Mobilenet_Model-Sucess-checking_with_imagenet_weights.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import datetime\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.applications.mobilenet import preprocess_input\n",
    "from tensorflow.keras.layers import Concatenate, UpSampling2D, Conv2D, Reshape, Flatten, Dense, GlobalAveragePooling2D\n",
    "from tensorflow.keras.losses import MeanSquaredError\n",
    "from tensorflow.keras.metrics import MeanSquaredError\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.losses import BinaryCrossentropy\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from PIL import Image\n",
    "\n",
    "from tqdm import tqdm\n",
    "\n",
    "from matplotlib import pyplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#loading the final npyfile\n",
    "data = np.load(r'D:\\mobilenetnpy_files_real\\train_set_undooped-pit-tgt-img-b_box_rescaled_tgtis1.npy', \n",
    "               allow_pickle = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>00436515-870c-4b36-a041-de91049b9ab4</td>\n",
       "      <td>1</td>\n",
       "      <td>[[2, 3, 3, 3, 3, 2, 2, 2, 2, 2, 4, 4, 4, 5, 5,...</td>\n",
       "      <td>[57.75, 33.25, 46.59375, 46.59375, 122.9375, 3...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>00704310-78a8-4b38-8475-49f4573b2dbb</td>\n",
       "      <td>1</td>\n",
       "      <td>[[40, 11, 2, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, ...</td>\n",
       "      <td>[70.65625, 126.21875, 35.0, 35.0, 152.03125, 1...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>00aecb01-a116-45a2-956c-08d2fa55433f</td>\n",
       "      <td>1</td>\n",
       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
       "      <td>[63.0, 70.4375, 20.5625, 20.5625, 119.65625, 6...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00c0b293-48e7-4e16-ac76-9269ba535a62</td>\n",
       "      <td>1</td>\n",
       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
       "      <td>[66.9375, 119.0, 36.75, 36.75, 142.1875, 111.7...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>00f08de1-517e-4652-a04f-d1dc9ee48593</td>\n",
       "      <td>1</td>\n",
       "      <td>[[61, 63, 62, 73, 63, 60, 65, 67, 70, 71, 73, ...</td>\n",
       "      <td>[39.59375, 40.25, 45.0625, 45.0625, 124.90625,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6007</th>\n",
       "      <td>c18d1138-ba74-4af5-af21-bdd4d2c96bb5</td>\n",
       "      <td>1</td>\n",
       "      <td>[[7, 11, 10, 10, 12, 9, 9, 9, 9, 8, 8, 9, 10, ...</td>\n",
       "      <td>[141.3125, 84.4375, 30.84375, 30.84375, 64.531...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6008</th>\n",
       "      <td>c1cddf32-b957-4753-acaa-472ab1447e86</td>\n",
       "      <td>1</td>\n",
       "      <td>[[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,...</td>\n",
       "      <td>[58.84375, 91.0, 42.21875, 42.21875, 167.5625,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6009</th>\n",
       "      <td>c1e73a4e-7afe-4ec5-8af6-ce8315d7a2f2</td>\n",
       "      <td>1</td>\n",
       "      <td>[[23, 24, 24, 24, 23, 23, 23, 22, 22, 22, 22, ...</td>\n",
       "      <td>[145.6875, 91.4375, 40.6875, 40.6875, 69.125, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6010</th>\n",
       "      <td>c1ec14ff-f6d7-4b38-b0cb-fe07041cbdc8</td>\n",
       "      <td>1</td>\n",
       "      <td>[[2, 1, 1, 2, 3, 4, 5, 5, 4, 5, 4, 4, 5, 6, 5,...</td>\n",
       "      <td>[133.21875, 101.5, 52.5, 52.5, 40.46875, 65.18...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6011</th>\n",
       "      <td>c1f7889a-9ea9-4acb-b64c-b737c929599a</td>\n",
       "      <td>1</td>\n",
       "      <td>[[91, 128, 112, 120, 122, 111, 101, 89, 66, 57...</td>\n",
       "      <td>[124.6875, 85.96875, 57.09375, 57.09375, 50.96...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6012 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         0  1  \\\n",
       "0     00436515-870c-4b36-a041-de91049b9ab4  1   \n",
       "1     00704310-78a8-4b38-8475-49f4573b2dbb  1   \n",
       "2     00aecb01-a116-45a2-956c-08d2fa55433f  1   \n",
       "3     00c0b293-48e7-4e16-ac76-9269ba535a62  1   \n",
       "4     00f08de1-517e-4652-a04f-d1dc9ee48593  1   \n",
       "...                                    ... ..   \n",
       "6007  c18d1138-ba74-4af5-af21-bdd4d2c96bb5  1   \n",
       "6008  c1cddf32-b957-4753-acaa-472ab1447e86  1   \n",
       "6009  c1e73a4e-7afe-4ec5-8af6-ce8315d7a2f2  1   \n",
       "6010  c1ec14ff-f6d7-4b38-b0cb-fe07041cbdc8  1   \n",
       "6011  c1f7889a-9ea9-4acb-b64c-b737c929599a  1   \n",
       "\n",
       "                                                      2  \\\n",
       "0     [[2, 3, 3, 3, 3, 2, 2, 2, 2, 2, 4, 4, 4, 5, 5,...   \n",
       "1     [[40, 11, 2, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, ...   \n",
       "2     [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   \n",
       "3     [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   \n",
       "4     [[61, 63, 62, 73, 63, 60, 65, 67, 70, 71, 73, ...   \n",
       "...                                                 ...   \n",
       "6007  [[7, 11, 10, 10, 12, 9, 9, 9, 9, 8, 8, 9, 10, ...   \n",
       "6008  [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,...   \n",
       "6009  [[23, 24, 24, 24, 23, 23, 23, 22, 22, 22, 22, ...   \n",
       "6010  [[2, 1, 1, 2, 3, 4, 5, 5, 4, 5, 4, 4, 5, 6, 5,...   \n",
       "6011  [[91, 128, 112, 120, 122, 111, 101, 89, 66, 57...   \n",
       "\n",
       "                                                      3  \n",
       "0     [57.75, 33.25, 46.59375, 46.59375, 122.9375, 3...  \n",
       "1     [70.65625, 126.21875, 35.0, 35.0, 152.03125, 1...  \n",
       "2     [63.0, 70.4375, 20.5625, 20.5625, 119.65625, 6...  \n",
       "3     [66.9375, 119.0, 36.75, 36.75, 142.1875, 111.7...  \n",
       "4     [39.59375, 40.25, 45.0625, 45.0625, 124.90625,...  \n",
       "...                                                 ...  \n",
       "6007  [141.3125, 84.4375, 30.84375, 30.84375, 64.531...  \n",
       "6008  [58.84375, 91.0, 42.21875, 42.21875, 167.5625,...  \n",
       "6009  [145.6875, 91.4375, 40.6875, 40.6875, 69.125, ...  \n",
       "6010  [133.21875, 101.5, 52.5, 52.5, 40.46875, 65.18...  \n",
       "6011  [124.6875, 85.96875, 57.09375, 57.09375, 50.96...  \n",
       "\n",
       "[6012 rows x 4 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "0 is patient id X_train will the 2, which are the images and y will 3 which is the bb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pre-processing the input\n",
    "from tensorflow.keras.applications.mobilenet import preprocess_input\n",
    "\n",
    "df['imgs'] = df[2].map(lambda x: preprocess_input(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df['imgs'].map(lambda x: np.expand_dims(x, axis = 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(224, 224, 1)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-0.6862745  -0.9137255  -0.9843137  ... -0.9843137  -0.9843137\n",
      "  -0.99215686]\n",
      " [-0.7019608  -0.9137255  -0.9843137  ... -0.9843137  -0.9843137\n",
      "  -0.99215686]\n",
      " [-0.69411767 -0.9137255  -0.9843137  ... -0.9843137  -0.9843137\n",
      "  -0.99215686]\n",
      " ...\n",
      " [-0.75686276 -0.88235295 -0.92156863 ... -0.9764706  -0.94509804\n",
      "  -0.90588236]\n",
      " [-0.5764706  -0.7254902  -0.77254903 ... -0.9529412  -0.8901961\n",
      "  -0.827451  ]\n",
      " [-0.3098039  -0.45098037 -0.4980392  ... -0.8745098  -0.8039216\n",
      "  -0.7176471 ]]\n",
      "(224, 224)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x26601d15b80>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot\n",
    "n = 1\n",
    "print(df['imgs'][n])\n",
    "print(df['imgs'][n].shape)\n",
    "pyplot.imshow(df['imgs'][n])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SWEET ! 🍬🍬🍬🍬\n",
    "\n",
    "everything seems good ! ✨✨"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6012"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X1 = np.zeros((int(df.shape[0]),224, 224,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(len(X)):\n",
    "    X1[i] = X[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = np.zeros((int(df.shape[0]), 16))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 57.75   ,  33.25   ,  46.59375,  46.59375, 122.9375 ,  33.25   ,\n",
       "        56.     ,  56.     ,   0.     ,   0.     ,   0.     ,   0.     ,\n",
       "         0.     ,   0.     ,   0.     ,   0.     ])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(len(df)):\n",
    "    for j in range(16):\n",
    "        preds[i][j] = df[3][i][j]\n",
    "        #print(type(y[i][j]))\n",
    "preds[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds_class = np.zeros((int(df.shape[0]), 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1.])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(len(df)):\n",
    "    for j in range(16):\n",
    "        preds_class[i] = df[1][i]\n",
    "        #print(type(y[i][j]))\n",
    "preds_class[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6012, 16)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = tf.constant(preds, shape = (6012,16), dtype = tf.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds_class = tf.constant(preds_class, shape = (6012,1), dtype = tf.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.applications.mobilenet import MobileNet\n",
    "from tensorflow.keras.layers import Concatenate, UpSampling2D, Conv2D, Reshape, Flatten, Dense, GlobalAveragePooling2D, Input\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.losses import MeanSquaredError\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "mob_net = MobileNet(input_shape = (224,224,3), weights = 'imagenet', include_top= False) \n",
    "\n",
    "input_tensor = Input(shape=(224,224,1))\n",
    "x = Conv2D(3,(3,3),padding='same')(input_tensor)    # x has a dimension of (IMG_SIZE,IMG_SIZE,3)\n",
    "out = mob_net (x) \n",
    "\n",
    "model = Model(inputs=input_tensor,outputs=out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = GlobalAveragePooling2D()(model.output)\n",
    "x = Dense(588,  activation = 'relu')(x)\n",
    "x = Dense(294, activation = 'relu')(x)\n",
    "x = Dense(147, activation = 'relu')(x)\n",
    "regr = Dense(16, activation = 'linear')(x)\n",
    "model1 = Model(inputs = model.input, outputs = regr, name = 'regr_op')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = Dense(1, activation = 'sigmoid', name = 'clf_op')(x)\n",
    "modelCLF_REGR = Model(inputs = model1.input, outputs = [regr, clf])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_2 (InputLayer)            [(None, 224, 224, 1) 0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d (Conv2D)                 (None, 224, 224, 3)  30          input_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "mobilenet_1.00_224 (Model)      (None, 7, 7, 1024)   3228864     conv2d[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling2d (Globa (None, 1024)         0           mobilenet_1.00_224[1][0]         \n",
      "__________________________________________________________________________________________________\n",
      "dense (Dense)                   (None, 588)          602700      global_average_pooling2d[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "dense_1 (Dense)                 (None, 294)          173166      dense[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "dense_2 (Dense)                 (None, 147)          43365       dense_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_3 (Dense)                 (None, 16)           2368        dense_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "clf_op (Dense)                  (None, 1)            148         dense_2[0][0]                    \n",
      "==================================================================================================\n",
      "Total params: 4,050,641\n",
      "Trainable params: 4,028,753\n",
      "Non-trainable params: 21,888\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "modelCLF_REGR.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.losses import BinaryCrossentropy\n",
    "from tensorflow.keras.metrics import BinaryAccuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = 'adam'\n",
    "modelCLF_REGR.compile(optimizer = optimizer,\n",
    "                loss = {'dense_3' : MeanSquaredError(), 'clf_op' : BinaryCrossentropy()},  \n",
    "                metrics = {'dense_3' :MeanSquaredError(), 'clf_op' :BinaryAccuracy()})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "import datetime\n",
    "\n",
    "\n",
    "checkpoint = ModelCheckpoint(\"model-{loss:.2f}.h5\", monitor = \"val_loss\", save_best_only = True,\n",
    "                             save_weights_only = True, mode = \"auto\")\n",
    "\n",
    "\n",
    "stop = EarlyStopping(monitor = \"val_dense_3_mean_squared_error\", patience = 5, mode = \"min\", \n",
    "                     restore_best_weights = True)\n",
    "\n",
    "\n",
    "reduce_lr = ReduceLROnPlateau(monitor = \"val_dense_3_mean_squared_error\", factor = 0.2, patience = 2, \n",
    "                              min_lr = 1e-6, mode = \"auto\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "301/301 [==============================] - 39s 128ms/step - loss: 631.9837 - dense_3_loss: 631.7765 - clf_op_loss: 0.2073 - dense_3_mean_squared_error: 631.7509 - clf_op_binary_accuracy: 0.9636 - val_loss: 575.3832 - val_dense_3_loss: 575.3832 - val_clf_op_loss: 2.8523e-07 - val_dense_3_mean_squared_error: 574.6342 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 544.0223 - dense_3_loss: 544.0221 - clf_op_loss: 2.9174e-04 - dense_3_mean_squared_error: 544.0096 - clf_op_binary_accuracy: 1.0000 - val_loss: 599.8358 - val_dense_3_loss: 599.8333 - val_clf_op_loss: 0.0026 - val_dense_3_mean_squared_error: 597.7442 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 518.6971 - dense_3_loss: 518.6967 - clf_op_loss: 4.1295e-04 - dense_3_mean_squared_error: 518.8724 - clf_op_binary_accuracy: 1.0000 - val_loss: 576.0008 - val_dense_3_loss: 576.0007 - val_clf_op_loss: 4.9769e-05 - val_dense_3_mean_squared_error: 574.2127 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 494.5413 - dense_3_loss: 494.5411 - clf_op_loss: 2.1679e-04 - dense_3_mean_squared_error: 494.6591 - clf_op_binary_accuracy: 1.0000 - val_loss: 556.6173 - val_dense_3_loss: 556.6172 - val_clf_op_loss: 1.9410e-05 - val_dense_3_mean_squared_error: 554.9973 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "301/301 [==============================] - 38s 127ms/step - loss: 470.5039 - dense_3_loss: 470.5036 - clf_op_loss: 4.4008e-04 - dense_3_mean_squared_error: 470.4516 - clf_op_binary_accuracy: 1.0000 - val_loss: 525.0692 - val_dense_3_loss: 525.0682 - val_clf_op_loss: 8.4847e-04 - val_dense_3_mean_squared_error: 522.4811 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "301/301 [==============================] - 37s 124ms/step - loss: 432.4716 - dense_3_loss: 432.4714 - clf_op_loss: 1.7489e-04 - dense_3_mean_squared_error: 432.6233 - clf_op_binary_accuracy: 1.0000 - val_loss: 602.0024 - val_dense_3_loss: 602.0020 - val_clf_op_loss: 4.1305e-04 - val_dense_3_mean_squared_error: 599.3011 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "301/301 [==============================] - 38s 127ms/step - loss: 397.4983 - dense_3_loss: 397.4983 - clf_op_loss: 5.0242e-05 - dense_3_mean_squared_error: 397.6411 - clf_op_binary_accuracy: 1.0000 - val_loss: 638.9695 - val_dense_3_loss: 638.9695 - val_clf_op_loss: 1.4054e-05 - val_dense_3_mean_squared_error: 638.0488 - val_clf_op_binary_accuracy: 1.0000 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 296.7630 - dense_3_loss: 296.7630 - clf_op_loss: 7.1767e-06 - dense_3_mean_squared_error: 296.5805 - clf_op_binary_accuracy: 1.0000 - val_loss: 611.1748 - val_dense_3_loss: 611.1747 - val_clf_op_loss: 1.4129e-05 - val_dense_3_mean_squared_error: 608.3192 - val_clf_op_binary_accuracy: 1.0000 - lr: 2.0000e-04\n",
      "Epoch 9/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 248.2602 - dense_3_loss: 248.2602 - clf_op_loss: 1.9555e-06 - dense_3_mean_squared_error: 248.0535 - clf_op_binary_accuracy: 1.0000 - val_loss: 639.1030 - val_dense_3_loss: 639.1030 - val_clf_op_loss: 7.2574e-07 - val_dense_3_mean_squared_error: 635.4824 - val_clf_op_binary_accuracy: 1.0000 - lr: 2.0000e-04\n",
      "Epoch 10/100\n",
      "301/301 [==============================] - 38s 126ms/step - loss: 202.0850 - dense_3_loss: 202.0850 - clf_op_loss: 5.0173e-07 - dense_3_mean_squared_error: 202.1026 - clf_op_binary_accuracy: 1.0000 - val_loss: 643.0269 - val_dense_3_loss: 643.0269 - val_clf_op_loss: 1.0364e-06 - val_dense_3_mean_squared_error: 639.0164 - val_clf_op_binary_accuracy: 1.0000 - lr: 4.0000e-05\n"
     ]
    }
   ],
   "source": [
    "history = modelCLF_REGR.fit(X1, [preds, preds_class],\n",
    "                      validation_split = 0.2,\n",
    "                      batch_size = 16,\n",
    "                      epochs=100, \n",
    "                      callbacks=[checkpoint, reduce_lr, stop],\n",
    "                     )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'dense_3_loss', 'clf_op_loss', 'dense_3_mean_squared_error', 'clf_op_binary_accuracy', 'val_loss', 'val_dense_3_loss', 'val_clf_op_loss', 'val_dense_3_mean_squared_error', 'val_clf_op_binary_accuracy', 'lr'])\n"
     ]
    }
   ],
   "source": [
    "print(history.history.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['dense_3_loss'])\n",
    "plt.plot(history.history['val_dense_3_loss'])\n",
    "plt.title('MSE by Epochs - BBox regression Loss')\n",
    "plt.ylabel('MSE')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['clf_op_loss'])\n",
    "plt.plot(history.history['val_clf_op_loss'])\n",
    "plt.title('Binary Cross Entropy by Epochs')\n",
    "plt.ylabel('BinaryCrEntropy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['clf_op_binary_accuracy'])\n",
    "plt.plot(history.history['val_clf_op_binary_accuracy'])\n",
    "plt.title('Binary Accuracy by Epochs')\n",
    "plt.ylabel('BinaryAccuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train', 'test'], loc='lower right')\n",
    "plt.show()\n"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}