0.0354\n",
            "Epoch 3/20\n",
            "1251/1251 [==============================] - 0s 162us/step - loss: 14.4342 - acc: 0.0408 - val_loss: 8.8655 - val_acc: 0.0745\n",
            "Epoch 4/20\n",
            "1251/1251 [==============================] - 0s 156us/step - loss: 12.6343 - acc: 0.0448 - val_loss: 4.6575 - val_acc: 0.0875\n",
            "Epoch 5/20\n",
            "1251/1251 [==============================] - 0s 158us/step - loss: 6.2508 - acc: 0.0568 - val_loss: 3.5101 - val_acc: 0.0670\n",
            "Epoch 6/20\n",
            "1251/1251 [==============================] - 0s 160us/step - loss: 3.7038 - acc: 0.0400 - val_loss: 3.5342 - val_acc: 0.0708\n",
            "Epoch 7/20\n",
            "1251/1251 [==============================] - 0s 157us/step - loss: 3.5596 - acc: 0.0504 - val_loss: 3.5258 - val_acc: 0.0745\n",
            "Epoch 8/20\n",
            "1251/1251 [==============================] - 0s 166us/step - loss: 3.5287 - acc: 0.0584 - val_loss: 3.4990 - val_acc: 0.0931\n",
            "Epoch 9/20\n",
            "1251/1251 [==============================] - 0s 156us/step - loss: 3.4772 - acc: 0.0703 - val_loss: 3.4506 - val_acc: 0.1285\n",
            "Epoch 10/20\n",
            "1251/1251 [==============================] - 0s 158us/step - loss: 3.4491 - acc: 0.0839 - val_loss: 3.3863 - val_acc: 0.1732\n",
            "Epoch 11/20\n",
            "1251/1251 [==============================] - 0s 158us/step - loss: 3.3993 - acc: 0.0935 - val_loss: 3.3144 - val_acc: 0.2104\n",
            "Epoch 12/20\n",
            "1251/1251 [==============================] - 0s 161us/step - loss: 3.2823 - acc: 0.1047 - val_loss: 3.1926 - val_acc: 0.2533\n",
            "Epoch 13/20\n",
            "1251/1251 [==============================] - 0s 161us/step - loss: 3.2136 - acc: 0.1351 - val_loss: 3.0564 - val_acc: 0.3091\n",
            "Epoch 14/20\n",
            "1251/1251 [==============================] - 0s 169us/step - loss: 3.1430 - acc: 0.1367 - val_loss: 2.9338 - val_acc: 0.3557\n",
            "Epoch 15/20\n",
            "1251/1251 [==============================] - 0s 161us/step - loss: 3.0495 - acc: 0.1735 - val_loss: 2.7583 - val_acc: 0.3911\n",
            "Epoch 16/20\n",
            "1251/1251 [==============================] - 0s 160us/step - loss: 2.9245 - acc: 0.1934 - val_loss: 2.5972 - val_acc: 0.4749\n",
            "Epoch 17/20\n",
            "1251/1251 [==============================] - 0s 160us/step - loss: 2.7549 - acc: 0.2342 - val_loss: 2.3643 - val_acc: 0.5196\n",
            "Epoch 18/20\n",
            "1251/1251 [==============================] - 0s 160us/step - loss: 2.6011 - acc: 0.2742 - val_loss: 2.0671 - val_acc: 0.6629\n",
            "Epoch 19/20\n",
            "1251/1251 [==============================] - 0s 162us/step - loss: 2.3614 - acc: 0.3317 - val_loss: 1.8110 - val_acc: 0.6853\n",
            "Epoch 20/20\n",
            "1251/1251 [==============================] - 0s 158us/step - loss: 2.2062 - acc: 0.3741 - val_loss: 1.5561 - val_acc: 0.7393\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7X2QnE15CTcr",
        "colab_type": "code",
        "outputId": "ab59fee7-0d06-4198-e952-cc5792f1a27b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        }
      },
      "source": [
        "# summarize history for accuracy\n",
        "plt.plot(history.history['acc'])\n",
        "plt.plot(history.history['val_acc'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()\n",
        "# summarize history for loss\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.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": 39,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEBHJcwoCEZE89/8D7Q3OTbFGY64AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nBNlMFqHKzvT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "img = train_x[500].reshape(-1,32,32,1)\n",
        "belongs_to_class = train_y[500]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xtl0LNcwK9Mi",
        "colab_type": "code",
        "outputId": "1f60eec7-395e-46d0-dac7-39db1207f96b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print(img.shape)"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1, 32, 32, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wfkrVqc61jYz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "x = model.predict(img)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nFRTZVJLSXUq",
        "colab_type": "code",
        "outputId": "43a6a3af-41fc-4df3-b468-a8aa64641e50",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        }
      },
      "source": [
        "print(x)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[0.01382816 0.01377816 0.0045843  0.01904106 0.15585573 0.00206233\n",
            "  0.01892174 0.00471483 0.02681716 0.00823716 0.01493545 0.00955976\n",
            "  0.3003277  0.05339213 0.00194533 0.00567723 0.02937835 0.03095538\n",
            "  0.01612666 0.00393626 0.00966468 0.01471926 0.01413449 0.00734408\n",
            "  0.02236115 0.03143315 0.01652023 0.01559718 0.00871305 0.01259963\n",
            "  0.01066462 0.01432278 0.02847598 0.04134345 0.01430448 0.00372703]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nK7scl2HR_ow",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "name = CATEGORIES[np.argmax(x)]\n",
        "bc = CATEGORIES[np.argmax(belongs_to_class)]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8r1OJfbnVpJt",
        "colab_type": "code",
        "outputId": "a57b1582-12a1-4a4c-966e-632090f05e4e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 303
        }
      },
      "source": [
        "%matplotlib inline\n",
        "imgplot = plt.imshow(train_x[500])\n",
        "i=np.argmax(history.history['acc'])\n",
        "acc= history.history['val_acc'][19]\n",
        "acc= acc*100\n",
        "acc = np.trunc(acc)\n",
        "\n",
        "print(\"predicted image is of: {} with accuracy {}\".format(name,acc))\n",
        "print(\"Actual class label is: {}\".format(bc))\n",
        "\n"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "predicted image is of: character_13_da with accuracy 73.0\n",
            "Actual class label is: character_13_da\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxQDvU2Yh6QAAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8S6af1hIWEPL",
        "colab_type": "code",
        "outputId": "30ee56ae-b7c4-4a5b-f04c-9825551cb6ca",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9821350877710937\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dvMNB9Kpqeic",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}