Best_Model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Yha2Y0RUEKUN",
    "outputId": "995efa53-0355-47d3-ebdd-5934117104a7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The tensorboard extension is already loaded. To reload it, use:\n",
      "  %reload_ext tensorboard\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "%load_ext tensorboard\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "from tqdm import tqdm\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Conv2D,Dense,Input,BatchNormalization,Flatten,MaxPooling2D,Activation,Dropout,concatenate,Softmax\n",
    "from tensorflow.keras.models import Model,Sequential \n",
    "from tensorflow.keras.callbacks import TensorBoard,ModelCheckpoint,EarlyStopping,ReduceLROnPlateau\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.callbacks import CSVLogger\n",
    "from tensorflow.keras.layers import GlobalAveragePooling2D, Concatenate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "kc4XCOLKEUG0",
    "outputId": "e405a528-aa91-4da1-fc53-caa495d28129"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "BEZtFvTjNS9K"
   },
   "outputs": [],
   "source": [
    "#!tar -xvf /content/drive/MyDrive/Assignment/images.tar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "x-YeCzHMEVwn"
   },
   "outputs": [],
   "source": [
    "#!unzip /content/drive/MyDrive/Assignment/dog-breed-identification.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "xjZAu82aEX4t"
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "function ConnectButton(){\n",
    "    console.log(\"Connect pushed\"); \n",
    "    document.querySelector(\"#top-toolbar > colab-connect-button\").shadowRoot.querySelector(\"#connect\").click() \n",
    "}\n",
    "\n",
    "var colab = setInterval(ConnectButton,60000);\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "YJy4vW2flrff",
    "outputId": "a270358e-dfa3-4bf8-9999-74c721b7c21a"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 120/120 [00:00<00:00, 46599.06it/s]\n"
     ]
    }
   ],
   "source": [
    "rootDir = 'Images/'\n",
    "for i in tqdm(os.listdir(rootDir)):\n",
    "  old_name = rootDir+i\n",
    "  lower_i = i.lower()\n",
    "  new_name = rootDir+lower_i\n",
    "  os.rename(old_name, new_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "L1u0euFeipb3"
   },
   "outputs": [],
   "source": [
    "train_data_dir = \"Images/\"\n",
    "train_datagen = ImageDataGenerator(\n",
    "    rescale= 1./255,\n",
    "    shear_range= 0.2,\n",
    "    zoom_range= 0.2,\n",
    "    horizontal_flip= True,\n",
    "    rotation_range= 20,\n",
    "    width_shift_range= 0.2,\n",
    "    height_shift_range= 0.2,   \n",
    "    validation_split=0.2)\n",
    "\n",
    "valid_datagen = ImageDataGenerator(\n",
    "    rescale= 1./255, \n",
    "    validation_split=0.2,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "7Ey9n3KQiaZh",
    "outputId": "cc1f0e12-b2d8-4494-edb1-7b8a7e31cde1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 16508 images belonging to 120 classes.\n"
     ]
    }
   ],
   "source": [
    "train_generator = train_datagen.flow_from_directory(  \n",
    "    train_data_dir,  \n",
    "    batch_size= 64,  \n",
    "    subset='training',\n",
    "    shuffle= True, \n",
    "    class_mode= 'categorical',\n",
    "    target_size= (331, 331), \n",
    "    color_mode= 'rgb',\n",
    "    seed= 42) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "yybUZIG3j4tl",
    "outputId": "32d55469-f4ed-4a42-9c38-f698bf5bd25d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4072 images belonging to 120 classes.\n"
     ]
    }
   ],
   "source": [
    "valid_generator = valid_datagen.flow_from_directory(\n",
    "    train_data_dir,\n",
    "    batch_size= 64,  \n",
    "    subset='validation',\n",
    "    shuffle= True, \n",
    "    class_mode= 'categorical',\n",
    "    target_size= (331, 331),\n",
    "    color_mode= 'rgb',\n",
    "    seed= 42\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 614
    },
    "id": "1tA-9it6kKUR",
    "outputId": "9cce15c6-fde7-4f89-afc0-d997b43720b9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_3 (InputLayer)           [(None, 331, 331, 3  0           []                               \n",
      "                                )]                                                                \n",
      "                                                                                                  \n",
      " inception_resnet_v2 (Functiona  (None, 9, 9, 1536)  54336736    ['input_3[0][0]']                \n",
      " l)                                                                                               \n",
      "                                                                                                  \n",
      " NASNet (Functional)            (None, 11, 11, 4032  84916818    ['input_3[0][0]']                \n",
      "                                )                                                                 \n",
      "                                                                                                  \n",
      " global_average_pooling2d (Glob  (None, 1536)        0           ['inception_resnet_v2[0][0]']    \n",
      " alAveragePooling2D)                                                                              \n",
      "                                                                                                  \n",
      " global_average_pooling2d_1 (Gl  (None, 4032)        0           ['NASNet[0][0]']                 \n",
      " obalAveragePooling2D)                                                                            \n",
      "                                                                                                  \n",
      " concatenate_4 (Concatenate)    (None, 5568)         0           ['global_average_pooling2d[0][0]'\n",
      "                                                                 , 'global_average_pooling2d_1[0][\n",
      "                                                                 0]']                             \n",
      "                                                                                                  \n",
      " dropout (Dropout)              (None, 5568)         0           ['concatenate_4[0][0]']          \n",
      "                                                                                                  \n",
      " dense (Dense)                  (None, 120)          668280      ['dropout[0][0]']                \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 139,921,834\n",
      "Trainable params: 668,280\n",
      "Non-trainable params: 139,253,554\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "IMG_HEIGHT = 331\n",
    "IMG_WIDTH = 331\n",
    "\n",
    "model_inception_resnet_v2 = tf.keras.applications.inception_resnet_v2.InceptionResNetV2(weights='imagenet', include_top=False, input_shape=(IMG_HEIGHT, IMG_WIDTH,3))\n",
    "\n",
    "model_nasnet = tf.keras.applications.nasnet.NASNetLarge(weights='imagenet', include_top=False, input_shape=(IMG_HEIGHT, IMG_WIDTH,3))\n",
    "\n",
    "model_inception_resnet_v2.trainable = False\n",
    "model_nasnet.trainable = False\n",
    "\n",
    "inputs = Input(shape=(IMG_HEIGHT, IMG_WIDTH, 3))\n",
    "\n",
    "x1 = model_inception_resnet_v2(inputs, training=False)\n",
    "x1 = GlobalAveragePooling2D()(x1)\n",
    "\n",
    "x2 = model_nasnet(inputs, training=False)\n",
    "x2 = GlobalAveragePooling2D()(x2)\n",
    "\n",
    "x = Concatenate()([x1, x2])\n",
    "x = Dropout(.7)(x)\n",
    "outputs = Dense(120, activation='softmax')(x)\n",
    "model = Model(inputs, outputs)\n",
    "\n",
    "display(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "WmpBcsDh3oGZ"
   },
   "outputs": [],
   "source": [
    "indices = train_generator.class_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "a5qxF7yMjl3c"
   },
   "outputs": [],
   "source": [
    "labels = dict()\n",
    "for k,v in indices.items():\n",
    "  labels[k.split('-')[1]] = v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "DG66FAgzkE0T",
    "outputId": "413fc25a-1a8d-4dc4-f522-b18ff6dda79f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['/content/drive/MyDrive/self_case_study/debun/labels.pkl']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import joblib\n",
    "root_dir = '/content/drive/MyDrive/self_case_study/debun/'\n",
    "joblib.dump(labels, root_dir+'labels.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "-ZolBTvxi-ho"
   },
   "outputs": [],
   "source": [
    "optimizer = tf.keras.optimizers.Adam(learning_rate=0.0001)\n",
    "model.compile(loss=\"categorical_crossentropy\", metrics=['accuracy'], optimizer=optimizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "gMyUfKTN18an"
   },
   "outputs": [],
   "source": [
    "earlystopping = tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=7,mode='max')\n",
    "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(monitor='val_accuracy', min_lr=0.0000001,patience=2)\n",
    "\n",
    "root_dir = '/content/drive/MyDrive/self_case_study/debun'\n",
    "from keras.callbacks import CSVLogger\n",
    "\n",
    "csv_logger = CSVLogger(root_dir+\"model_ensemble_history_log_stanford_data.csv\", append=True,separator=\",\")\n",
    "\n",
    "checkpoint = ModelCheckpoint(root_dir+'model_ensemble_weights_stanford_data.h5', save_weights_only=True, save_best_only=True, \\\n",
    "                                       mode='max', monitor='val_accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "u1ZmD7K8kkMS",
    "outputId": "f12f0aa5-e677-4c58-be12-1397cbe331e0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "257/257 [==============================] - 889s 3s/step - loss: 2.8930 - accuracy: 0.5919 - val_loss: 1.1685 - val_accuracy: 0.9340 - lr: 1.0000e-04\n",
      "Epoch 2/50\n",
      "257/257 [==============================] - 844s 3s/step - loss: 0.8370 - accuracy: 0.8942 - val_loss: 0.4199 - val_accuracy: 0.9377 - lr: 1.0000e-04\n",
      "Epoch 3/50\n",
      "257/257 [==============================] - 846s 3s/step - loss: 0.4610 - accuracy: 0.9104 - val_loss: 0.2844 - val_accuracy: 0.9390 - lr: 1.0000e-04\n",
      "Epoch 4/50\n",
      "257/257 [==============================] - 844s 3s/step - loss: 0.3499 - accuracy: 0.9187 - val_loss: 0.2362 - val_accuracy: 0.9415 - lr: 1.0000e-04\n",
      "Epoch 5/50\n",
      "257/257 [==============================] - 843s 3s/step - loss: 0.3110 - accuracy: 0.9172 - val_loss: 0.2144 - val_accuracy: 0.9420 - lr: 1.0000e-04\n",
      "Epoch 6/50\n",
      "257/257 [==============================] - 844s 3s/step - loss: 0.2794 - accuracy: 0.9250 - val_loss: 0.2035 - val_accuracy: 0.9422 - lr: 1.0000e-04\n",
      "Epoch 7/50\n",
      "257/257 [==============================] - 845s 3s/step - loss: 0.2635 - accuracy: 0.9228 - val_loss: 0.1979 - val_accuracy: 0.9425 - lr: 1.0000e-04\n",
      "Epoch 8/50\n",
      "257/257 [==============================] - 845s 3s/step - loss: 0.2478 - accuracy: 0.9254 - val_loss: 0.1932 - val_accuracy: 0.9432 - lr: 1.0000e-04\n",
      "Epoch 9/50\n",
      "257/257 [==============================] - 835s 3s/step - loss: 0.2435 - accuracy: 0.9237 - val_loss: 0.1913 - val_accuracy: 0.9425 - lr: 1.0000e-04\n",
      "Epoch 10/50\n",
      "257/257 [==============================] - 834s 3s/step - loss: 0.2424 - accuracy: 0.9245 - val_loss: 0.1889 - val_accuracy: 0.9432 - lr: 1.0000e-04\n",
      "Epoch 11/50\n",
      "257/257 [==============================] - 834s 3s/step - loss: 0.2288 - accuracy: 0.9284 - val_loss: 0.1876 - val_accuracy: 0.9430 - lr: 1.0000e-05\n",
      "Epoch 12/50\n",
      "257/257 [==============================] - 873s 3s/step - loss: 0.2307 - accuracy: 0.9288 - val_loss: 0.1885 - val_accuracy: 0.9430 - lr: 1.0000e-05\n",
      "Epoch 13/50\n",
      "257/257 [==============================] - 842s 3s/step - loss: 0.2237 - accuracy: 0.9321 - val_loss: 0.1868 - val_accuracy: 0.9435 - lr: 1.0000e-06\n",
      "Epoch 14/50\n",
      "257/257 [==============================] - 835s 3s/step - loss: 0.2231 - accuracy: 0.9313 - val_loss: 0.1878 - val_accuracy: 0.9430 - lr: 1.0000e-06\n",
      "Epoch 15/50\n",
      "257/257 [==============================] - 835s 3s/step - loss: 0.2267 - accuracy: 0.9295 - val_loss: 0.1875 - val_accuracy: 0.9435 - lr: 1.0000e-06\n",
      "Epoch 16/50\n",
      "257/257 [==============================] - 836s 3s/step - loss: 0.2297 - accuracy: 0.9276 - val_loss: 0.1861 - val_accuracy: 0.9432 - lr: 1.0000e-07\n",
      "Epoch 17/50\n",
      "257/257 [==============================] - 836s 3s/step - loss: 0.2251 - accuracy: 0.9310 - val_loss: 0.1879 - val_accuracy: 0.9432 - lr: 1.0000e-07\n",
      "Epoch 18/50\n",
      "257/257 [==============================] - 835s 3s/step - loss: 0.2204 - accuracy: 0.9324 - val_loss: 0.1877 - val_accuracy: 0.9432 - lr: 1.0000e-07\n",
      "Epoch 19/50\n",
      "257/257 [==============================] - 835s 3s/step - loss: 0.2208 - accuracy: 0.9300 - val_loss: 0.1877 - val_accuracy: 0.9430 - lr: 1.0000e-07\n",
      "Epoch 20/50\n",
      "257/257 [==============================] - 834s 3s/step - loss: 0.2296 - accuracy: 0.9297 - val_loss: 0.1881 - val_accuracy: 0.9432 - lr: 1.0000e-07\n"
     ]
    }
   ],
   "source": [
    "STEP_SIZE_TRAIN=train_generator.n//train_generator.batch_size\n",
    "STEP_SIZE_VALID=valid_generator.n//valid_generator.batch_size\n",
    "\n",
    "history = model.fit(train_generator,\n",
    "                      steps_per_epoch=STEP_SIZE_TRAIN,\n",
    "                      validation_data=valid_generator,\n",
    "                      validation_steps=STEP_SIZE_VALID,\n",
    "                      epochs=50, callbacks = [earlystopping, reduce_lr, csv_logger, checkpoint])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "FkvsxG3zxlqz"
   },
   "outputs": [],
   "source": [
    "history = pd.read_csv(root_dir+\"model_ensemble_history_log_stanford_data.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "QbHyrAXKxt4Z"
   },
   "outputs": [],
   "source": [
    "def plot_performance_metric(history):\n",
    "  # Plot training & validation iou_score values\n",
    "  plt.figure(figsize=(30, 5))\n",
    "  plt.subplot(121)\n",
    "  plt.plot(history['accuracy'])\n",
    "  plt.plot(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",
    "\n",
    "  # Plot training & validation loss values\n",
    "  plt.subplot(122)\n",
    "  plt.plot(history['loss'])\n",
    "  plt.plot(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": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 138
    },
    "id": "BM7lofdxxl0j",
    "outputId": "a8eb96d5-a749-48d1-be4d-59ae678de1f3"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_performance_metric(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "7_6ahJRs4Rn5"
   },
   "outputs": [],
   "source": [
    "breed = []\n",
    "for i in indices.keys():\n",
    "  breed.append(i.split('-')[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iZFLoZ-_yKfg"
   },
   "outputs": [],
   "source": [
    "#!unzip /content/drive/MyDrive/Assignment/dog-breed-identification.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ZuSku6OVx4hj"
   },
   "outputs": [],
   "source": [
    "root_dir = '/content/drive/MyDrive/self_case_study/debun'\n",
    "filepath=root_dir+'model_ensemble_weights_stanford_data.h5'\n",
    "model.load_weights(filepath)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "7OcW9BFnnHcu"
   },
   "outputs": [],
   "source": [
    "model.save(root_dir+'/model_ensemble_weights_stanford_data.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "b1kXmyBR2zgx"
   },
   "outputs": [],
   "source": [
    "root_dir = '/content/drive/MyDrive/self_case_study/debun/'\n",
    "model = tf.keras.models.load_model(root_dir+'model_ensemble_weights_stanford_data.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "lB5Ml0UEx4mk",
    "outputId": "306e44e8-e79e-4b54-8c0e-2a021a753d44"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10357/10357 [00:54<00:00, 188.73it/s]\n"
     ]
    }
   ],
   "source": [
    "import shutil\n",
    "os.makedirs(\"root/data/test/\")\n",
    "for i in tqdm(range(len(os.listdir('test')))):\n",
    "    dir = os.listdir('test')[i]\n",
    "    source = \"test/\"+dir\n",
    "\n",
    "    # Destination path\n",
    "    destination = \"root/data/test/\"\n",
    "    shutil.copy(source, destination)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Ad18BmwNx4rZ",
    "outputId": "2a125753-ce75-48ca-c8ff-101afd3ae814"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 10357 images belonging to 1 classes.\n"
     ]
    }
   ],
   "source": [
    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "test_data = test_datagen.flow_from_directory(\n",
    "        'root/data/',\n",
    "        target_size=(331, 331),\n",
    "        batch_size=64,\n",
    "        class_mode=None,\n",
    "        shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "id": "7VRBC6QAPDHl"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "a = os.listdir(\"Images\")[0]\n",
    "os.makedirs('test_root/test')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "tKzpy2xJx933"
   },
   "outputs": [],
   "source": [
    "prediction = model.predict(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "m63fs5p4x99B"
   },
   "outputs": [],
   "source": [
    "id = []\n",
    "for i in sorted(os.listdir('test')):\n",
    "  id.append(i.split('.')[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "Era2KZESpqhF",
    "outputId": "ac349a34-9a88-4475-9030-c39f8ab4281b"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-d48a07c6-c72d-4e17-9f08-dc6d75db41ac\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>breed</th>\n",
       "      <th>path</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>000bec180eb18c7604dcecc8fe0dba07</td>\n",
       "      <td>boston_bull</td>\n",
       "      <td>000bec180eb18c7604dcecc8fe0dba07.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>001513dfcb2ffafc82cccf4d8bbaba97</td>\n",
       "      <td>dingo</td>\n",
       "      <td>001513dfcb2ffafc82cccf4d8bbaba97.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>001cdf01b096e06d78e9e5112d419397</td>\n",
       "      <td>pekinese</td>\n",
       "      <td>001cdf01b096e06d78e9e5112d419397.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00214f311d5d2247d5dfe4fe24b2303d</td>\n",
       "      <td>bluetick</td>\n",
       "      <td>00214f311d5d2247d5dfe4fe24b2303d.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0021f9ceb3235effd7fcde7f7538ed62</td>\n",
       "      <td>golden_retriever</td>\n",
       "      <td>0021f9ceb3235effd7fcde7f7538ed62.jpg</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d48a07c6-c72d-4e17-9f08-dc6d75db41ac')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-d48a07c6-c72d-4e17-9f08-dc6d75db41ac button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-d48a07c6-c72d-4e17-9f08-dc6d75db41ac');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                                 id             breed  \\\n",
       "0  000bec180eb18c7604dcecc8fe0dba07       boston_bull   \n",
       "1  001513dfcb2ffafc82cccf4d8bbaba97             dingo   \n",
       "2  001cdf01b096e06d78e9e5112d419397          pekinese   \n",
       "3  00214f311d5d2247d5dfe4fe24b2303d          bluetick   \n",
       "4  0021f9ceb3235effd7fcde7f7538ed62  golden_retriever   \n",
       "\n",
       "                                   path  \n",
       "0  000bec180eb18c7604dcecc8fe0dba07.jpg  \n",
       "1  001513dfcb2ffafc82cccf4d8bbaba97.jpg  \n",
       "2  001cdf01b096e06d78e9e5112d419397.jpg  \n",
       "3  00214f311d5d2247d5dfe4fe24b2303d.jpg  \n",
       "4  0021f9ceb3235effd7fcde7f7538ed62.jpg  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(\"labels.csv\")\n",
    "data['path'] = data['id'].values+'.jpg'\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Y70SmA8k6W_A",
    "outputId": "bafff962-753f-430f-a351-ec51b18b0818"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10357/10357 [00:19<00:00, 543.00it/s]\n"
     ]
    }
   ],
   "source": [
    "df = pd.DataFrame(columns=breed)\n",
    "for i in tqdm(range(10357)):\n",
    "  row = []\n",
    "  #row.append(id[i])\n",
    "  for x in prediction[i]:\n",
    "    row.append(x)\n",
    "  df.loc[len(df.index)] = row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "aO4sg4zQ91Jl"
   },
   "outputs": [],
   "source": [
    "submission = pd.DataFrame(columns=['id'])\n",
    "submission['id'] = id\n",
    "for i in sorted(np.unique(data['breed'])):\n",
    "  submission[i] = df[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "iH3J67kix-CY",
    "outputId": "1939f1dc-235e-43af-f301-f3404813cebd"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 10357/10357 [01:49<00:00, 94.71it/s]\n"
     ]
    }
   ],
   "source": [
    "submission = pd.DataFrame(columns=['id']+sorted(np.unique(data['breed'])))\n",
    "\n",
    "for i in tqdm(range(10357)):\n",
    "  row = []\n",
    "  row.append(id[i])\n",
    "  \n",
    "  for x in prediction[i]:\n",
    "    row.append(x)\n",
    "  submission.loc[len(submission.index)] = row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Vh0u4ukpyCs0"
   },
   "outputs": [],
   "source": [
    "submission.to_csv('model_ensemble_2.4_stanford_data.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "Iqwxot1l3SxY"
   },
   "outputs": [],
   "source": [
    "os.makedirs('root/test')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "0mai1K4l3MCf",
    "outputId": "725003bd-f591-43f0-ed20-dba47fb25623"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.google.colaboratory.intrinsic+json": {
       "type": "string"
      },
      "text/plain": [
       "'root/test/test_img.jpg'"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import shutil\n",
    "source = '/content/4366ae84914916e5e60738e0d1d3846e.jpg'\n",
    "shutil.move(source,'root/test/test_img.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "id": "aoL6f6GZ6Xfr",
    "outputId": "c19a92c8-9212-415e-db12-2b80edce5f81"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1 images belonging to 1 classes.\n"
     ]
    },
    {
     "data": {
      "application/vnd.google.colaboratory.intrinsic+json": {
       "type": "string"
      },
      "text/plain": [
       "\"Given Dog's breed is airedale\""
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import joblib\n",
    "def function1(image_dir):\n",
    "          test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "          test_data = test_datagen.flow_from_directory(\n",
    "                  image_dir,\n",
    "                  target_size=(331, 331),\n",
    "                  batch_size=64,\n",
    "                  class_mode=None,\n",
    "                  shuffle=False)\n",
    "          labels = joblib.load('/content/drive/MyDrive/self_case_study/debun/labels.pkl')\n",
    "          #model = joblib.load('C://Users//Debasish Pradhan//.spyder-py3//Deployment//pkl_files//debunfinal_model.pkl')\n",
    "          model = tf.keras.models.load_model('/content/drive/MyDrive/self_case_study/debun/model_ensemble_weights_stanford_data.h5')\n",
    "\n",
    "          prediction = model.predict(test_data)\n",
    "          predicted_class = np.argmax(prediction[0])\n",
    "          #print('hhhhhhhhhhh',predicted_class)\n",
    "          output = [key for key,value in labels.items() if value==predicted_class][0]\n",
    "          return f\"Given Dog's breed is {output}\"\n",
    "      \n",
    "output = function1('root')\n",
    "output"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Best Model.ipynb",
   "provenance": []
  },
  "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}