cnn-isl_classification.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "aYN5KViMUIKW"
      },
      "outputs": [],
      "source": [
        "training_ds_path=\"/Users/vaishaalikondapalli/Desktop/Allimgs/Train\"\n",
        "testing_ds_path=\"/Users/vaishaalikondapalli/Desktop/Allimgs/Test\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "Xa6UnMGhUITi"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import keras\n",
        "from keras.layers import *\n",
        "from keras.models import *\n",
        "from keras.preprocessing import image"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "Bu_VZAAZUIWQ"
      },
      "outputs": [],
      "source": [
        "#model definition\n",
        "model = Sequential()\n",
        "model.add(Conv2D(32,kernel_size=(3,3),activation='relu', input_shape=(224,224,3)))\n",
        "model.add(Conv2D(64,(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(64,(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Conv2D(128,(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(64,activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(1,activation='sigmoid'))\n",
        "\n",
        "model.compile(loss=keras.losses.binary_crossentropy, optimizer='adam',metrics=['accuracy'])\n",
        "\n",
        "#used to detect feature in very tiny areas of the image.Its gonna learn a hidden pattern.As you go deeper into the layered network,\n",
        "#the receptive feild of the CNN increases. So, the feature it extracts are quite a bigger part of the original picture.\n",
        "#(3x3) is the standard kernel size picked"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yhsnYvquUIYr",
        "outputId": "b2f1fabb-ca82-42c1-fcec-5c9a7cc46749"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_4 (Conv2D)           (None, 222, 222, 32)      896       \n",
            "                                                                 \n",
            " conv2d_5 (Conv2D)           (None, 220, 220, 64)      18496     \n",
            "                                                                 \n",
            " max_pooling2d_3 (MaxPoolin  (None, 110, 110, 64)      0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_4 (Dropout)         (None, 110, 110, 64)      0         \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 108, 108, 64)      36928     \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPoolin  (None, 54, 54, 64)        0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_5 (Dropout)         (None, 54, 54, 64)        0         \n",
            "                                                                 \n",
            " conv2d_7 (Conv2D)           (None, 52, 52, 128)       73856     \n",
            "                                                                 \n",
            " max_pooling2d_5 (MaxPoolin  (None, 26, 26, 128)       0         \n",
            " g2D)                                                            \n",
            "                                                                 \n",
            " dropout_6 (Dropout)         (None, 26, 26, 128)       0         \n",
            "                                                                 \n",
            " flatten_1 (Flatten)         (None, 86528)             0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 64)                5537856   \n",
            "                                                                 \n",
            " dropout_7 (Dropout)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 1)                 65        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 5668097 (21.62 MB)\n",
            "Trainable params: 5668097 (21.62 MB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "r7TyLzYSUIbd"
      },
      "outputs": [],
      "source": [
        "#Train from scratch\n",
        "#so we'll use the keras image data generator library to make the data ready for the mode\n",
        "\n",
        "train_datagen = image.ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True,\n",
        "    rotation_range=20,       # Rotate the image by 20 degrees\n",
        "    width_shift_range=0.2,   # Shift the image horizontally\n",
        "    height_shift_range=0.2,  # Shift the image vertically\n",
        "    brightness_range=[0.8,1.2], # Adjust brightness\n",
        "    channel_shift_range=0.2,  # Randomly shift color channels\n",
        "    fill_mode='nearest'      # How to fill in new pixels after a transformation\n",
        ")\n",
        "test_dataset=image.ImageDataGenerator(rescale=1./255)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yd6igAYzUIeM",
        "outputId": "d138a05a-481a-4026-affc-5dfac70ae78c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 283 images belonging to 7 classes.\n"
          ]
        }
      ],
      "source": [
        "#creating actual generator function\n",
        "train_generator=train_datagen.flow_from_directory(\n",
        "    '/Users/vaishaalikondapalli/Desktop/Allimgs/Train',\n",
        "    target_size=(224,224),\n",
        "    batch_size=8,\n",
        "    class_mode='categorical'\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bJhxOeQtUIgw",
        "outputId": "0cac12f2-5b0c-427b-c36d-8fedd580b339"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'Hello': 0,\n",
              " 'Help': 1,\n",
              " 'PhoneNumber': 2,\n",
              " 'Please': 3,\n",
              " 'Sorry': 4,\n",
              " 'ThankYou': 5,\n",
              " 'Water': 6}"
            ]
          },
          "execution_count": 24,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "train_generator.class_indices"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u7r8we6SUIjK",
        "outputId": "1b0e284d-9cab-47a5-82fd-2cedebf9d67f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 163 images belonging to 7 classes.\n"
          ]
        }
      ],
      "source": [
        "#creating the validation function\n",
        "validation_generator=test_dataset.flow_from_directory(\n",
        "    '/Users/vaishaalikondapalli/Desktop/Allimgs/Test',\n",
        "    target_size=(224,224),\n",
        "    batch_size=8,\n",
        "    class_mode='categorical'\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZcUb85_iUIlQ",
        "outputId": "966b3232-bce2-4929-be19-444959abc8c6"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/var/folders/13/p_dy58694092f9vhw3lnlmwc0000gn/T/ipykernel_54412/2018266594.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  hist = model.fit_generator(\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "8/8 [==============================] - 10s 1s/step - loss: 0.7942 - accuracy: 0.7902 - val_loss: 0.4326 - val_accuracy: 0.8571\n",
            "Epoch 2/100\n",
            "8/8 [==============================] - 8s 985ms/step - loss: 0.4834 - accuracy: 0.8571 - val_loss: 0.5897 - val_accuracy: 0.8571\n",
            "Epoch 3/100\n",
            "8/8 [==============================] - 8s 991ms/step - loss: 0.4496 - accuracy: 0.8571 - val_loss: 0.4540 - val_accuracy: 0.8571\n",
            "Epoch 4/100\n",
            "8/8 [==============================] - 8s 957ms/step - loss: 0.4373 - accuracy: 0.8571 - val_loss: 0.4749 - val_accuracy: 0.8571\n",
            "Epoch 5/100\n",
            "8/8 [==============================] - 8s 962ms/step - loss: 0.4333 - accuracy: 0.8571 - val_loss: 0.4493 - val_accuracy: 0.8571\n",
            "Epoch 6/100\n",
            "8/8 [==============================] - 7s 920ms/step - loss: 0.4321 - accuracy: 0.8571 - val_loss: 0.4634 - val_accuracy: 0.8571\n",
            "Epoch 7/100\n",
            "8/8 [==============================] - 8s 971ms/step - loss: 0.4288 - accuracy: 0.8571 - val_loss: 0.4518 - val_accuracy: 0.8571\n",
            "Epoch 8/100\n",
            "8/8 [==============================] - 8s 998ms/step - loss: 0.4247 - accuracy: 0.8571 - val_loss: 0.4268 - val_accuracy: 0.8571\n",
            "Epoch 9/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4263 - accuracy: 0.8571 - val_loss: 0.4417 - val_accuracy: 0.8571\n",
            "Epoch 10/100\n",
            "8/8 [==============================] - 8s 983ms/step - loss: 0.4279 - accuracy: 0.8571 - val_loss: 0.4267 - val_accuracy: 0.8571\n",
            "Epoch 11/100\n",
            "8/8 [==============================] - 8s 972ms/step - loss: 0.4291 - accuracy: 0.8571 - val_loss: 0.4248 - val_accuracy: 0.8571\n",
            "Epoch 12/100\n",
            "8/8 [==============================] - 8s 972ms/step - loss: 0.4188 - accuracy: 0.8571 - val_loss: 0.4450 - val_accuracy: 0.8571\n",
            "Epoch 13/100\n",
            "8/8 [==============================] - 8s 974ms/step - loss: 0.4325 - accuracy: 0.8571 - val_loss: 0.4227 - val_accuracy: 0.8571\n",
            "Epoch 14/100\n",
            "8/8 [==============================] - 7s 905ms/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4472 - val_accuracy: 0.8571\n",
            "Epoch 15/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4200 - accuracy: 0.8571 - val_loss: 0.4244 - val_accuracy: 0.8571\n",
            "Epoch 16/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4236 - accuracy: 0.8571 - val_loss: 0.4222 - val_accuracy: 0.8571\n",
            "Epoch 17/100\n",
            "8/8 [==============================] - 8s 945ms/step - loss: 0.4228 - accuracy: 0.8571 - val_loss: 0.4329 - val_accuracy: 0.8571\n",
            "Epoch 18/100\n",
            "8/8 [==============================] - 7s 891ms/step - loss: 0.4292 - accuracy: 0.8571 - val_loss: 0.4347 - val_accuracy: 0.8571\n",
            "Epoch 19/100\n",
            "8/8 [==============================] - 8s 928ms/step - loss: 0.4221 - accuracy: 0.8571 - val_loss: 0.4161 - val_accuracy: 0.8571\n",
            "Epoch 20/100\n",
            "8/8 [==============================] - 8s 962ms/step - loss: 0.4248 - accuracy: 0.8571 - val_loss: 0.4412 - val_accuracy: 0.8571\n",
            "Epoch 21/100\n",
            "8/8 [==============================] - 7s 916ms/step - loss: 0.4312 - accuracy: 0.8571 - val_loss: 0.4116 - val_accuracy: 0.8571\n",
            "Epoch 22/100\n",
            "8/8 [==============================] - 8s 977ms/step - loss: 0.4310 - accuracy: 0.8571 - val_loss: 0.4409 - val_accuracy: 0.8571\n",
            "Epoch 23/100\n",
            "8/8 [==============================] - 8s 995ms/step - loss: 0.4275 - accuracy: 0.8571 - val_loss: 0.4217 - val_accuracy: 0.8571\n",
            "Epoch 24/100\n",
            "8/8 [==============================] - 8s 988ms/step - loss: 0.4304 - accuracy: 0.8571 - val_loss: 0.4174 - val_accuracy: 0.8571\n",
            "Epoch 25/100\n",
            "8/8 [==============================] - 8s 992ms/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4359 - val_accuracy: 0.8571\n",
            "Epoch 26/100\n",
            "8/8 [==============================] - 8s 983ms/step - loss: 0.4256 - accuracy: 0.8571 - val_loss: 0.4105 - val_accuracy: 0.8571\n",
            "Epoch 27/100\n",
            "8/8 [==============================] - 8s 970ms/step - loss: 0.4296 - accuracy: 0.8571 - val_loss: 0.4487 - val_accuracy: 0.8571\n",
            "Epoch 28/100\n",
            "8/8 [==============================] - 8s 975ms/step - loss: 0.4289 - accuracy: 0.8571 - val_loss: 0.4127 - val_accuracy: 0.8571\n",
            "Epoch 29/100\n",
            "8/8 [==============================] - 8s 958ms/step - loss: 0.4222 - accuracy: 0.8571 - val_loss: 0.4375 - val_accuracy: 0.8571\n",
            "Epoch 30/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4211 - accuracy: 0.8571 - val_loss: 0.4129 - val_accuracy: 0.8571\n",
            "Epoch 31/100\n",
            "8/8 [==============================] - 8s 967ms/step - loss: 0.4226 - accuracy: 0.8571 - val_loss: 0.4257 - val_accuracy: 0.8571\n",
            "Epoch 32/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4219 - accuracy: 0.8571 - val_loss: 0.4221 - val_accuracy: 0.8571\n",
            "Epoch 33/100\n",
            "8/8 [==============================] - 8s 951ms/step - loss: 0.4236 - accuracy: 0.8571 - val_loss: 0.4147 - val_accuracy: 0.8571\n",
            "Epoch 34/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4234 - accuracy: 0.8571 - val_loss: 0.4206 - val_accuracy: 0.8571\n",
            "Epoch 35/100\n",
            "8/8 [==============================] - 8s 987ms/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4164 - val_accuracy: 0.8571\n",
            "Epoch 36/100\n",
            "8/8 [==============================] - 8s 968ms/step - loss: 0.4240 - accuracy: 0.8571 - val_loss: 0.4201 - val_accuracy: 0.8571\n",
            "Epoch 37/100\n",
            "8/8 [==============================] - 8s 953ms/step - loss: 0.4212 - accuracy: 0.8571 - val_loss: 0.4183 - val_accuracy: 0.8571\n",
            "Epoch 38/100\n",
            "8/8 [==============================] - 8s 929ms/step - loss: 0.4232 - accuracy: 0.8571 - val_loss: 0.4180 - val_accuracy: 0.8571\n",
            "Epoch 39/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4243 - accuracy: 0.8571 - val_loss: 0.4192 - val_accuracy: 0.8571\n",
            "Epoch 40/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4238 - accuracy: 0.8571 - val_loss: 0.4194 - val_accuracy: 0.8571\n",
            "Epoch 41/100\n",
            "8/8 [==============================] - 8s 980ms/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4173 - val_accuracy: 0.8571\n",
            "Epoch 42/100\n",
            "8/8 [==============================] - 8s 987ms/step - loss: 0.4213 - accuracy: 0.8571 - val_loss: 0.4168 - val_accuracy: 0.8571\n",
            "Epoch 43/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4210 - accuracy: 0.8571 - val_loss: 0.4241 - val_accuracy: 0.8571\n",
            "Epoch 44/100\n",
            "8/8 [==============================] - 8s 956ms/step - loss: 0.4260 - accuracy: 0.8571 - val_loss: 0.4121 - val_accuracy: 0.8571\n",
            "Epoch 45/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4180 - accuracy: 0.8571 - val_loss: 0.4186 - val_accuracy: 0.8571\n",
            "Epoch 46/100\n",
            "8/8 [==============================] - 7s 914ms/step - loss: 0.4248 - accuracy: 0.8571 - val_loss: 0.4150 - val_accuracy: 0.8571\n",
            "Epoch 47/100\n",
            "8/8 [==============================] - 7s 1s/step - loss: 0.4251 - accuracy: 0.8571 - val_loss: 0.4204 - val_accuracy: 0.8571\n",
            "Epoch 48/100\n",
            "8/8 [==============================] - 8s 982ms/step - loss: 0.4244 - accuracy: 0.8571 - val_loss: 0.4145 - val_accuracy: 0.8571\n",
            "Epoch 49/100\n",
            "8/8 [==============================] - 8s 999ms/step - loss: 0.4244 - accuracy: 0.8571 - val_loss: 0.4133 - val_accuracy: 0.8571\n",
            "Epoch 50/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4231 - accuracy: 0.8571 - val_loss: 0.4239 - val_accuracy: 0.8571\n",
            "Epoch 51/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4263 - accuracy: 0.8571 - val_loss: 0.4119 - val_accuracy: 0.8571\n",
            "Epoch 52/100\n",
            "8/8 [==============================] - 8s 972ms/step - loss: 0.4264 - accuracy: 0.8571 - val_loss: 0.4185 - val_accuracy: 0.8571\n",
            "Epoch 53/100\n",
            "8/8 [==============================] - 8s 913ms/step - loss: 0.4196 - accuracy: 0.8571 - val_loss: 0.4135 - val_accuracy: 0.8571\n",
            "Epoch 54/100\n",
            "8/8 [==============================] - 8s 979ms/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4201 - val_accuracy: 0.8571\n",
            "Epoch 55/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4245 - accuracy: 0.8571 - val_loss: 0.4139 - val_accuracy: 0.8571\n",
            "Epoch 56/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4227 - accuracy: 0.8571 - val_loss: 0.4166 - val_accuracy: 0.8571\n",
            "Epoch 57/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4194 - accuracy: 0.8571 - val_loss: 0.4165 - val_accuracy: 0.8571\n",
            "Epoch 58/100\n",
            "8/8 [==============================] - 8s 971ms/step - loss: 0.4214 - accuracy: 0.8571 - val_loss: 0.4111 - val_accuracy: 0.8571\n",
            "Epoch 59/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4258 - accuracy: 0.8571 - val_loss: 0.4170 - val_accuracy: 0.8571\n",
            "Epoch 60/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4206 - accuracy: 0.8571 - val_loss: 0.4107 - val_accuracy: 0.8571\n",
            "Epoch 61/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4270 - accuracy: 0.8571 - val_loss: 0.4323 - val_accuracy: 0.8571\n",
            "Epoch 62/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4254 - accuracy: 0.8571 - val_loss: 0.4104 - val_accuracy: 0.8571\n",
            "Epoch 63/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4268 - accuracy: 0.8571 - val_loss: 0.4201 - val_accuracy: 0.8571\n",
            "Epoch 64/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4228 - accuracy: 0.8571 - val_loss: 0.4102 - val_accuracy: 0.8571\n",
            "Epoch 65/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4228 - accuracy: 0.8571 - val_loss: 0.4149 - val_accuracy: 0.8571\n",
            "Epoch 66/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4202 - accuracy: 0.8571 - val_loss: 0.4150 - val_accuracy: 0.8571\n",
            "Epoch 67/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4232 - accuracy: 0.8571 - val_loss: 0.4107 - val_accuracy: 0.8571\n",
            "Epoch 68/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4214 - accuracy: 0.8571 - val_loss: 0.4177 - val_accuracy: 0.8571\n",
            "Epoch 69/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4194 - accuracy: 0.8571 - val_loss: 0.4110 - val_accuracy: 0.8571\n",
            "Epoch 70/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4187 - accuracy: 0.8571 - val_loss: 0.4121 - val_accuracy: 0.8571\n",
            "Epoch 71/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4209 - accuracy: 0.8571 - val_loss: 0.4148 - val_accuracy: 0.8571\n",
            "Epoch 72/100\n",
            "8/8 [==============================] - 8s 949ms/step - loss: 0.4212 - accuracy: 0.8571 - val_loss: 0.4130 - val_accuracy: 0.8571\n",
            "Epoch 73/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4212 - accuracy: 0.8571 - val_loss: 0.4111 - val_accuracy: 0.8571\n",
            "Epoch 74/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4243 - accuracy: 0.8571 - val_loss: 0.4129 - val_accuracy: 0.8571\n",
            "Epoch 75/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4233 - accuracy: 0.8571 - val_loss: 0.4108 - val_accuracy: 0.8571\n",
            "Epoch 76/100\n",
            "8/8 [==============================] - 8s 940ms/step - loss: 0.4177 - accuracy: 0.8571 - val_loss: 0.4117 - val_accuracy: 0.8571\n",
            "Epoch 77/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4218 - accuracy: 0.8571 - val_loss: 0.4113 - val_accuracy: 0.8571\n",
            "Epoch 78/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4221 - accuracy: 0.8571 - val_loss: 0.4114 - val_accuracy: 0.8571\n",
            "Epoch 79/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4192 - accuracy: 0.8571 - val_loss: 0.4129 - val_accuracy: 0.8571\n",
            "Epoch 80/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4195 - accuracy: 0.8571 - val_loss: 0.4137 - val_accuracy: 0.8571\n",
            "Epoch 81/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4228 - accuracy: 0.8571 - val_loss: 0.4105 - val_accuracy: 0.8571\n",
            "Epoch 82/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4244 - accuracy: 0.8571 - val_loss: 0.4144 - val_accuracy: 0.8571\n",
            "Epoch 83/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4204 - accuracy: 0.8571 - val_loss: 0.4139 - val_accuracy: 0.8571\n",
            "Epoch 84/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4207 - accuracy: 0.8571 - val_loss: 0.4110 - val_accuracy: 0.8571\n",
            "Epoch 85/100\n",
            "8/8 [==============================] - 7s 906ms/step - loss: 0.4212 - accuracy: 0.8571 - val_loss: 0.4147 - val_accuracy: 0.8571\n",
            "Epoch 86/100\n",
            "8/8 [==============================] - 8s 984ms/step - loss: 0.4188 - accuracy: 0.8571 - val_loss: 0.4113 - val_accuracy: 0.8571\n",
            "Epoch 87/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4205 - accuracy: 0.8571 - val_loss: 0.4112 - val_accuracy: 0.8571\n",
            "Epoch 88/100\n",
            "8/8 [==============================] - 9s 1s/step - loss: 0.4201 - accuracy: 0.8571 - val_loss: 0.4117 - val_accuracy: 0.8571\n",
            "Epoch 89/100\n",
            "8/8 [==============================] - 8s 987ms/step - loss: 0.4158 - accuracy: 0.8571 - val_loss: 0.4115 - val_accuracy: 0.8571\n",
            "Epoch 90/100\n",
            "8/8 [==============================] - 8s 972ms/step - loss: 0.4205 - accuracy: 0.8571 - val_loss: 0.4141 - val_accuracy: 0.8571\n",
            "Epoch 91/100\n",
            "8/8 [==============================] - 8s 957ms/step - loss: 0.4206 - accuracy: 0.8571 - val_loss: 0.4103 - val_accuracy: 0.8571\n",
            "Epoch 92/100\n",
            "8/8 [==============================] - 8s 966ms/step - loss: 0.4185 - accuracy: 0.8571 - val_loss: 0.4140 - val_accuracy: 0.8571\n",
            "Epoch 93/100\n",
            "8/8 [==============================] - 8s 987ms/step - loss: 0.4199 - accuracy: 0.8571 - val_loss: 0.4108 - val_accuracy: 0.8571\n",
            "Epoch 94/100\n",
            "8/8 [==============================] - 8s 960ms/step - loss: 0.4295 - accuracy: 0.8571 - val_loss: 0.4135 - val_accuracy: 0.8571\n",
            "Epoch 95/100\n",
            "8/8 [==============================] - 7s 895ms/step - loss: 0.4206 - accuracy: 0.8571 - val_loss: 0.4112 - val_accuracy: 0.8571\n",
            "Epoch 96/100\n",
            "8/8 [==============================] - 8s 998ms/step - loss: 0.4191 - accuracy: 0.8571 - val_loss: 0.4109 - val_accuracy: 0.8571\n",
            "Epoch 97/100\n",
            "8/8 [==============================] - 8s 981ms/step - loss: 0.4226 - accuracy: 0.8571 - val_loss: 0.4162 - val_accuracy: 0.8571\n",
            "Epoch 98/100\n",
            "8/8 [==============================] - 8s 979ms/step - loss: 0.4211 - accuracy: 0.8571 - val_loss: 0.4104 - val_accuracy: 0.8571\n",
            "Epoch 99/100\n",
            "8/8 [==============================] - 8s 965ms/step - loss: 0.4218 - accuracy: 0.8571 - val_loss: 0.4118 - val_accuracy: 0.8571\n",
            "Epoch 100/100\n",
            "8/8 [==============================] - 8s 1s/step - loss: 0.4176 - accuracy: 0.8571 - val_loss: 0.4126 - val_accuracy: 0.8571\n"
          ]
        }
      ],
      "source": [
        "hist = model.fit_generator(\n",
        "    train_generator,\n",
        "    steps_per_epoch=8,\n",
        "    epochs=100,\n",
        "    validation_data = validation_generator,\n",
        "    validation_steps=2\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "fJ2jq2S4Updx"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/var/folders/13/p_dy58694092f9vhw3lnlmwc0000gn/T/ipykernel_54412/326654268.py:1: UserWarning: `Model.evaluate_generator` is deprecated and will be removed in a future version. Please use `Model.evaluate`, which supports generators.\n",
            "  model.evaluate_generator(train_generator)\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "[0.4129108488559723, 0.857142984867096]"
            ]
          },
          "execution_count": 27,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model.evaluate_generator(train_generator)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "yTk1k9nUUpg2"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/var/folders/13/p_dy58694092f9vhw3lnlmwc0000gn/T/ipykernel_54412/3378472596.py:1: UserWarning: `Model.evaluate_generator` is deprecated and will be removed in a future version. Please use `Model.evaluate`, which supports generators.\n",
            "  model.evaluate_generator(validation_generator)\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "[0.41261357069015503, 0.8571427464485168]"
            ]
          },
          "execution_count": 28,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model.evaluate_generator(validation_generator)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "XcC8Z05hUpj8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Assets written to: saved_model/my_model/assets\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Assets written to: saved_model/my_model/assets\n"
          ]
        }
      ],
      "source": [
        "# Save the entire model as a SavedModel.\n",
        "!mkdir -p saved_model\n",
        "model.save('saved_model/my_model')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "KAGW0Z_4VL0Q"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "# accuracy plot\n",
        "plt.plot(hist.history['accuracy'])\n",
        "plt.plot(hist.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "# loss plot\n",
        "plt.plot(hist.history['loss'])\n",
        "plt.plot(hist.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()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 79ms/step\n",
            "Predicted class: Hello\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "from keras.preprocessing import image\n",
        "\n",
        "# Load the image\n",
        "img_path = '/Users/vaishaalikondapalli/Desktop/Allimgs/Test/Hello/Hello.5503eb44-54dc-11ef-8c18-d22f8fcef282.png'  # Update this to your actual image path\n",
        "img = image.load_img(img_path, target_size=(224, 224))\n",
        "img_array = image.img_to_array(img)\n",
        "img_array = np.expand_dims(img_array, axis=0)\n",
        "img_array = img_array * 1./255  # Rescale\n",
        "\n",
        "# Predict the class\n",
        "prediction = model.predict(img_array)\n",
        "predicted_class = np.argmax(prediction, axis=1)\n",
        "\n",
        "# Map the predicted class index to the actual class label\n",
        "class_indices = train_generator.class_indices\n",
        "class_labels = {v: k for k, v in class_indices.items()}\n",
        "predicted_label = class_labels[predicted_class[0]]\n",
        "\n",
        "print(f\"Predicted class: {predicted_label}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 32ms/step\n",
            "Predicted class: Hello\n"
          ]
        }
      ],
      "source": [
        "# Load the image\n",
        "img_path = '/Users/vaishaalikondapalli/Desktop/Allimgs/Test/Water/Water.a8fbde40-5654-11ef-b97a-d22f8fcef282.png'  # Update this to your actual image path\n",
        "img = image.load_img(img_path, target_size=(224, 224))\n",
        "img_array = image.img_to_array(img)\n",
        "img_array = np.expand_dims(img_array, axis=0)\n",
        "img_array = img_array * 1./255  # Rescale\n",
        "\n",
        "# Predict the class\n",
        "prediction = model.predict(img_array)\n",
        "predicted_class = np.argmax(prediction, axis=1)\n",
        "\n",
        "# Map the predicted class index to the actual class label\n",
        "class_indices = train_generator.class_indices\n",
        "class_labels = {v: k for k, v in class_indices.items()}\n",
        "predicted_label = class_labels[predicted_class[0]]\n",
        "\n",
        "print(f\"Predicted class: {predicted_label}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 31ms/step\n",
            "Predicted class: Hello\n"
          ]
        }
      ],
      "source": [
        "# Load the image\n",
        "img_path = '/Users/vaishaalikondapalli/Desktop/Allimgs/Test/ThankYou/ThankYou.886edde0-54dc-11ef-8c18-d22f8fcef282.png'  # Update this to your actual image path\n",
        "img = image.load_img(img_path, target_size=(224, 224))\n",
        "img_array = image.img_to_array(img)\n",
        "img_array = np.expand_dims(img_array, axis=0)\n",
        "img_array = img_array * 1./255  # Rescale\n",
        "\n",
        "# Predict the class\n",
        "prediction = model.predict(img_array)\n",
        "predicted_class = np.argmax(prediction, axis=1)\n",
        "\n",
        "# Map the predicted class index to the actual class label\n",
        "class_indices = train_generator.class_indices\n",
        "class_labels = {v: k for k, v in class_indices.items()}\n",
        "predicted_label = class_labels[predicted_class[0]]\n",
        "\n",
        "print(f\"Predicted class: {predicted_label}\")"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.9.12"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}