2-product-detection/[v3] Product Detection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow.keras as tfk\n",
    "\n",
    "from tensorflow.keras.preprocessing import image as tfk_image\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator as tfk_ImageDataGenerator\n",
    "\n",
    "from tensorflow.python.keras.applications.efficientnet import EfficientNetB7 as tfk_BaseModel\n",
    "from tensorflow.python.keras.applications.efficientnet import preprocess_input as tfk_preprocess_input\n",
    "\n",
    "import datetime, os\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns; sns.set();\n",
    "from PIL import Image\n",
    "\n",
    "from sklearn.utils import class_weight\n",
    "from sklearn.metrics import confusion_matrix, log_loss, accuracy_score\n",
    "from sklearn.preprocessing import OneHotEncoder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd_train = pd.read_csv('shopee-product-detection-dataset/train.csv').sort_values(by=['filename'])\n",
    "pd_confusing_images = pd.read_csv('shopee-product-detection-dataset/confusing_images__loss_limit=9.csv').sort_values(by=['filename'])\n",
    "\n",
    "pd_train['dir_filename']            = pd_train.apply(lambda x: '{:02d}/{}'.format(x.category, x.filename), axis=1)\n",
    "pd_confusing_images['dir_filename'] = pd_confusing_images.apply(lambda x: '{:02d}/{}'.format(x.category, x.filename), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>filename</th>\n",
       "      <th>category</th>\n",
       "      <th>dir_filename</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>50947</th>\n",
       "      <td>00017f8bf9d8a34dab8f9707a378dbe5.jpg</td>\n",
       "      <td>38</td>\n",
       "      <td>38/00017f8bf9d8a34dab8f9707a378dbe5.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31622</th>\n",
       "      <td>00022314f9c58eb20d248ed239db3a52.jpg</td>\n",
       "      <td>16</td>\n",
       "      <td>16/00022314f9c58eb20d248ed239db3a52.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34234</th>\n",
       "      <td>0002d098348d8358c44032903e117ade.jpg</td>\n",
       "      <td>28</td>\n",
       "      <td>28/0002d098348d8358c44032903e117ade.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12787</th>\n",
       "      <td>000367a6ec58e9d1e4bc340d8d871dab.jpg</td>\n",
       "      <td>34</td>\n",
       "      <td>34/000367a6ec58e9d1e4bc340d8d871dab.jpg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38157</th>\n",
       "      <td>00042ddb52b8b5046c4f2eaee8c948ed.jpg</td>\n",
       "      <td>10</td>\n",
       "      <td>10/00042ddb52b8b5046c4f2eaee8c948ed.jpg</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   filename category  \\\n",
       "50947  00017f8bf9d8a34dab8f9707a378dbe5.jpg       38   \n",
       "31622  00022314f9c58eb20d248ed239db3a52.jpg       16   \n",
       "34234  0002d098348d8358c44032903e117ade.jpg       28   \n",
       "12787  000367a6ec58e9d1e4bc340d8d871dab.jpg       34   \n",
       "38157  00042ddb52b8b5046c4f2eaee8c948ed.jpg       10   \n",
       "\n",
       "                                  dir_filename  \n",
       "50947  38/00017f8bf9d8a34dab8f9707a378dbe5.jpg  \n",
       "31622  16/00022314f9c58eb20d248ed239db3a52.jpg  \n",
       "34234  28/0002d098348d8358c44032903e117ade.jpg  \n",
       "12787  34/000367a6ec58e9d1e4bc340d8d871dab.jpg  \n",
       "38157  10/00042ddb52b8b5046c4f2eaee8c948ed.jpg  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mask = ~pd_train.set_index('dir_filename').index.isin(pd_confusing_images.set_index('dir_filename').index)\n",
    "pd_train_clean = pd_train.loc[mask].copy()\n",
    "pd_train_clean['category'] = pd_train_clean['category'].astype('str')\n",
    "pd_train_clean.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['0', '1', '10', '11', '12', '13', '14', '15', '16', '17', '18',\n",
       "       '19', '2', '20', '21', '22', '23', '24', '25', '26', '27', '28',\n",
       "       '29', '3', '30', '31', '32', '33', '34', '35', '36', '37', '38',\n",
       "       '39', '4', '40', '41', '5', '6', '7', '8', '9'], dtype=object)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes = np.unique(pd_train_clean['category'])\n",
    "classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_shape = (256, 256, 3)\n",
    "# input_shape = (512, 512, 3)\n",
    "batch_size=128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 74790 validated image filenames belonging to 42 classes.\n",
      "Found 8310 validated image filenames belonging to 42 classes.\n"
     ]
    }
   ],
   "source": [
    "train_datagen = tfk_ImageDataGenerator(\n",
    "    preprocessing_function=tfk_preprocess_input,\n",
    "    validation_split=0.1,\n",
    "    rotation_range=45,\n",
    "    shear_range=20,\n",
    "    zoom_range=[0.9, 1.1],\n",
    "    width_shift_range=0.2,\n",
    "    height_shift_range=0.2,\n",
    "    horizontal_flip=True,\n",
    "    vertical_flip=True\n",
    ")\n",
    "\n",
    "train_generator = train_datagen.flow_from_dataframe(\n",
    "    pd_train_clean,\n",
    "    directory='shopee-product-detection-dataset/train/',\n",
    "    x_col='dir_filename',\n",
    "    y_col='category',\n",
    "    target_size=input_shape[:2],\n",
    "    color_mode='rgb',\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical',\n",
    "    shuffle=True,\n",
    "    subset='training'\n",
    ")\n",
    "\n",
    "val_generator = train_datagen.flow_from_dataframe(\n",
    "    pd_train_clean,\n",
    "    directory='shopee-product-detection-dataset/train/',\n",
    "    x_col='dir_filename',\n",
    "    y_col='category',\n",
    "    target_size=input_shape[:2],\n",
    "    color_mode='rgb',\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical',\n",
    "    shuffle=True,\n",
    "    subset='validation'\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "813"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "base_model = tfk_BaseModel(input_shape=input_shape, weights='imagenet', include_top=False)\n",
    "len(base_model.layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in base_model.layers[:-16]:\n",
    "  layer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tfk.Sequential([\n",
    "    base_model,\n",
    "    tfk.layers.GlobalMaxPooling2D(),\n",
    "    tfk.layers.Dense(512, activation='relu'),\n",
    "    #tfk.layers.Dense(256, activation='relu'),\n",
    "    tfk.layers.Dense(128, activation='relu'),\n",
    "    tfk.layers.Dense(42, activation='softmax', name='output_layer'),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "efficientnetb7 (Model)       (None, 8, 8, 2560)        64097687  \n",
      "_________________________________________________________________\n",
      "global_max_pooling2d (Global (None, 2560)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 512)               1311232   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 128)               65664     \n",
      "_________________________________________________________________\n",
      "output_layer (Dense)         (None, 42)                5418      \n",
      "=================================================================\n",
      "Total params: 65,480,001\n",
      "Trainable params: 6,759,754\n",
      "Non-trainable params: 58,720,247\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tfk.utils.plot_model(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer=tfk.optimizers.Adam(learning_rate=1e-3),\n",
    "              loss=tfk.losses.CategoricalCrossentropy(label_smoothing=0.1),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tfk.models.load_model('model_clean_1.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "584 64\n"
     ]
    }
   ],
   "source": [
    "class_weights = class_weight.compute_class_weight('balanced',\n",
    "                                                  np.unique(train_generator.classes),\n",
    "                                                  train_generator.classes)\n",
    "class_weights = {i : class_weights[i] for i in range(len(class_weights))}\n",
    "\n",
    "train_step_size = train_generator.n//train_generator.batch_size\n",
    "val_step_size   = val_generator.n//val_generator.batch_size\n",
    "print(train_step_size, val_step_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(\n",
    "    monitor='val_accuracy',\n",
    "    factor=0.2,\n",
    "    patience=5,\n",
    "    min_lr=1e-7\n",
    ")\n",
    "\n",
    "early_stop = tfk.callbacks.EarlyStopping(\n",
    "    monitor='val_accuracy',\n",
    "    patience=15,\n",
    "    restore_best_weights=True\n",
    ")\n",
    "\n",
    "checkpoint_dir = os.path.join('.model_files')\n",
    "if not os.path.exists(checkpoint_dir):\n",
    "    os.makedirs(checkpoint_dir)\n",
    "checkpoint = tfk.callbacks.ModelCheckpoint(\n",
    "    filepath = checkpoint_dir + '/ckpt/',\n",
    "    save_freq ='epoch',\n",
    "    save_best_only=True\n",
    ")\n",
    "\n",
    "log_dir = os.path.join(\".model_files/logs/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\"))\n",
    "if not os.path.exists(log_dir):\n",
    "    os.makedirs(log_dir)\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(\n",
    "    log_dir=log_dir,\n",
    "    histogram_freq=1,\n",
    "    update_freq=100,\n",
    "    profile_batch = 100000000\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.4189 - accuracy: 0.7679WARNING:tensorflow:From c:\\users\\franz cesista\\appdata\\local\\programs\\python\\python36\\lib\\site-packages\\tensorflow\\python\\ops\\resource_variable_ops.py:1817: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "If using Keras pass *_constraint arguments to layers.\n",
      "INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 635s 3s/step - loss: 1.4189 - accuracy: 0.7679 - val_loss: 1.1747 - val_accuracy: 0.8426 - lr: 0.0010\n",
      "Epoch 10/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.4151 - accuracy: 0.7684INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 662s 3s/step - loss: 1.4151 - accuracy: 0.7684 - val_loss: 1.1579 - val_accuracy: 0.8633 - lr: 0.0010\n",
      "Epoch 11/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.4045 - accuracy: 0.7719INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 732s 4s/step - loss: 1.4045 - accuracy: 0.7719 - val_loss: 1.1540 - val_accuracy: 0.8578 - lr: 0.0010\n",
      "Epoch 12/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.3627 - accuracy: 0.7840INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 686s 3s/step - loss: 1.3627 - accuracy: 0.7840 - val_loss: 1.0951 - val_accuracy: 0.8784 - lr: 0.0010\n",
      "Epoch 13/100\n",
      "200/200 [==============================] - 697s 3s/step - loss: 1.3595 - accuracy: 0.7843 - val_loss: 1.1424 - val_accuracy: 0.8598 - lr: 0.0010\n",
      "Epoch 14/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.3536 - accuracy: 0.7860INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 719s 4s/step - loss: 1.3536 - accuracy: 0.7860 - val_loss: 1.0880 - val_accuracy: 0.8813 - lr: 0.0010\n",
      "Epoch 15/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.3132 - accuracy: 0.8014INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 731s 4s/step - loss: 1.3132 - accuracy: 0.8014 - val_loss: 1.0823 - val_accuracy: 0.8800 - lr: 0.0010\n",
      "Epoch 16/100\n",
      "200/200 [==============================] - 761s 4s/step - loss: 1.3119 - accuracy: 0.8005 - val_loss: 1.0970 - val_accuracy: 0.8723 - lr: 0.0010\n",
      "Epoch 17/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.3207 - accuracy: 0.7969INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 805s 4s/step - loss: 1.3207 - accuracy: 0.7969 - val_loss: 1.0758 - val_accuracy: 0.8852 - lr: 0.0010\n",
      "Epoch 18/100\n",
      "200/200 [==============================] - 717s 4s/step - loss: 1.2826 - accuracy: 0.8075 - val_loss: 1.0817 - val_accuracy: 0.8710 - lr: 0.0010\n",
      "Epoch 19/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.2722 - accuracy: 0.8120INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 819s 4s/step - loss: 1.2722 - accuracy: 0.8120 - val_loss: 1.0363 - val_accuracy: 0.9008 - lr: 0.0010\n",
      "Epoch 20/100\n",
      "200/200 [==============================] - 762s 4s/step - loss: 1.2629 - accuracy: 0.8129 - val_loss: 1.0543 - val_accuracy: 0.8828 - lr: 0.0010\n",
      "Epoch 21/100\n",
      "200/200 [==============================] - 785s 4s/step - loss: 1.2390 - accuracy: 0.8228 - val_loss: 1.0676 - val_accuracy: 0.8808 - lr: 0.0010\n",
      "Epoch 22/100\n",
      "200/200 [==============================] - 783s 4s/step - loss: 1.2409 - accuracy: 0.8228 - val_loss: 1.0488 - val_accuracy: 0.8879 - lr: 0.0010\n",
      "Epoch 23/100\n",
      "200/200 [==============================] - 847s 4s/step - loss: 1.2408 - accuracy: 0.8191 - val_loss: 1.0537 - val_accuracy: 0.8832 - lr: 0.0010\n",
      "Epoch 24/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.2120 - accuracy: 0.8302INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 968s 5s/step - loss: 1.2120 - accuracy: 0.8302 - val_loss: 1.0273 - val_accuracy: 0.8941 - lr: 0.0010\n",
      "Epoch 25/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.1775 - accuracy: 0.8427INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 1039s 5s/step - loss: 1.1775 - accuracy: 0.8427 - val_loss: 1.0131 - val_accuracy: 0.9047 - lr: 2.0000e-04\n",
      "Epoch 26/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.1552 - accuracy: 0.8477INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 978s 5s/step - loss: 1.1552 - accuracy: 0.8477 - val_loss: 0.9846 - val_accuracy: 0.9125 - lr: 2.0000e-04\n",
      "Epoch 27/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.1278 - accuracy: 0.8587INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 1010s 5s/step - loss: 1.1278 - accuracy: 0.8587 - val_loss: 0.9800 - val_accuracy: 0.9169 - lr: 2.0000e-04\n",
      "Epoch 28/100\n",
      "200/200 [==============================] - 834s 4s/step - loss: 1.1266 - accuracy: 0.8578 - val_loss: 0.9849 - val_accuracy: 0.9129 - lr: 2.0000e-04\n",
      "Epoch 29/100\n",
      "200/200 [==============================] - 812s 4s/step - loss: 1.1171 - accuracy: 0.8618 - val_loss: 0.9966 - val_accuracy: 0.9113 - lr: 2.0000e-04\n",
      "Epoch 30/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.1179 - accuracy: 0.8616INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 953s 5s/step - loss: 1.1179 - accuracy: 0.8616 - val_loss: 0.9763 - val_accuracy: 0.9165 - lr: 2.0000e-04\n",
      "Epoch 31/100\n",
      "200/200 [==============================] - 902s 5s/step - loss: 1.1088 - accuracy: 0.8646 - val_loss: 0.9930 - val_accuracy: 0.9059 - lr: 2.0000e-04\n",
      "Epoch 32/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.1130 - accuracy: 0.8635INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 982s 5s/step - loss: 1.1130 - accuracy: 0.8635 - val_loss: 0.9578 - val_accuracy: 0.9289 - lr: 2.0000e-04\n",
      "Epoch 33/100\n",
      "200/200 [==============================] - 920s 5s/step - loss: 1.1042 - accuracy: 0.8664 - val_loss: 0.9736 - val_accuracy: 0.9176 - lr: 2.0000e-04\n",
      "Epoch 34/100\n",
      "200/200 [==============================] - 931s 5s/step - loss: 1.0983 - accuracy: 0.8698 - val_loss: 0.9840 - val_accuracy: 0.9164 - lr: 2.0000e-04\n",
      "Epoch 35/100\n",
      "200/200 [==============================] - 1017s 5s/step - loss: 1.0929 - accuracy: 0.8683 - val_loss: 0.9712 - val_accuracy: 0.9227 - lr: 2.0000e-04\n",
      "Epoch 36/100\n",
      "200/200 [==============================] - 932s 5s/step - loss: 1.0891 - accuracy: 0.8745 - val_loss: 0.9685 - val_accuracy: 0.9204 - lr: 2.0000e-04\n",
      "Epoch 37/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.0917 - accuracy: 0.8724INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 1030s 5s/step - loss: 1.0917 - accuracy: 0.8724 - val_loss: 0.9508 - val_accuracy: 0.9238 - lr: 2.0000e-04\n",
      "Epoch 38/100\n",
      "200/200 [==============================] - 1057s 5s/step - loss: 1.0721 - accuracy: 0.8763 - val_loss: 0.9986 - val_accuracy: 0.9027 - lr: 4.0000e-05\n",
      "Epoch 39/100\n",
      "200/200 [==============================] - 1003s 5s/step - loss: 1.0706 - accuracy: 0.8779 - val_loss: 0.9548 - val_accuracy: 0.9310 - lr: 4.0000e-05\n",
      "Epoch 40/100\n",
      "200/200 [==============================] - 900s 5s/step - loss: 1.0653 - accuracy: 0.8804 - val_loss: 0.9687 - val_accuracy: 0.9227 - lr: 4.0000e-05\n",
      "Epoch 41/100\n",
      "200/200 [==============================] - 937s 5s/step - loss: 1.0671 - accuracy: 0.8803 - val_loss: 0.9559 - val_accuracy: 0.9223 - lr: 4.0000e-05\n",
      "Epoch 42/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.0713 - accuracy: 0.8761INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 998s 5s/step - loss: 1.0713 - accuracy: 0.8761 - val_loss: 0.9507 - val_accuracy: 0.9262 - lr: 4.0000e-05\n",
      "Epoch 43/100\n",
      "200/200 [==============================] - 881s 4s/step - loss: 1.0662 - accuracy: 0.8821 - val_loss: 0.9660 - val_accuracy: 0.9204 - lr: 4.0000e-05\n",
      "Epoch 44/100\n",
      "200/200 [==============================] - 907s 5s/step - loss: 1.0589 - accuracy: 0.8840 - val_loss: 0.9941 - val_accuracy: 0.9062 - lr: 4.0000e-05\n",
      "Epoch 45/100\n",
      "200/200 [==============================] - 880s 4s/step - loss: 1.0604 - accuracy: 0.8848 - val_loss: 0.9643 - val_accuracy: 0.9207 - lr: 8.0000e-06\n",
      "Epoch 46/100\n",
      "200/200 [==============================] - 871s 4s/step - loss: 1.0619 - accuracy: 0.8814 - val_loss: 0.9595 - val_accuracy: 0.9200 - lr: 8.0000e-06\n",
      "Epoch 47/100\n",
      "200/200 [==============================] - 982s 5s/step - loss: 1.0651 - accuracy: 0.8819 - val_loss: 0.9667 - val_accuracy: 0.9223 - lr: 8.0000e-06\n",
      "Epoch 48/100\n",
      "200/200 [==============================] - 876s 4s/step - loss: 1.0574 - accuracy: 0.8838 - val_loss: 0.9549 - val_accuracy: 0.9230 - lr: 8.0000e-06\n",
      "Epoch 49/100\n",
      "200/200 [==============================] - 890s 4s/step - loss: 1.0570 - accuracy: 0.8825 - val_loss: 0.9709 - val_accuracy: 0.9165 - lr: 8.0000e-06\n",
      "Epoch 50/100\n",
      "200/200 [==============================] - 977s 5s/step - loss: 1.0559 - accuracy: 0.8852 - val_loss: 0.9632 - val_accuracy: 0.9211 - lr: 1.6000e-06\n",
      "Epoch 51/100\n",
      "200/200 [==============================] - ETA: 0s - loss: 1.0601 - accuracy: 0.8830INFO:tensorflow:Assets written to: .model_files/ckpt/assets\n",
      "200/200 [==============================] - 1105s 6s/step - loss: 1.0601 - accuracy: 0.8830 - val_loss: 0.9412 - val_accuracy: 0.9309 - lr: 1.6000e-06\n",
      "Epoch 52/100\n",
      "200/200 [==============================] - 873s 4s/step - loss: 1.0607 - accuracy: 0.8816 - val_loss: 0.9493 - val_accuracy: 0.9294 - lr: 1.6000e-06\n",
      "Epoch 53/100\n",
      "200/200 [==============================] - 851s 4s/step - loss: 1.0552 - accuracy: 0.8840 - val_loss: 0.9487 - val_accuracy: 0.9227 - lr: 1.6000e-06\n",
      "Epoch 54/100\n",
      "200/200 [==============================] - 845s 4s/step - loss: 1.0596 - accuracy: 0.8832 - val_loss: 0.9666 - val_accuracy: 0.9156 - lr: 1.6000e-06\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_generator,\n",
    "                    initial_epoch=8,\n",
    "                    epochs=100,\n",
    "                    steps_per_epoch=200,\n",
    "                    validation_steps=20,\n",
    "                    validation_data=val_generator,\n",
    "                    class_weight=class_weights,\n",
    "                    callbacks=[reduce_lr,\n",
    "                              early_stop, \n",
    "                              checkpoint, \n",
    "                              tensorboard_callback\n",
    "                              ]\n",
    "                   )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('model_clean_2.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tfk.models.load_model('model_clean_2.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 83100 validated image filenames belonging to 42 classes.\n"
     ]
    }
   ],
   "source": [
    "clean_datagen = tfk_ImageDataGenerator(\n",
    "    preprocessing_function=tfk_preprocess_input,\n",
    ")\n",
    "\n",
    "clean_generator = clean_datagen.flow_from_dataframe(\n",
    "    pd_train_clean,\n",
    "    directory='shopee-product-detection-dataset/train/',\n",
    "    x_col='dir_filename',\n",
    "    y_col='category',\n",
    "    target_size=input_shape[:2],\n",
    "    color_mode='rgb',\n",
    "    batch_size=batch_size,\n",
    "    class_mode='categorical',\n",
    "    shuffle=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step: 10\n",
      "step: 20\n",
      "step: 30\n",
      "step: 40\n",
      "step: 50\n",
      "Wall time: 2min 7s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "y_pred = []\n",
    "y_true = []\n",
    "\n",
    "for i in range(50):\n",
    "    if (i+1) % 10 == 0:\n",
    "        print(\"step:\", i+1)\n",
    "    x, y = next(iter(clean_generator))\n",
    "    y_true.extend(y)\n",
    "    y_pred.extend(model.predict(x))\n",
    "\n",
    "y_pred = np.array(y_pred)\n",
    "y_true = np.array(y_true)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# y_pred_ = y_pred.copy()\n",
    "# \n",
    "# for i, c in enumerate(classes):\n",
    "#     c = int(c)\n",
    "#     y_pred_[:, c] = y_pred[:, i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "matrix = confusion_matrix(y_true.argmax(axis=1), y_pred.argmax(axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x648 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (16,9))\n",
    "sns.heatmap(matrix, linecolor='grey', linewidth=0.1)\n",
    "plt.savefig('confusion_matrix.png', \n",
    "            #transparent=True,\n",
    "            bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "def label_smoother(x, factor=0.1):\n",
    "    xx = x.copy()\n",
    "    xx[x == 1] = 1 - factor\n",
    "    xx[x == 0] = factor / (y_true.shape[1] - 1)\n",
    "    return xx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 1.4597827\n",
      "Accuracy: 0.963125\n"
     ]
    }
   ],
   "source": [
    "loss     = tfk.losses.categorical_crossentropy(y_pred, label_smoother(y_true)).numpy().mean()\n",
    "accuracy = accuracy_score(y_pred.argmax(axis=1), y_true.argmax(axis=1))\n",
    "\n",
    "print(\"Loss:\", loss)\n",
    "print(\"Accuracy:\", accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generate Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>filename</th>\n",
       "      <th>category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>485</th>\n",
       "      <td>0004b03ad7eabfb3989727c461310a84.jpg</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6397</th>\n",
       "      <td>0004d34271d0a0253f8a95d0746bae05.jpg</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2181</th>\n",
       "      <td>000d4ad3d0a372fd6eea2fee6bf738ef.jpg</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2626</th>\n",
       "      <td>0012f125ad07123e69468ae1d1d59de4.jpg</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2433</th>\n",
       "      <td>0015831b249b856e569e883a9dc6d4fa.jpg</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  filename  category\n",
       "485   0004b03ad7eabfb3989727c461310a84.jpg        43\n",
       "6397  0004d34271d0a0253f8a95d0746bae05.jpg        43\n",
       "2181  000d4ad3d0a372fd6eea2fee6bf738ef.jpg        43\n",
       "2626  0012f125ad07123e69468ae1d1d59de4.jpg        43\n",
       "2433  0015831b249b856e569e883a9dc6d4fa.jpg        43"
      ]
     },
     "execution_count": 278,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd_test = pd.read_csv('shopee-product-detection-dataset/test.csv').sort_values(by=['filename'])\n",
    "pd_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 12186 validated image filenames.\n"
     ]
    }
   ],
   "source": [
    "test_datagen = tfk_ImageDataGenerator(\n",
    "    preprocessing_function=tfk_preprocess_input\n",
    ")\n",
    "\n",
    "test_generator = test_datagen.flow_from_dataframe(\n",
    "    pd_test,\n",
    "    directory='shopee-product-detection-dataset/test/',\n",
    "    x_col='filename',\n",
    "    target_size=input_shape[:2],\n",
    "    color_mode='rgb',\n",
    "    batch_size=batch_size,\n",
    "    class_mode=None,\n",
    "    shuffle=False,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "96/96 [==============================] - 165s 2s/step\n"
     ]
    }
   ],
   "source": [
    "y_pred = model.predict(test_generator, verbose=1, steps=test_generator.n//batch_size+1)\n",
    "y_pred = y_pred[:test_generator.n]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>filename</th>\n",
       "      <th>category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>485</th>\n",
       "      <td>0004b03ad7eabfb3989727c461310a84.jpg</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6397</th>\n",
       "      <td>0004d34271d0a0253f8a95d0746bae05.jpg</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2181</th>\n",
       "      <td>000d4ad3d0a372fd6eea2fee6bf738ef.jpg</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2626</th>\n",
       "      <td>0012f125ad07123e69468ae1d1d59de4.jpg</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2433</th>\n",
       "      <td>0015831b249b856e569e883a9dc6d4fa.jpg</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6002</th>\n",
       "      <td>001640e7947935ad28d00f740c3eaefc.jpg</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8633</th>\n",
       "      <td>00196f6ef552b2c4d78365818acdbce2.jpg</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2935</th>\n",
       "      <td>001d40e7718c79ed31750d6c988a90fc.jpg</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12092</th>\n",
       "      <td>001eeb075533bafbfb8e4a6f9e4b8765.jpg</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10161</th>\n",
       "      <td>0022bd7f6b85fe98edd3291d9824ee6c.jpg</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   filename category\n",
       "485    0004b03ad7eabfb3989727c461310a84.jpg       12\n",
       "6397   0004d34271d0a0253f8a95d0746bae05.jpg       16\n",
       "2181   000d4ad3d0a372fd6eea2fee6bf738ef.jpg       40\n",
       "2626   0012f125ad07123e69468ae1d1d59de4.jpg       22\n",
       "2433   0015831b249b856e569e883a9dc6d4fa.jpg       12\n",
       "6002   001640e7947935ad28d00f740c3eaefc.jpg        1\n",
       "8633   00196f6ef552b2c4d78365818acdbce2.jpg       40\n",
       "2935   001d40e7718c79ed31750d6c988a90fc.jpg       30\n",
       "12092  001eeb075533bafbfb8e4a6f9e4b8765.jpg       37\n",
       "10161  0022bd7f6b85fe98edd3291d9824ee6c.jpg        9"
      ]
     },
     "execution_count": 282,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd_submission = pd_test.copy()\n",
    "pd_submission['category'] = list(map(lambda x : classes[x], y_pred.argmax(axis=1)))\n",
    "pd_submission.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd_submission.to_csv('submission_2.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(256, 256, 3)"
      ]
     },
     "execution_count": 297,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image = Image.open('shopee-product-detection-dataset/test/000d4ad3d0a372fd6eea2fee6bf738ef.jpg')\n",
    "#image = Image.open('shopee-product-detection-dataset/train/24/0a2f8630539a6028bb636f1bba990fb1.jpg')\n",
    "image = image.resize(input_shape[:2])\n",
    "image = np.asarray(image)\n",
    "image = tfk_preprocess_input(image)\n",
    "image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x279c06c1828>"
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {},
   "outputs": [],
   "source": [
    "res = model.predict(np.expand_dims(image, axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 300,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.00263054, 0.00547645, 0.01066356, 0.00936576, 0.00568813,\n",
       "        0.01023893, 0.01240693, 0.01074474, 0.002801  , 0.0039913 ,\n",
       "        0.00477088, 0.00920071, 0.00228364, 0.06766966, 0.00348781,\n",
       "        0.00132776, 0.00257775, 0.00269085, 0.00320876, 0.00121288,\n",
       "        0.00842926, 0.0039152 , 0.00244225, 0.00485736, 0.0021809 ,\n",
       "        0.01233215, 0.01175185, 0.01738959, 0.002115  , 0.02983152,\n",
       "        0.00948595, 0.00653848, 0.05547606, 0.01896148, 0.00681085,\n",
       "        0.6060216 , 0.00262597, 0.00854894, 0.0060868 , 0.00229492,\n",
       "        0.00390173, 0.00556413]], dtype=float32)"
      ]
     },
     "execution_count": 300,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 301,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "35"
      ]
     },
     "execution_count": 301,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_pred = res.argmax(axis=-1)[0]\n",
    "class_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'40'"
      ]
     },
     "execution_count": 302,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes[class_pred]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}