blood-cell-classification-nn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import PIL\n",
    "import tensorflow as tf\n",
    "import pathlib\n",
    "import itertools\n",
    "\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.models import Sequential\n",
    "from sklearn.metrics import confusion_matrix, classification_report, plot_confusion_matrix\n",
    "from keras.utils.np_utils import to_categorical\n",
    "from sklearn.preprocessing import OneHotEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9957"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict_characters = {1: 'NEUTROPHIL',\n",
    "                   2: 'EOSINOPHIL', 3: 'MONOCYTE', 4: 'LYMPHOCYTE'}\n",
    "dict_characters2 = {0: 'Mononuclear', 1: 'Polynuclear'}\n",
    "\n",
    "labels = pd.read_csv('../input/blood-cells/dataset2-master/dataset2-master/labels.csv')\n",
    "    \n",
    "train_data_dir = pathlib.Path('../input/blood-cells/dataset2-master/dataset2-master/images/TRAIN')\n",
    "test_data_dir = pathlib.Path('../input/blood-cells/dataset2-master/dataset2-master/images/TEST')\n",
    "image_count = len(list(train_data_dir.glob('*/*.jpeg')))\n",
    "image_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 64\n",
    "img_height = 180\n",
    "img_width = 180"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 9957 files belonging to 4 classes.\n",
      "Using 7966 files for training.\n"
     ]
    }
   ],
   "source": [
    "train_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    train_data_dir,\n",
    "    validation_split=0.2,\n",
    "    subset=\"training\",\n",
    "    seed=2020,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 9957 files belonging to 4 classes.\n",
      "Using 1991 files for validation.\n"
     ]
    }
   ],
   "source": [
    "val_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    train_data_dir,\n",
    "    validation_split=0.2,\n",
    "    subset=\"validation\",\n",
    "    seed=2020,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2487 files belonging to 4 classes.\n"
     ]
    }
   ],
   "source": [
    "test_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
    "    test_data_dir,\n",
    "    seed=2020,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['EOSINOPHIL', 'LYMPHOCYTE', 'MONOCYTE', 'NEUTROPHIL']\n"
     ]
    }
   ],
   "source": [
    "class_names = train_ds.class_names\n",
    "print(test_ds.class_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(64, 180, 180, 3)\n",
      "(64,)\n"
     ]
    }
   ],
   "source": [
    "AUTOTUNE = tf.data.experimental.AUTOTUNE\n",
    "\n",
    "train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)\n",
    "val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "tf.random.set_seed(\n",
    "    2020\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_classes = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "random_flip_1 (RandomFlip)   (None, 180, 180, 3)       0         \n",
      "_________________________________________________________________\n",
      "random_rotation_1 (RandomRot (None, 180, 180, 3)       0         \n",
      "_________________________________________________________________\n",
      "random_zoom_1 (RandomZoom)   (None, 180, 180, 3)       0         \n",
      "_________________________________________________________________\n",
      "rescaling (Rescaling)        (None, 180, 180, 3)       0         \n",
      "_________________________________________________________________\n",
      "conv2d (Conv2D)              (None, 180, 180, 16)      448       \n",
      "_________________________________________________________________\n",
      "max_pooling2d (MaxPooling2D) (None, 90, 90, 16)        0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 90, 90, 16)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 90, 90, 32)        4640      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 45, 45, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 45, 45, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 45, 45, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 22, 22, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 22, 22, 64)        0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 30976)             0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 128)               3965056   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 5)                 645       \n",
      "=================================================================\n",
      "Total params: 3,989,285\n",
      "Trainable params: 3,989,285\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "model = Sequential([\n",
    "  layers.experimental.preprocessing.RandomFlip(\"horizontal\", input_shape=(img_height, img_width, 3)),\n",
    "  layers.experimental.preprocessing.RandomRotation(0.1),\n",
    "  layers.experimental.preprocessing.RandomZoom(0.1),\n",
    "  layers.experimental.preprocessing.Rescaling(1./255),\n",
    "  layers.Conv2D(16, 3, padding='same', activation='relu'),\n",
    "  layers.MaxPooling2D(),\n",
    "  layers.Dropout(0.2),\n",
    "  layers.Conv2D(32, 3, padding='same', activation='relu'),\n",
    "  layers.MaxPooling2D(),\n",
    "  layers.Dropout(0.2),\n",
    "  layers.Conv2D(64, 3, padding='same', activation='relu'),\n",
    "  layers.MaxPooling2D(),\n",
    "  layers.Dropout(0.2),\n",
    "  layers.Flatten(),\n",
    "  layers.Dense(128, activation='relu'),\n",
    "  layers.Dense(num_classes)\n",
    "])\n",
    "# model._name = \"Model\"\n",
    "# model.load_weights('model.h5')\n",
    "model.compile(optimizer='adam',\n",
    "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "              metrics=['accuracy'])\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "125/125 [==============================] - 10s 82ms/step - loss: 1.5670 - accuracy: 0.2509 - val_loss: 1.4027 - val_accuracy: 0.2486\n",
      "Epoch 2/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 1.3910 - accuracy: 0.2535 - val_loss: 1.3954 - val_accuracy: 0.2501\n",
      "Epoch 3/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 1.3877 - accuracy: 0.2665 - val_loss: 1.4058 - val_accuracy: 0.2501\n",
      "Epoch 4/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 1.3521 - accuracy: 0.3192 - val_loss: 1.3434 - val_accuracy: 0.2938\n",
      "Epoch 5/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 1.1930 - accuracy: 0.4435 - val_loss: 1.3427 - val_accuracy: 0.3963\n",
      "Epoch 6/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 1.0611 - accuracy: 0.5132 - val_loss: 1.4083 - val_accuracy: 0.3586\n",
      "Epoch 7/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.9682 - accuracy: 0.5613 - val_loss: 1.2837 - val_accuracy: 0.4319\n",
      "Epoch 8/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.8636 - accuracy: 0.5979 - val_loss: 1.1279 - val_accuracy: 0.4957\n",
      "Epoch 9/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.7386 - accuracy: 0.6542 - val_loss: 0.7329 - val_accuracy: 0.6459\n",
      "Epoch 10/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.6303 - accuracy: 0.7001 - val_loss: 0.8362 - val_accuracy: 0.5987\n",
      "Epoch 11/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.5984 - accuracy: 0.7130 - val_loss: 1.0206 - val_accuracy: 0.5907\n",
      "Epoch 12/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.5480 - accuracy: 0.7363 - val_loss: 0.7722 - val_accuracy: 0.6685\n",
      "Epoch 13/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.4997 - accuracy: 0.7668 - val_loss: 0.7935 - val_accuracy: 0.6303\n",
      "Epoch 14/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.4515 - accuracy: 0.7916 - val_loss: 0.8199 - val_accuracy: 0.6539\n",
      "Epoch 15/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.4213 - accuracy: 0.8133 - val_loss: 0.4825 - val_accuracy: 0.7549\n",
      "Epoch 16/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.3815 - accuracy: 0.8339 - val_loss: 0.4651 - val_accuracy: 0.7830\n",
      "Epoch 17/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.3851 - accuracy: 0.8320 - val_loss: 0.5805 - val_accuracy: 0.7524\n",
      "Epoch 18/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.3476 - accuracy: 0.8505 - val_loss: 0.5721 - val_accuracy: 0.7594\n",
      "Epoch 19/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.2974 - accuracy: 0.8755 - val_loss: 0.3542 - val_accuracy: 0.8488\n",
      "Epoch 20/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.3070 - accuracy: 0.8715 - val_loss: 0.3254 - val_accuracy: 0.8694\n",
      "Epoch 21/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.3101 - accuracy: 0.8667 - val_loss: 0.3098 - val_accuracy: 0.8830\n",
      "Epoch 22/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.2856 - accuracy: 0.8810 - val_loss: 0.2354 - val_accuracy: 0.9076\n",
      "Epoch 23/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.2472 - accuracy: 0.8942 - val_loss: 0.2001 - val_accuracy: 0.9282\n",
      "Epoch 24/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.2497 - accuracy: 0.8913 - val_loss: 0.3456 - val_accuracy: 0.8574\n",
      "Epoch 25/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.2308 - accuracy: 0.9043 - val_loss: 0.9359 - val_accuracy: 0.7273\n",
      "Epoch 26/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.2355 - accuracy: 0.9037 - val_loss: 0.2899 - val_accuracy: 0.8860\n",
      "Epoch 27/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.2310 - accuracy: 0.9055 - val_loss: 0.1854 - val_accuracy: 0.9292\n",
      "Epoch 28/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.2218 - accuracy: 0.9106 - val_loss: 0.2145 - val_accuracy: 0.9166\n",
      "Epoch 29/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1987 - accuracy: 0.9212 - val_loss: 0.3289 - val_accuracy: 0.8795\n",
      "Epoch 30/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1894 - accuracy: 0.9230 - val_loss: 0.1815 - val_accuracy: 0.9332\n",
      "Epoch 31/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1838 - accuracy: 0.9251 - val_loss: 0.2171 - val_accuracy: 0.9116\n",
      "Epoch 32/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1811 - accuracy: 0.9276 - val_loss: 0.3193 - val_accuracy: 0.8734\n",
      "Epoch 33/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1985 - accuracy: 0.9214 - val_loss: 0.1796 - val_accuracy: 0.9277\n",
      "Epoch 34/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1655 - accuracy: 0.9347 - val_loss: 0.2772 - val_accuracy: 0.9061\n",
      "Epoch 35/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1833 - accuracy: 0.9263 - val_loss: 0.3354 - val_accuracy: 0.8594\n",
      "Epoch 36/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1869 - accuracy: 0.9246 - val_loss: 0.1522 - val_accuracy: 0.9397\n",
      "Epoch 37/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1587 - accuracy: 0.9348 - val_loss: 0.2305 - val_accuracy: 0.9051\n",
      "Epoch 38/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1630 - accuracy: 0.9347 - val_loss: 0.1850 - val_accuracy: 0.9287\n",
      "Epoch 39/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1489 - accuracy: 0.9395 - val_loss: 0.1278 - val_accuracy: 0.9468\n",
      "Epoch 40/200\n",
      "125/125 [==============================] - 5s 36ms/step - loss: 0.1407 - accuracy: 0.9401 - val_loss: 0.1246 - val_accuracy: 0.9538\n",
      "Epoch 41/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1359 - accuracy: 0.9443 - val_loss: 0.4434 - val_accuracy: 0.8609\n",
      "Epoch 42/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1511 - accuracy: 0.9431 - val_loss: 0.1400 - val_accuracy: 0.9473\n",
      "Epoch 43/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1203 - accuracy: 0.9517 - val_loss: 0.1664 - val_accuracy: 0.9347\n",
      "Epoch 44/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1293 - accuracy: 0.9484 - val_loss: 0.1268 - val_accuracy: 0.9478\n",
      "Epoch 45/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1349 - accuracy: 0.9473 - val_loss: 0.1687 - val_accuracy: 0.9322\n",
      "Epoch 46/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1305 - accuracy: 0.9494 - val_loss: 0.1189 - val_accuracy: 0.9533\n",
      "Epoch 47/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1336 - accuracy: 0.9470 - val_loss: 0.1126 - val_accuracy: 0.9548\n",
      "Epoch 48/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1039 - accuracy: 0.9600 - val_loss: 0.1001 - val_accuracy: 0.9633\n",
      "Epoch 49/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1082 - accuracy: 0.9586 - val_loss: 0.0985 - val_accuracy: 0.9674\n",
      "Epoch 50/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1039 - accuracy: 0.9600 - val_loss: 0.1102 - val_accuracy: 0.9558\n",
      "Epoch 51/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1314 - accuracy: 0.9493 - val_loss: 0.0847 - val_accuracy: 0.9669\n",
      "Epoch 52/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1053 - accuracy: 0.9587 - val_loss: 0.0890 - val_accuracy: 0.9663\n",
      "Epoch 53/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1020 - accuracy: 0.9602 - val_loss: 0.1313 - val_accuracy: 0.9483\n",
      "Epoch 54/200\n",
      "125/125 [==============================] - 5s 36ms/step - loss: 0.1018 - accuracy: 0.9640 - val_loss: 0.1093 - val_accuracy: 0.9578\n",
      "Epoch 55/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0958 - accuracy: 0.9636 - val_loss: 0.0974 - val_accuracy: 0.9638\n",
      "Epoch 56/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0938 - accuracy: 0.9618 - val_loss: 0.0944 - val_accuracy: 0.9663\n",
      "Epoch 57/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0860 - accuracy: 0.9664 - val_loss: 0.0910 - val_accuracy: 0.9633\n",
      "Epoch 58/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0953 - accuracy: 0.9636 - val_loss: 0.2234 - val_accuracy: 0.9337\n",
      "Epoch 59/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0929 - accuracy: 0.9636 - val_loss: 0.0837 - val_accuracy: 0.9643\n",
      "Epoch 60/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0938 - accuracy: 0.9622 - val_loss: 0.0692 - val_accuracy: 0.9744\n",
      "Epoch 61/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0951 - accuracy: 0.9623 - val_loss: 0.0929 - val_accuracy: 0.9608\n",
      "Epoch 62/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.1251 - accuracy: 0.9524 - val_loss: 0.1359 - val_accuracy: 0.9523\n",
      "Epoch 63/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0839 - accuracy: 0.9692 - val_loss: 0.0608 - val_accuracy: 0.9764\n",
      "Epoch 64/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1112 - accuracy: 0.9573 - val_loss: 0.1039 - val_accuracy: 0.9578\n",
      "Epoch 65/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0954 - accuracy: 0.9643 - val_loss: 0.0706 - val_accuracy: 0.9724\n",
      "Epoch 66/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0858 - accuracy: 0.9657 - val_loss: 0.0742 - val_accuracy: 0.9694\n",
      "Epoch 67/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0762 - accuracy: 0.9714 - val_loss: 0.0693 - val_accuracy: 0.9754\n",
      "Epoch 68/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0854 - accuracy: 0.9677 - val_loss: 0.0940 - val_accuracy: 0.9643\n",
      "Epoch 69/200\n",
      "125/125 [==============================] - 5s 39ms/step - loss: 0.2450 - accuracy: 0.9091 - val_loss: 0.1526 - val_accuracy: 0.9382\n",
      "Epoch 70/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1195 - accuracy: 0.9537 - val_loss: 0.0982 - val_accuracy: 0.9548\n",
      "Epoch 71/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1001 - accuracy: 0.9597 - val_loss: 0.1741 - val_accuracy: 0.9437\n",
      "Epoch 72/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0811 - accuracy: 0.9671 - val_loss: 0.0996 - val_accuracy: 0.9623\n",
      "Epoch 73/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1019 - accuracy: 0.9583 - val_loss: 0.0831 - val_accuracy: 0.9648\n",
      "Epoch 74/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0945 - accuracy: 0.9630 - val_loss: 0.0948 - val_accuracy: 0.9633\n",
      "Epoch 75/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0916 - accuracy: 0.9651 - val_loss: 0.0729 - val_accuracy: 0.9724\n",
      "Epoch 76/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0798 - accuracy: 0.9687 - val_loss: 0.0798 - val_accuracy: 0.9674\n",
      "Epoch 77/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0814 - accuracy: 0.9690 - val_loss: 0.0652 - val_accuracy: 0.9749\n",
      "Epoch 78/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0887 - accuracy: 0.9660 - val_loss: 0.1204 - val_accuracy: 0.9493\n",
      "Epoch 79/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0820 - accuracy: 0.9687 - val_loss: 0.0583 - val_accuracy: 0.9784\n",
      "Epoch 80/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0881 - accuracy: 0.9666 - val_loss: 0.1178 - val_accuracy: 0.9523\n",
      "Epoch 81/200\n",
      "125/125 [==============================] - 5s 37ms/step - loss: 0.0722 - accuracy: 0.9708 - val_loss: 0.1695 - val_accuracy: 0.9322\n",
      "Epoch 82/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0966 - accuracy: 0.9621 - val_loss: 0.0810 - val_accuracy: 0.9679\n",
      "Epoch 83/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0823 - accuracy: 0.9694 - val_loss: 0.0624 - val_accuracy: 0.9724\n",
      "Epoch 84/200\n",
      "125/125 [==============================] - 5s 36ms/step - loss: 0.0738 - accuracy: 0.9721 - val_loss: 0.0596 - val_accuracy: 0.9769\n",
      "Epoch 85/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0627 - accuracy: 0.9770 - val_loss: 0.1078 - val_accuracy: 0.9548\n",
      "Epoch 86/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0602 - accuracy: 0.9779 - val_loss: 0.0985 - val_accuracy: 0.9633\n",
      "Epoch 87/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0580 - accuracy: 0.9794 - val_loss: 0.0702 - val_accuracy: 0.9794\n",
      "Epoch 88/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0812 - accuracy: 0.9714 - val_loss: 0.0877 - val_accuracy: 0.9674\n",
      "Epoch 89/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0824 - accuracy: 0.9689 - val_loss: 0.0837 - val_accuracy: 0.9648\n",
      "Epoch 90/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0612 - accuracy: 0.9763 - val_loss: 0.0629 - val_accuracy: 0.9739\n",
      "Epoch 91/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0637 - accuracy: 0.9763 - val_loss: 0.1097 - val_accuracy: 0.9578\n",
      "Epoch 92/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0612 - accuracy: 0.9760 - val_loss: 0.2319 - val_accuracy: 0.9136\n",
      "Epoch 93/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0797 - accuracy: 0.9716 - val_loss: 0.0576 - val_accuracy: 0.9789\n",
      "Epoch 94/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0518 - accuracy: 0.9810 - val_loss: 0.0437 - val_accuracy: 0.9829\n",
      "Epoch 95/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0610 - accuracy: 0.9770 - val_loss: 0.0946 - val_accuracy: 0.9694\n",
      "Epoch 96/200\n",
      "125/125 [==============================] - 5s 37ms/step - loss: 0.0537 - accuracy: 0.9789 - val_loss: 0.0644 - val_accuracy: 0.9739\n",
      "Epoch 97/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0601 - accuracy: 0.9763 - val_loss: 0.0397 - val_accuracy: 0.9849\n",
      "Epoch 98/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1147 - accuracy: 0.9608 - val_loss: 0.0520 - val_accuracy: 0.9779\n",
      "Epoch 99/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0512 - accuracy: 0.9812 - val_loss: 0.0796 - val_accuracy: 0.9689\n",
      "Epoch 100/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0514 - accuracy: 0.9817 - val_loss: 0.4814 - val_accuracy: 0.9121\n",
      "Epoch 101/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0728 - accuracy: 0.9750 - val_loss: 0.0506 - val_accuracy: 0.9809\n",
      "Epoch 102/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0526 - accuracy: 0.9814 - val_loss: 0.0457 - val_accuracy: 0.9829\n",
      "Epoch 103/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0629 - accuracy: 0.9764 - val_loss: 0.0525 - val_accuracy: 0.9804\n",
      "Epoch 104/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0558 - accuracy: 0.9782 - val_loss: 0.0778 - val_accuracy: 0.9674\n",
      "Epoch 105/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0811 - accuracy: 0.9682 - val_loss: 0.0517 - val_accuracy: 0.9824\n",
      "Epoch 106/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0566 - accuracy: 0.9797 - val_loss: 0.0343 - val_accuracy: 0.9884\n",
      "Epoch 107/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0443 - accuracy: 0.9834 - val_loss: 0.0382 - val_accuracy: 0.9869\n",
      "Epoch 108/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0491 - accuracy: 0.9814 - val_loss: 0.0632 - val_accuracy: 0.9779\n",
      "Epoch 109/200\n",
      "125/125 [==============================] - 5s 36ms/step - loss: 0.0600 - accuracy: 0.9778 - val_loss: 0.0414 - val_accuracy: 0.9834\n",
      "Epoch 110/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0486 - accuracy: 0.9828 - val_loss: 0.0402 - val_accuracy: 0.9890\n",
      "Epoch 111/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0398 - accuracy: 0.9853 - val_loss: 0.0682 - val_accuracy: 0.9784\n",
      "Epoch 112/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0582 - accuracy: 0.9797 - val_loss: 0.4233 - val_accuracy: 0.8518\n",
      "Epoch 113/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0991 - accuracy: 0.9640 - val_loss: 0.0398 - val_accuracy: 0.9874\n",
      "Epoch 114/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0600 - accuracy: 0.9772 - val_loss: 0.0475 - val_accuracy: 0.9824\n",
      "Epoch 115/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0408 - accuracy: 0.9841 - val_loss: 0.0350 - val_accuracy: 0.9869\n",
      "Epoch 116/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0467 - accuracy: 0.9833 - val_loss: 0.0762 - val_accuracy: 0.9699\n",
      "Epoch 117/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0541 - accuracy: 0.9826 - val_loss: 0.0547 - val_accuracy: 0.9824\n",
      "Epoch 118/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0569 - accuracy: 0.9799 - val_loss: 0.0368 - val_accuracy: 0.9879\n",
      "Epoch 119/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0510 - accuracy: 0.9818 - val_loss: 0.0332 - val_accuracy: 0.9869\n",
      "Epoch 120/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0652 - accuracy: 0.9764 - val_loss: 0.0417 - val_accuracy: 0.9839\n",
      "Epoch 121/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0382 - accuracy: 0.9859 - val_loss: 0.0727 - val_accuracy: 0.9774\n",
      "Epoch 122/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0493 - accuracy: 0.9809 - val_loss: 0.0630 - val_accuracy: 0.9784\n",
      "Epoch 123/200\n",
      "125/125 [==============================] - 5s 36ms/step - loss: 0.0393 - accuracy: 0.9861 - val_loss: 0.0735 - val_accuracy: 0.9734\n",
      "Epoch 124/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0392 - accuracy: 0.9846 - val_loss: 0.0435 - val_accuracy: 0.9844\n",
      "Epoch 125/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0368 - accuracy: 0.9852 - val_loss: 0.0516 - val_accuracy: 0.9799\n",
      "Epoch 126/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.1100 - accuracy: 0.9650 - val_loss: 0.0521 - val_accuracy: 0.9829\n",
      "Epoch 127/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0533 - accuracy: 0.9813 - val_loss: 0.0732 - val_accuracy: 0.9729\n",
      "Epoch 128/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0519 - accuracy: 0.9817 - val_loss: 0.0387 - val_accuracy: 0.9879\n",
      "Epoch 129/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0341 - accuracy: 0.9872 - val_loss: 0.0571 - val_accuracy: 0.9804\n",
      "Epoch 130/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.1606 - accuracy: 0.9443 - val_loss: 0.0733 - val_accuracy: 0.9709\n",
      "Epoch 131/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0431 - accuracy: 0.9844 - val_loss: 0.0702 - val_accuracy: 0.9724\n",
      "Epoch 132/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0391 - accuracy: 0.9851 - val_loss: 0.0649 - val_accuracy: 0.9729\n",
      "Epoch 133/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0393 - accuracy: 0.9859 - val_loss: 0.0395 - val_accuracy: 0.9834\n",
      "Epoch 134/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0452 - accuracy: 0.9839 - val_loss: 0.0553 - val_accuracy: 0.9774\n",
      "Epoch 135/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0339 - accuracy: 0.9882 - val_loss: 0.0475 - val_accuracy: 0.9809\n",
      "Epoch 136/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0244 - accuracy: 0.9920 - val_loss: 0.3257 - val_accuracy: 0.9533\n",
      "Epoch 137/200\n",
      "125/125 [==============================] - 5s 42ms/step - loss: 0.0591 - accuracy: 0.9808 - val_loss: 0.0834 - val_accuracy: 0.9694\n",
      "Epoch 138/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0400 - accuracy: 0.9849 - val_loss: 0.1141 - val_accuracy: 0.9578\n",
      "Epoch 139/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0468 - accuracy: 0.9829 - val_loss: 0.0356 - val_accuracy: 0.9864\n",
      "Epoch 140/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0302 - accuracy: 0.9895 - val_loss: 0.0399 - val_accuracy: 0.9859\n",
      "Epoch 141/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0318 - accuracy: 0.9887 - val_loss: 0.0535 - val_accuracy: 0.9809\n",
      "Epoch 142/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0779 - accuracy: 0.9718 - val_loss: 0.0629 - val_accuracy: 0.9804\n",
      "Epoch 143/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0517 - accuracy: 0.9819 - val_loss: 0.0777 - val_accuracy: 0.9714\n",
      "Epoch 144/200\n",
      "125/125 [==============================] - 5s 37ms/step - loss: 0.0291 - accuracy: 0.9898 - val_loss: 0.0473 - val_accuracy: 0.9849\n",
      "Epoch 145/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0311 - accuracy: 0.9888 - val_loss: 0.0369 - val_accuracy: 0.9884\n",
      "Epoch 146/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0345 - accuracy: 0.9867 - val_loss: 0.0843 - val_accuracy: 0.9734\n",
      "Epoch 147/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0870 - accuracy: 0.9685 - val_loss: 0.0396 - val_accuracy: 0.9869\n",
      "Epoch 148/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0371 - accuracy: 0.9867 - val_loss: 0.0393 - val_accuracy: 0.9874\n",
      "Epoch 149/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0301 - accuracy: 0.9892 - val_loss: 0.0275 - val_accuracy: 0.9890\n",
      "Epoch 150/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0552 - accuracy: 0.9820 - val_loss: 0.0361 - val_accuracy: 0.9879\n",
      "Epoch 151/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0362 - accuracy: 0.9868 - val_loss: 0.0515 - val_accuracy: 0.9794\n",
      "Epoch 152/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0272 - accuracy: 0.9897 - val_loss: 0.0679 - val_accuracy: 0.9779\n",
      "Epoch 153/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0270 - accuracy: 0.9886 - val_loss: 0.0299 - val_accuracy: 0.9895\n",
      "Epoch 154/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0645 - accuracy: 0.9778 - val_loss: 0.0444 - val_accuracy: 0.9854\n",
      "Epoch 155/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0283 - accuracy: 0.9905 - val_loss: 0.4674 - val_accuracy: 0.9437\n",
      "Epoch 156/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0377 - accuracy: 0.9872 - val_loss: 0.0339 - val_accuracy: 0.9874\n",
      "Epoch 157/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0297 - accuracy: 0.9893 - val_loss: 0.0359 - val_accuracy: 0.9844\n",
      "Epoch 158/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0265 - accuracy: 0.9915 - val_loss: 0.0949 - val_accuracy: 0.9734\n",
      "Epoch 159/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.1493 - accuracy: 0.9538 - val_loss: 0.0619 - val_accuracy: 0.9784\n",
      "Epoch 160/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0335 - accuracy: 0.9892 - val_loss: 0.0511 - val_accuracy: 0.9834\n",
      "Epoch 161/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0268 - accuracy: 0.9925 - val_loss: 0.0351 - val_accuracy: 0.9895\n",
      "Epoch 162/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0295 - accuracy: 0.9900 - val_loss: 0.0724 - val_accuracy: 0.9744\n",
      "Epoch 163/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0409 - accuracy: 0.9849 - val_loss: 0.0602 - val_accuracy: 0.9769\n",
      "Epoch 164/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0317 - accuracy: 0.9879 - val_loss: 0.0331 - val_accuracy: 0.9884\n",
      "Epoch 165/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0280 - accuracy: 0.9895 - val_loss: 0.0277 - val_accuracy: 0.9890\n",
      "Epoch 166/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0217 - accuracy: 0.9931 - val_loss: 0.0239 - val_accuracy: 0.9910\n",
      "Epoch 167/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0444 - accuracy: 0.9841 - val_loss: 0.0475 - val_accuracy: 0.9829\n",
      "Epoch 168/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0253 - accuracy: 0.9907 - val_loss: 0.0450 - val_accuracy: 0.9839\n",
      "Epoch 169/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0267 - accuracy: 0.9896 - val_loss: 0.0640 - val_accuracy: 0.9769\n",
      "Epoch 170/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0462 - accuracy: 0.9849 - val_loss: 0.0508 - val_accuracy: 0.9839\n",
      "Epoch 171/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0242 - accuracy: 0.9911 - val_loss: 0.0379 - val_accuracy: 0.9874\n",
      "Epoch 172/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0260 - accuracy: 0.9898 - val_loss: 0.0659 - val_accuracy: 0.9779\n",
      "Epoch 173/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0442 - accuracy: 0.9846 - val_loss: 0.0464 - val_accuracy: 0.9834\n",
      "Epoch 174/200\n",
      "125/125 [==============================] - 4s 36ms/step - loss: 0.0302 - accuracy: 0.9881 - val_loss: 0.0428 - val_accuracy: 0.9844\n",
      "Epoch 175/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0239 - accuracy: 0.9905 - val_loss: 0.0314 - val_accuracy: 0.9890\n",
      "Epoch 176/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0434 - accuracy: 0.9858 - val_loss: 0.0269 - val_accuracy: 0.9900\n",
      "Epoch 177/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0292 - accuracy: 0.9903 - val_loss: 0.0322 - val_accuracy: 0.9884\n",
      "Epoch 178/200\n",
      "125/125 [==============================] - 5s 37ms/step - loss: 0.0222 - accuracy: 0.9930 - val_loss: 0.0250 - val_accuracy: 0.9900\n",
      "Epoch 179/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0372 - accuracy: 0.9873 - val_loss: 0.0405 - val_accuracy: 0.9849\n",
      "Epoch 180/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0288 - accuracy: 0.9903 - val_loss: 0.0483 - val_accuracy: 0.9819\n",
      "Epoch 181/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0468 - accuracy: 0.9842 - val_loss: 0.0969 - val_accuracy: 0.9684\n",
      "Epoch 182/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0246 - accuracy: 0.9906 - val_loss: 0.0563 - val_accuracy: 0.9804\n",
      "Epoch 183/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0227 - accuracy: 0.9912 - val_loss: 0.0359 - val_accuracy: 0.9869\n",
      "Epoch 184/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0462 - accuracy: 0.9831 - val_loss: 0.0959 - val_accuracy: 0.9699\n",
      "Epoch 185/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0287 - accuracy: 0.9893 - val_loss: 0.0424 - val_accuracy: 0.9854\n",
      "Epoch 186/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0240 - accuracy: 0.9902 - val_loss: 0.1181 - val_accuracy: 0.9653\n",
      "Epoch 187/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0278 - accuracy: 0.9903 - val_loss: 0.0546 - val_accuracy: 0.9814\n",
      "Epoch 188/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0474 - accuracy: 0.9847 - val_loss: 0.0246 - val_accuracy: 0.9930\n",
      "Epoch 189/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0269 - accuracy: 0.9890 - val_loss: 0.0292 - val_accuracy: 0.9905\n",
      "Epoch 190/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0345 - accuracy: 0.9867 - val_loss: 0.0467 - val_accuracy: 0.9824\n",
      "Epoch 191/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0285 - accuracy: 0.9886 - val_loss: 0.0396 - val_accuracy: 0.9839\n",
      "Epoch 192/200\n",
      "125/125 [==============================] - 5s 37ms/step - loss: 0.0502 - accuracy: 0.9828 - val_loss: 0.2977 - val_accuracy: 0.9181\n",
      "Epoch 193/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0349 - accuracy: 0.9893 - val_loss: 0.0306 - val_accuracy: 0.9915\n",
      "Epoch 194/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0208 - accuracy: 0.9928 - val_loss: 0.0351 - val_accuracy: 0.9864\n",
      "Epoch 195/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0286 - accuracy: 0.9902 - val_loss: 0.2074 - val_accuracy: 0.9714\n",
      "Epoch 196/200\n",
      "125/125 [==============================] - 4s 35ms/step - loss: 0.0369 - accuracy: 0.9871 - val_loss: 0.0291 - val_accuracy: 0.9900\n",
      "Epoch 197/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0229 - accuracy: 0.9931 - val_loss: 0.0783 - val_accuracy: 0.9719\n",
      "Epoch 198/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0182 - accuracy: 0.9937 - val_loss: 0.0409 - val_accuracy: 0.9864\n",
      "Epoch 199/200\n",
      "125/125 [==============================] - 4s 33ms/step - loss: 0.0179 - accuracy: 0.9941 - val_loss: 0.0464 - val_accuracy: 0.9844\n",
      "Epoch 200/200\n",
      "125/125 [==============================] - 4s 34ms/step - loss: 0.0269 - accuracy: 0.9920 - val_loss: 0.0306 - val_accuracy: 0.9895\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_ds, validation_data = val_ds, epochs = epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs_range = range(epochs)\n",
    "\n",
    "plt.figure(figsize=(20, 10))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
    "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(epochs_range, loss, label='Training Loss')\n",
    "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(cm, classes,\n",
    "                        normalize=False,\n",
    "                        title='Confusion matrix',\n",
    "                        cmap=plt.cm.YlGnBu):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "#     print(cm)\n",
    "\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "            horizontalalignment=\"center\",\n",
    "            color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2487 images belonging to 4 classes.\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "  EOSINOPHIL       0.88      0.85      0.86       623\n",
      "  LYMPHOCYTE       1.00      1.00      1.00       620\n",
      "    MONOCYTE       1.00      0.75      0.86       620\n",
      "  NEUTROPHIL       0.69      0.88      0.77       624\n",
      "\n",
      "    accuracy                           0.87      2487\n",
      "   macro avg       0.89      0.87      0.87      2487\n",
      "weighted avg       0.89      0.87      0.87      2487\n",
      "\n"
     ]
    }
   ],
   "source": [
    "test_generator = tf.keras.preprocessing.image.ImageDataGenerator()\n",
    "test_data_generator = test_generator.flow_from_directory(\n",
    "    test_data_dir, # Put your path here\n",
    "     target_size=(img_width, img_height),\n",
    "    batch_size=64,\n",
    "    shuffle=False)\n",
    "test_steps_per_epoch = np.math.ceil(test_data_generator.samples / test_data_generator.batch_size)\n",
    "\n",
    "true_classes = test_data_generator.classes\n",
    "class_labels = list(test_data_generator.class_indices.keys())  \n",
    "predictions = model.predict_generator(test_data_generator, steps=test_steps_per_epoch)\n",
    "predicted_classes= np.argmax(predictions, axis=1)\n",
    "\n",
    "report = classification_report(true_classes, predicted_classes, target_names = class_labels)\n",
    "print(report)\n",
    "cm = confusion_matrix(y_true=true_classes, y_pred=predicted_classes)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7, 7))\n",
    "plot_confusion_matrix(cm=cm, classes=class_labels, title='Confusion Matrix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.utils import plot_model\n",
    "\n",
    "plot_model(model, to_file=\"model.png\", show_shapes=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.get_config()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save_weights('model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in model.layers:\n",
    "    if 'conv' in layer.name:\n",
    "        weights, bias= layer.get_weights()\n",
    "#         print(layer.name, filters.shape)\n",
    "        \n",
    "        #normalize filter values between  0 and 1 for visualization\n",
    "        f_min, f_max = weights.min(), weights.max()\n",
    "        filters = (weights - f_min) / (f_max - f_min)  \n",
    "        print(filters.shape[3])\n",
    "        filter_cnt=1\n",
    "        plt.figure(figsize=(10, 20))\n",
    "\n",
    "        #plotting all the filters\n",
    "        for i in range(filters.shape[3]):\n",
    "            #get the filters\n",
    "            filt=filters[:,:,:, i]\n",
    "\n",
    "            #plotting each of the channel, color image RGB channels\n",
    "            for j in range(filters.shape[0]):\n",
    "                ax= plt.subplot(filters.shape[3], filters.shape[0], filter_cnt  )\n",
    "                ax.set_xticks([])\n",
    "                ax.set_yticks([])\n",
    "                plt.imshow(filt[:,:, j])\n",
    "                filter_cnt+=1\n",
    "        plt.savefig(str(filters.shape[3])+\".jpg\")\n",
    "        plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "img_path='../input/blood-cells/dataset-master/dataset-master/JPEGImages/BloodImage_00100.jpg'\n",
    "# Define a new Model, Input= image \n",
    "# Output= intermediate representations for all layers in the  \n",
    "# previous model after the first.\n",
    "successive_outputs = [layer.output for layer in model.layers[1:]]#visualization_model = Model(img_input, successive_outputs)\n",
    "visualization_model = tf.keras.models.Model(inputs = model.input, outputs = successive_outputs)#Load the input image\n",
    "img = load_img(img_path, target_size=(180, 180))# Convert ht image to Array of dimension (150,150,3)\n",
    "x   = img_to_array(img)                           \n",
    "x   = x.reshape((1,) + x.shape)# Rescale by 1/255\n",
    "x /= 255.0# Let's run input image through our vislauization network\n",
    "# to obtain all intermediate representations for the image.\n",
    "successive_feature_maps = visualization_model.predict(x)# Retrieve are the names of the layers, so can have them as part of our plot\n",
    "layer_names = [layer.name for layer in model.layers]\n",
    "for layer_name, feature_map in zip(layer_names, successive_feature_maps):\n",
    "  print(feature_map.shape)\n",
    "  if len(feature_map.shape) == 4:\n",
    "    \n",
    "    # Plot Feature maps for the conv / maxpool layers, not the fully-connected layers\n",
    "   \n",
    "    n_features = feature_map.shape[-1]  # number of features in the feature map\n",
    "    size       = feature_map.shape[ 1]  # feature map shape (1, size, size, n_features)\n",
    "    \n",
    "    # We will tile our images in this matrix\n",
    "    display_grid = np.zeros((size, size * n_features))\n",
    "    \n",
    "    # Postprocess the feature to be visually palatable\n",
    "    for i in range(n_features):\n",
    "      x  = feature_map[0, :, :, i]\n",
    "      x -= x.mean()\n",
    "      x /= x.std ()\n",
    "      x *=  64\n",
    "      x += 128\n",
    "      x  = np.clip(x, 0, 255).astype('uint8')\n",
    "      # Tile each filter into a horizontal grid\n",
    "      display_grid[:, i * size : (i + 1) * size] = x# Display the grid\n",
    "    scale = 20. / n_features\n",
    "    plt.figure( figsize=(scale * n_features, scale) )\n",
    "    plt.title ( layer_name )\n",
    "    plt.grid  ( False )\n",
    "    plt.imshow( display_grid, aspect='auto', cmap='viridis' )"
   ]
  }
 ],
 "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": "3.6.10"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}