MNiST Fashion Dataset | A case of Deep Neural Network.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "9de529b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\"Data Source\"#https://www.tensorflow.org/tutorials/keras/classification\n",
    "\"Label Classification\"#https://knowyourdata-tfds.withgoogle.com/#dataset=fashion_mnist&tab=STATS&select=default_segment.fashion_mnist.label.value&expanded_groups=fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5b2baad9",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_mnist=keras.datasets.fashion_mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a83339c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
      "29515/29515 [==============================] - 0s 1us/step\n",
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
      "26421880/26421880 [==============================] - 2s 0us/step\n",
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
      "5148/5148 [==============================] - 0s 0us/step\n",
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
      "4422102/4422102 [==============================] - 0s 0us/step\n"
     ]
    }
   ],
   "source": [
    "(train_images, train_labels), (test_images, test_labels) = data_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "2e7f8700",
   "metadata": {},
   "outputs": [],
   "source": [
    "(X_train, y_train), (X_test, y_test) = data_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "f4ec72ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "Class=['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
    "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "12598885",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "83df0527",
   "metadata": {},
   "outputs": [],
   "source": [
    "#no RGB value hence this is a grey scale image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "464b56e0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 28, 28)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d6e1f61f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1593e3d60>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 200x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(2,2))\n",
    "plt.imshow(X_train[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "30e2b9a5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "d41900f4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Ankle boot'"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_train[1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "cae23e3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sample_image(index):\n",
    "    plt.figure(figsize=(2,2))\n",
    "    plt.imshow(X_train[index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "7eb18389",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 200x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sample_image(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "e54e6fa5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Ankle boot'"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_train[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "1fdebc41",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 200x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sample_image(7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "51d08763",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Pullover'"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_train[7]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "21950909",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(Class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "792e0f71",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x1000 with 36 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "for i in range(36):\n",
    "    plt.subplot(6,6,i+1)\n",
    "    plt.xticks([])\n",
    "    plt.yticks([])\n",
    "    plt.grid(False)\n",
    "    plt.imshow(X_train[i], cmap=plt.cm.binary)\n",
    "    plt.tight_layout\n",
    "    plt.xlabel(Class[y_train[i]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "94e4f350",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_scaled=X_train/255\n",
    "X_test_scaled=X_test/255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "497a9ebe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 0, 0, ..., 0, 0, 1],\n",
       "       [1, 0, 0, ..., 0, 0, 0],\n",
       "       [1, 0, 0, ..., 0, 0, 0],\n",
       "       ...,\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [1, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0]])"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_categorical=keras.utils.to_categorical(y_train,num_classes=len(Class),dtype='int64')\n",
    "y_train_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "af0f49db",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test_categorical=keras.utils.to_categorical(y_test,num_classes=len(Class),dtype='int64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "947c13db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2980 - accuracy: 0.1593\n",
      "Epoch 2/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 2.2959 - accuracy: 0.1734\n",
      "Epoch 3/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 2.2934 - accuracy: 0.1885\n",
      "Epoch 4/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2903 - accuracy: 0.2097\n",
      "Epoch 5/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2858 - accuracy: 0.2304\n",
      "Epoch 6/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2794 - accuracy: 0.2688\n",
      "Epoch 7/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2696 - accuracy: 0.2866\n",
      "Epoch 8/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 2.2537 - accuracy: 0.3270\n",
      "Epoch 9/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 2.2260 - accuracy: 0.3424\n",
      "Epoch 10/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 2.1759 - accuracy: 0.3444\n",
      "Epoch 11/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 2.0893 - accuracy: 0.3289\n",
      "Epoch 12/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 1.9675 - accuracy: 0.3552\n",
      "Epoch 13/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.8454 - accuracy: 0.3803\n",
      "Epoch 14/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.7461 - accuracy: 0.4040\n",
      "Epoch 15/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 1.6608 - accuracy: 0.4343\n",
      "Epoch 16/500\n",
      "1875/1875 [==============================] - 16s 9ms/step - loss: 1.5742 - accuracy: 0.4565\n",
      "Epoch 17/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 1.4809 - accuracy: 0.4875\n",
      "Epoch 18/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.3872 - accuracy: 0.5081\n",
      "Epoch 19/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.3060 - accuracy: 0.5256\n",
      "Epoch 20/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 1.2407 - accuracy: 0.5403\n",
      "Epoch 21/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 1.1910 - accuracy: 0.5526\n",
      "Epoch 22/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 1.1513 - accuracy: 0.5630\n",
      "Epoch 23/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.1228 - accuracy: 0.5694\n",
      "Epoch 24/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 1.0989 - accuracy: 0.5753\n",
      "Epoch 25/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 1.0782 - accuracy: 0.5814\n",
      "Epoch 26/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 1.0590 - accuracy: 0.5891\n",
      "Epoch 27/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 1.0434 - accuracy: 0.5925\n",
      "Epoch 28/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 1.0277 - accuracy: 0.5963\n",
      "Epoch 29/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 1.0138 - accuracy: 0.6023\n",
      "Epoch 30/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 1.0030 - accuracy: 0.6043\n",
      "Epoch 31/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9935 - accuracy: 0.6094\n",
      "Epoch 32/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9835 - accuracy: 0.6118\n",
      "Epoch 33/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9743 - accuracy: 0.6154\n",
      "Epoch 34/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9635 - accuracy: 0.6189\n",
      "Epoch 35/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9558 - accuracy: 0.6241\n",
      "Epoch 36/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9464 - accuracy: 0.6284\n",
      "Epoch 37/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9408 - accuracy: 0.6313\n",
      "Epoch 38/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9342 - accuracy: 0.6366\n",
      "Epoch 39/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9277 - accuracy: 0.6375\n",
      "Epoch 40/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9212 - accuracy: 0.6420\n",
      "Epoch 41/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9120 - accuracy: 0.6472\n",
      "Epoch 42/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.9054 - accuracy: 0.6526\n",
      "Epoch 43/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.9014 - accuracy: 0.6546\n",
      "Epoch 44/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8945 - accuracy: 0.6593\n",
      "Epoch 45/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8851 - accuracy: 0.6633\n",
      "Epoch 46/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8768 - accuracy: 0.6686\n",
      "Epoch 47/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8724 - accuracy: 0.6722\n",
      "Epoch 48/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8624 - accuracy: 0.6771\n",
      "Epoch 49/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8620 - accuracy: 0.6791\n",
      "Epoch 50/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8514 - accuracy: 0.6848\n",
      "Epoch 51/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8406 - accuracy: 0.6900\n",
      "Epoch 52/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8352 - accuracy: 0.6935\n",
      "Epoch 53/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.8258 - accuracy: 0.6954\n",
      "Epoch 54/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.8207 - accuracy: 0.6983\n",
      "Epoch 55/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.8099 - accuracy: 0.7032\n",
      "Epoch 56/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.8037 - accuracy: 0.7039\n",
      "Epoch 57/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.8054 - accuracy: 0.7044\n",
      "Epoch 58/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.7954 - accuracy: 0.7077\n",
      "Epoch 59/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.7877 - accuracy: 0.7102\n",
      "Epoch 60/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7779 - accuracy: 0.7130\n",
      "Epoch 61/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.7755 - accuracy: 0.7135\n",
      "Epoch 62/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7694 - accuracy: 0.7157\n",
      "Epoch 63/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.7663 - accuracy: 0.7177\n",
      "Epoch 64/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7562 - accuracy: 0.7188\n",
      "Epoch 65/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7525 - accuracy: 0.7201\n",
      "Epoch 66/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7481 - accuracy: 0.7229\n",
      "Epoch 67/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.7454 - accuracy: 0.7245\n",
      "Epoch 68/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7405 - accuracy: 0.7247\n",
      "Epoch 69/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7355 - accuracy: 0.7262\n",
      "Epoch 70/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7324 - accuracy: 0.7261\n",
      "Epoch 71/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.7275 - accuracy: 0.7282\n",
      "Epoch 72/500\n",
      "1875/1875 [==============================] - 1840s 982ms/step - loss: 0.7243 - accuracy: 0.7280\n",
      "Epoch 73/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.7218 - accuracy: 0.7290\n",
      "Epoch 74/500\n",
      "1875/1875 [==============================] - 16s 8ms/step - loss: 0.7144 - accuracy: 0.7319\n",
      "Epoch 75/500\n",
      "1875/1875 [==============================] - 16s 9ms/step - loss: 0.7136 - accuracy: 0.7316\n",
      "Epoch 76/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.7112 - accuracy: 0.7329\n",
      "Epoch 77/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.7052 - accuracy: 0.7363\n",
      "Epoch 78/500\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.7010 - accuracy: 0.7370\n",
      "Epoch 79/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.7016 - accuracy: 0.7361\n",
      "Epoch 80/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6982 - accuracy: 0.7398\n",
      "Epoch 81/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6979 - accuracy: 0.7369\n",
      "Epoch 82/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.6903 - accuracy: 0.7393\n",
      "Epoch 83/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.6904 - accuracy: 0.7415\n",
      "Epoch 84/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6847 - accuracy: 0.7424\n",
      "Epoch 85/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6839 - accuracy: 0.7439\n",
      "Epoch 86/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6799 - accuracy: 0.7432\n",
      "Epoch 87/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6796 - accuracy: 0.7451\n",
      "Epoch 88/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6767 - accuracy: 0.7449\n",
      "Epoch 89/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6747 - accuracy: 0.7456\n",
      "Epoch 90/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6710 - accuracy: 0.7473\n",
      "Epoch 91/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6731 - accuracy: 0.7470\n",
      "Epoch 92/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6709 - accuracy: 0.7464\n",
      "Epoch 93/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6685 - accuracy: 0.7477\n",
      "Epoch 94/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6662 - accuracy: 0.7502\n",
      "Epoch 95/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6642 - accuracy: 0.7489\n",
      "Epoch 96/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6589 - accuracy: 0.7504\n",
      "Epoch 97/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6595 - accuracy: 0.7515\n",
      "Epoch 98/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6576 - accuracy: 0.7512\n",
      "Epoch 99/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6573 - accuracy: 0.7523\n",
      "Epoch 100/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6567 - accuracy: 0.7551\n",
      "Epoch 101/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6536 - accuracy: 0.7541\n",
      "Epoch 102/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6529 - accuracy: 0.7549\n",
      "Epoch 103/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6517 - accuracy: 0.7560\n",
      "Epoch 104/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6502 - accuracy: 0.7575\n",
      "Epoch 105/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6462 - accuracy: 0.7579\n",
      "Epoch 106/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6462 - accuracy: 0.7599\n",
      "Epoch 107/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6451 - accuracy: 0.7578\n",
      "Epoch 108/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6415 - accuracy: 0.7595\n",
      "Epoch 109/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6408 - accuracy: 0.7601\n",
      "Epoch 110/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6415 - accuracy: 0.7602\n",
      "Epoch 111/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6378 - accuracy: 0.7627\n",
      "Epoch 112/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.6361 - accuracy: 0.7642\n",
      "Epoch 113/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6363 - accuracy: 0.7622\n",
      "Epoch 114/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6360 - accuracy: 0.7621\n",
      "Epoch 115/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6325 - accuracy: 0.7634\n",
      "Epoch 116/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.6314 - accuracy: 0.7639\n",
      "Epoch 117/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.6324 - accuracy: 0.7651\n",
      "Epoch 118/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.6291 - accuracy: 0.7657\n",
      "Epoch 119/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6275 - accuracy: 0.7657\n",
      "Epoch 120/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6258 - accuracy: 0.7663\n",
      "Epoch 121/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6262 - accuracy: 0.7661\n",
      "Epoch 122/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6261 - accuracy: 0.7661\n",
      "Epoch 123/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6233 - accuracy: 0.7672\n",
      "Epoch 124/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6231 - accuracy: 0.7682\n",
      "Epoch 125/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6228 - accuracy: 0.7690\n",
      "Epoch 126/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6214 - accuracy: 0.7697\n",
      "Epoch 127/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6197 - accuracy: 0.7705\n",
      "Epoch 128/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6206 - accuracy: 0.7690\n",
      "Epoch 129/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6161 - accuracy: 0.7717\n",
      "Epoch 130/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6194 - accuracy: 0.7700\n",
      "Epoch 131/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6129 - accuracy: 0.7732\n",
      "Epoch 132/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6145 - accuracy: 0.7724\n",
      "Epoch 133/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6113 - accuracy: 0.7738\n",
      "Epoch 134/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6110 - accuracy: 0.7723\n",
      "Epoch 135/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6113 - accuracy: 0.7739\n",
      "Epoch 136/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6092 - accuracy: 0.7745\n",
      "Epoch 137/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6080 - accuracy: 0.7740\n",
      "Epoch 138/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6076 - accuracy: 0.7755\n",
      "Epoch 139/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6061 - accuracy: 0.7767\n",
      "Epoch 140/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6066 - accuracy: 0.7753\n",
      "Epoch 141/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6068 - accuracy: 0.7761\n",
      "Epoch 142/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.6029 - accuracy: 0.7760\n",
      "Epoch 143/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6025 - accuracy: 0.7770\n",
      "Epoch 144/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.6055 - accuracy: 0.7764\n",
      "Epoch 145/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6024 - accuracy: 0.7769\n",
      "Epoch 146/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6003 - accuracy: 0.7784\n",
      "Epoch 147/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.6011 - accuracy: 0.7772\n",
      "Epoch 148/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5975 - accuracy: 0.7802\n",
      "Epoch 149/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5953 - accuracy: 0.7800\n",
      "Epoch 150/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5985 - accuracy: 0.7798\n",
      "Epoch 151/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5957 - accuracy: 0.7814\n",
      "Epoch 152/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5932 - accuracy: 0.7804\n",
      "Epoch 153/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.5937 - accuracy: 0.7797\n",
      "Epoch 154/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5939 - accuracy: 0.7816\n",
      "Epoch 155/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5916 - accuracy: 0.7808\n",
      "Epoch 156/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5887 - accuracy: 0.7837\n",
      "Epoch 157/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5888 - accuracy: 0.7836\n",
      "Epoch 158/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5891 - accuracy: 0.7825\n",
      "Epoch 159/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.5886 - accuracy: 0.7851\n",
      "Epoch 160/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5885 - accuracy: 0.7850\n",
      "Epoch 161/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.5873 - accuracy: 0.7850\n",
      "Epoch 162/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5866 - accuracy: 0.7841\n",
      "Epoch 163/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.5870 - accuracy: 0.7839\n",
      "Epoch 164/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5823 - accuracy: 0.7843\n",
      "Epoch 165/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5836 - accuracy: 0.7847\n",
      "Epoch 166/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5813 - accuracy: 0.7853\n",
      "Epoch 167/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5803 - accuracy: 0.7880\n",
      "Epoch 168/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5823 - accuracy: 0.7862\n",
      "Epoch 169/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5822 - accuracy: 0.7861\n",
      "Epoch 170/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5799 - accuracy: 0.7866\n",
      "Epoch 171/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5793 - accuracy: 0.7879\n",
      "Epoch 172/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5793 - accuracy: 0.7867\n",
      "Epoch 173/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5772 - accuracy: 0.7875\n",
      "Epoch 174/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5750 - accuracy: 0.7886\n",
      "Epoch 175/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5740 - accuracy: 0.7875\n",
      "Epoch 176/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.5740 - accuracy: 0.7897\n",
      "Epoch 177/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5759 - accuracy: 0.7887\n",
      "Epoch 178/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5712 - accuracy: 0.7904\n",
      "Epoch 179/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5711 - accuracy: 0.7910\n",
      "Epoch 180/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5713 - accuracy: 0.7912\n",
      "Epoch 181/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5713 - accuracy: 0.7890\n",
      "Epoch 182/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5696 - accuracy: 0.7907\n",
      "Epoch 183/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5695 - accuracy: 0.7906\n",
      "Epoch 184/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5675 - accuracy: 0.7930\n",
      "Epoch 185/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5660 - accuracy: 0.7935\n",
      "Epoch 186/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5681 - accuracy: 0.7905\n",
      "Epoch 187/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5655 - accuracy: 0.7933\n",
      "Epoch 188/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5634 - accuracy: 0.7938\n",
      "Epoch 189/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5650 - accuracy: 0.7935\n",
      "Epoch 190/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5618 - accuracy: 0.7929\n",
      "Epoch 191/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5623 - accuracy: 0.7924\n",
      "Epoch 192/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5621 - accuracy: 0.7934\n",
      "Epoch 193/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5607 - accuracy: 0.7949\n",
      "Epoch 194/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5598 - accuracy: 0.7966\n",
      "Epoch 195/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5605 - accuracy: 0.7939\n",
      "Epoch 196/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5576 - accuracy: 0.7962\n",
      "Epoch 197/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5587 - accuracy: 0.7949\n",
      "Epoch 198/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5558 - accuracy: 0.7973\n",
      "Epoch 199/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5573 - accuracy: 0.7971\n",
      "Epoch 200/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5557 - accuracy: 0.7958\n",
      "Epoch 201/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5554 - accuracy: 0.7967\n",
      "Epoch 202/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5560 - accuracy: 0.7966\n",
      "Epoch 203/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5538 - accuracy: 0.7981\n",
      "Epoch 204/500\n",
      "1875/1875 [==============================] - 20s 11ms/step - loss: 0.5539 - accuracy: 0.7973\n",
      "Epoch 205/500\n",
      "1875/1875 [==============================] - 18s 9ms/step - loss: 0.5529 - accuracy: 0.7981\n",
      "Epoch 206/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5500 - accuracy: 0.7995\n",
      "Epoch 207/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5502 - accuracy: 0.7994\n",
      "Epoch 208/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5499 - accuracy: 0.7993\n",
      "Epoch 209/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5500 - accuracy: 0.7992\n",
      "Epoch 210/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5488 - accuracy: 0.7993\n",
      "Epoch 211/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5485 - accuracy: 0.7998\n",
      "Epoch 212/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.5469 - accuracy: 0.7998\n",
      "Epoch 213/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5473 - accuracy: 0.7996\n",
      "Epoch 214/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5451 - accuracy: 0.8015\n",
      "Epoch 215/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5444 - accuracy: 0.8013\n",
      "Epoch 216/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5436 - accuracy: 0.8027\n",
      "Epoch 217/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5435 - accuracy: 0.8022\n",
      "Epoch 218/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5420 - accuracy: 0.8018\n",
      "Epoch 219/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5428 - accuracy: 0.8009\n",
      "Epoch 220/500\n",
      "1875/1875 [==============================] - 120s 64ms/step - loss: 0.5407 - accuracy: 0.8031\n",
      "Epoch 221/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5402 - accuracy: 0.8033\n",
      "Epoch 222/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5429 - accuracy: 0.8017\n",
      "Epoch 223/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5405 - accuracy: 0.8012\n",
      "Epoch 224/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5382 - accuracy: 0.8033\n",
      "Epoch 225/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5375 - accuracy: 0.8039\n",
      "Epoch 226/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5390 - accuracy: 0.8030\n",
      "Epoch 227/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5360 - accuracy: 0.8059\n",
      "Epoch 228/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5361 - accuracy: 0.8035\n",
      "Epoch 229/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.5379 - accuracy: 0.8034\n",
      "Epoch 230/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5338 - accuracy: 0.8050\n",
      "Epoch 231/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.5359 - accuracy: 0.8043\n",
      "Epoch 232/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5335 - accuracy: 0.8064\n",
      "Epoch 233/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5334 - accuracy: 0.8064\n",
      "Epoch 234/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5323 - accuracy: 0.8066\n",
      "Epoch 235/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.5329 - accuracy: 0.8054\n",
      "Epoch 236/500\n",
      "1875/1875 [==============================] - 19s 10ms/step - loss: 0.5322 - accuracy: 0.8065\n",
      "Epoch 237/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5318 - accuracy: 0.8056\n",
      "Epoch 238/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5306 - accuracy: 0.8068\n",
      "Epoch 239/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5297 - accuracy: 0.8072\n",
      "Epoch 240/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5288 - accuracy: 0.8071\n",
      "Epoch 241/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5292 - accuracy: 0.8071\n",
      "Epoch 242/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5286 - accuracy: 0.8064\n",
      "Epoch 243/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5254 - accuracy: 0.8069\n",
      "Epoch 244/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5278 - accuracy: 0.8072\n",
      "Epoch 245/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5269 - accuracy: 0.8068\n",
      "Epoch 246/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5279 - accuracy: 0.8072\n",
      "Epoch 247/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5243 - accuracy: 0.8095\n",
      "Epoch 248/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.5238 - accuracy: 0.8081\n",
      "Epoch 249/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5229 - accuracy: 0.8100\n",
      "Epoch 250/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5230 - accuracy: 0.8090\n",
      "Epoch 251/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5251 - accuracy: 0.8092\n",
      "Epoch 252/500\n",
      "1875/1875 [==============================] - 18s 9ms/step - loss: 0.5228 - accuracy: 0.8107\n",
      "Epoch 253/500\n",
      "1875/1875 [==============================] - 17s 9ms/step - loss: 0.5191 - accuracy: 0.8107\n",
      "Epoch 254/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5215 - accuracy: 0.8102\n",
      "Epoch 255/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5218 - accuracy: 0.8097\n",
      "Epoch 256/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5204 - accuracy: 0.8102\n",
      "Epoch 257/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5194 - accuracy: 0.8105\n",
      "Epoch 258/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5215 - accuracy: 0.8106\n",
      "Epoch 259/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5194 - accuracy: 0.8104\n",
      "Epoch 260/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5186 - accuracy: 0.8101\n",
      "Epoch 261/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5160 - accuracy: 0.8114\n",
      "Epoch 262/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5181 - accuracy: 0.8118\n",
      "Epoch 263/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5156 - accuracy: 0.8110\n",
      "Epoch 264/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5152 - accuracy: 0.8135\n",
      "Epoch 265/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5155 - accuracy: 0.8123\n",
      "Epoch 266/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5156 - accuracy: 0.8126\n",
      "Epoch 267/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5150 - accuracy: 0.8124\n",
      "Epoch 268/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5141 - accuracy: 0.8127\n",
      "Epoch 269/500\n",
      "1875/1875 [==============================] - 703s 375ms/step - loss: 0.5138 - accuracy: 0.8124\n",
      "Epoch 270/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5128 - accuracy: 0.8124\n",
      "Epoch 271/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5135 - accuracy: 0.8134\n",
      "Epoch 272/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5109 - accuracy: 0.8148\n",
      "Epoch 273/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5128 - accuracy: 0.8134\n",
      "Epoch 274/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5110 - accuracy: 0.8131\n",
      "Epoch 275/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5108 - accuracy: 0.8149\n",
      "Epoch 276/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5096 - accuracy: 0.8141\n",
      "Epoch 277/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5100 - accuracy: 0.8140\n",
      "Epoch 278/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5088 - accuracy: 0.8132\n",
      "Epoch 279/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5073 - accuracy: 0.8146\n",
      "Epoch 280/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5108 - accuracy: 0.8139\n",
      "Epoch 281/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5080 - accuracy: 0.8141\n",
      "Epoch 282/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5082 - accuracy: 0.8144\n",
      "Epoch 283/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5075 - accuracy: 0.8141\n",
      "Epoch 284/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5073 - accuracy: 0.8157\n",
      "Epoch 285/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5066 - accuracy: 0.8151\n",
      "Epoch 286/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5051 - accuracy: 0.8160\n",
      "Epoch 287/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5058 - accuracy: 0.8166\n",
      "Epoch 288/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5027 - accuracy: 0.8169\n",
      "Epoch 289/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5051 - accuracy: 0.8161\n",
      "Epoch 290/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5035 - accuracy: 0.8159\n",
      "Epoch 291/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.5036 - accuracy: 0.8161\n",
      "Epoch 292/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5022 - accuracy: 0.8161\n",
      "Epoch 293/500\n",
      "1875/1875 [==============================] - 388s 207ms/step - loss: 0.5020 - accuracy: 0.8164\n",
      "Epoch 294/500\n",
      "1875/1875 [==============================] - 18s 9ms/step - loss: 0.5027 - accuracy: 0.8173\n",
      "Epoch 295/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5020 - accuracy: 0.8162\n",
      "Epoch 296/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5015 - accuracy: 0.8193\n",
      "Epoch 297/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5010 - accuracy: 0.8173\n",
      "Epoch 298/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.5029 - accuracy: 0.8161\n",
      "Epoch 299/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4978 - accuracy: 0.8187\n",
      "Epoch 300/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4998 - accuracy: 0.8177\n",
      "Epoch 301/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4999 - accuracy: 0.8179\n",
      "Epoch 302/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.5015 - accuracy: 0.8174\n",
      "Epoch 303/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4987 - accuracy: 0.8179\n",
      "Epoch 304/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4971 - accuracy: 0.8194\n",
      "Epoch 305/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4966 - accuracy: 0.8194\n",
      "Epoch 306/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4997 - accuracy: 0.8176\n",
      "Epoch 307/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4972 - accuracy: 0.8195\n",
      "Epoch 308/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4962 - accuracy: 0.8196\n",
      "Epoch 309/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4961 - accuracy: 0.8198\n",
      "Epoch 310/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4916 - accuracy: 0.8206\n",
      "Epoch 311/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4949 - accuracy: 0.8202\n",
      "Epoch 312/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4969 - accuracy: 0.8188\n",
      "Epoch 313/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4933 - accuracy: 0.8193\n",
      "Epoch 314/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4957 - accuracy: 0.8199\n",
      "Epoch 315/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4929 - accuracy: 0.8204\n",
      "Epoch 316/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4936 - accuracy: 0.8202\n",
      "Epoch 317/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4929 - accuracy: 0.8199\n",
      "Epoch 318/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4954 - accuracy: 0.8200\n",
      "Epoch 319/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4921 - accuracy: 0.8197\n",
      "Epoch 320/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4930 - accuracy: 0.8208\n",
      "Epoch 321/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4929 - accuracy: 0.8207\n",
      "Epoch 322/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4907 - accuracy: 0.8204\n",
      "Epoch 323/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4924 - accuracy: 0.8200\n",
      "Epoch 324/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.4911 - accuracy: 0.8217\n",
      "Epoch 325/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4900 - accuracy: 0.8203\n",
      "Epoch 326/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4917 - accuracy: 0.8201\n",
      "Epoch 327/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4911 - accuracy: 0.8212\n",
      "Epoch 328/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4899 - accuracy: 0.8220\n",
      "Epoch 329/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4873 - accuracy: 0.8234\n",
      "Epoch 330/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4869 - accuracy: 0.8230\n",
      "Epoch 331/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4891 - accuracy: 0.8224\n",
      "Epoch 332/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4874 - accuracy: 0.8216\n",
      "Epoch 333/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4858 - accuracy: 0.8233\n",
      "Epoch 334/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4873 - accuracy: 0.8222\n",
      "Epoch 335/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4874 - accuracy: 0.8235\n",
      "Epoch 336/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4877 - accuracy: 0.8216\n",
      "Epoch 337/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4850 - accuracy: 0.8230\n",
      "Epoch 338/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4863 - accuracy: 0.8227\n",
      "Epoch 339/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4856 - accuracy: 0.8239\n",
      "Epoch 340/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4840 - accuracy: 0.8230\n",
      "Epoch 341/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4841 - accuracy: 0.8234\n",
      "Epoch 342/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4850 - accuracy: 0.8234\n",
      "Epoch 343/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4821 - accuracy: 0.8238\n",
      "Epoch 344/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4829 - accuracy: 0.8243\n",
      "Epoch 345/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4841 - accuracy: 0.8232\n",
      "Epoch 346/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4835 - accuracy: 0.8238\n",
      "Epoch 347/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4801 - accuracy: 0.8252\n",
      "Epoch 348/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4834 - accuracy: 0.8238\n",
      "Epoch 349/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4832 - accuracy: 0.8248\n",
      "Epoch 350/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4806 - accuracy: 0.8240\n",
      "Epoch 351/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4821 - accuracy: 0.8237\n",
      "Epoch 352/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4807 - accuracy: 0.8251\n",
      "Epoch 353/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4796 - accuracy: 0.8244\n",
      "Epoch 354/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4800 - accuracy: 0.8244\n",
      "Epoch 355/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4792 - accuracy: 0.8254\n",
      "Epoch 356/500\n",
      "1875/1875 [==============================] - 3413s 2s/step - loss: 0.4785 - accuracy: 0.8258\n",
      "Epoch 357/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4802 - accuracy: 0.8252\n",
      "Epoch 358/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.4799 - accuracy: 0.8254\n",
      "Epoch 359/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4778 - accuracy: 0.8259\n",
      "Epoch 360/500\n",
      "1875/1875 [==============================] - 1274s 680ms/step - loss: 0.4782 - accuracy: 0.8266\n",
      "Epoch 361/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4768 - accuracy: 0.8261\n",
      "Epoch 362/500\n",
      "1875/1875 [==============================] - 1910s 1s/step - loss: 0.4769 - accuracy: 0.8262\n",
      "Epoch 363/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4783 - accuracy: 0.8260\n",
      "Epoch 364/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4787 - accuracy: 0.8250\n",
      "Epoch 365/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4760 - accuracy: 0.8266\n",
      "Epoch 366/500\n",
      "1875/1875 [==============================] - 97s 52ms/step - loss: 0.4794 - accuracy: 0.8239\n",
      "Epoch 367/500\n",
      "1875/1875 [==============================] - 16s 8ms/step - loss: 0.4769 - accuracy: 0.8254\n",
      "Epoch 368/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4755 - accuracy: 0.8266\n",
      "Epoch 369/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4762 - accuracy: 0.8256\n",
      "Epoch 370/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4756 - accuracy: 0.8267\n",
      "Epoch 371/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4751 - accuracy: 0.8270\n",
      "Epoch 372/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4738 - accuracy: 0.8287\n",
      "Epoch 373/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4757 - accuracy: 0.8266\n",
      "Epoch 374/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4754 - accuracy: 0.8266\n",
      "Epoch 375/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4747 - accuracy: 0.8265\n",
      "Epoch 376/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4725 - accuracy: 0.8284\n",
      "Epoch 377/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4737 - accuracy: 0.8272\n",
      "Epoch 378/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4744 - accuracy: 0.8269\n",
      "Epoch 379/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4714 - accuracy: 0.8300\n",
      "Epoch 380/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4724 - accuracy: 0.8285\n",
      "Epoch 381/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4751 - accuracy: 0.8270\n",
      "Epoch 382/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4705 - accuracy: 0.8284\n",
      "Epoch 383/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4711 - accuracy: 0.8283\n",
      "Epoch 384/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4701 - accuracy: 0.8277\n",
      "Epoch 385/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4719 - accuracy: 0.8277\n",
      "Epoch 386/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4708 - accuracy: 0.8294\n",
      "Epoch 387/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.4698 - accuracy: 0.8293\n",
      "Epoch 388/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4682 - accuracy: 0.8297\n",
      "Epoch 389/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4705 - accuracy: 0.8281\n",
      "Epoch 390/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4708 - accuracy: 0.8300\n",
      "Epoch 391/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4693 - accuracy: 0.8293\n",
      "Epoch 392/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.4697 - accuracy: 0.8286\n",
      "Epoch 393/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4684 - accuracy: 0.8301\n",
      "Epoch 394/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4667 - accuracy: 0.8301\n",
      "Epoch 395/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4691 - accuracy: 0.8295\n",
      "Epoch 396/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4680 - accuracy: 0.8297\n",
      "Epoch 397/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4671 - accuracy: 0.8303\n",
      "Epoch 398/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4676 - accuracy: 0.8293\n",
      "Epoch 399/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4665 - accuracy: 0.8300\n",
      "Epoch 400/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4655 - accuracy: 0.8297\n",
      "Epoch 401/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4653 - accuracy: 0.8299\n",
      "Epoch 402/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.4678 - accuracy: 0.8303\n",
      "Epoch 403/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4654 - accuracy: 0.8306\n",
      "Epoch 404/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4660 - accuracy: 0.8305\n",
      "Epoch 405/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4643 - accuracy: 0.8323\n",
      "Epoch 406/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4662 - accuracy: 0.8299\n",
      "Epoch 407/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4663 - accuracy: 0.8304\n",
      "Epoch 408/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4650 - accuracy: 0.8309\n",
      "Epoch 409/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4659 - accuracy: 0.8289\n",
      "Epoch 410/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4661 - accuracy: 0.8309\n",
      "Epoch 411/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4636 - accuracy: 0.8321\n",
      "Epoch 412/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4633 - accuracy: 0.8319\n",
      "Epoch 413/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4633 - accuracy: 0.8299\n",
      "Epoch 414/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4640 - accuracy: 0.8309\n",
      "Epoch 415/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4637 - accuracy: 0.8318\n",
      "Epoch 416/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4612 - accuracy: 0.8335\n",
      "Epoch 417/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4640 - accuracy: 0.8320\n",
      "Epoch 418/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4619 - accuracy: 0.8320\n",
      "Epoch 419/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4609 - accuracy: 0.8325\n",
      "Epoch 420/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4614 - accuracy: 0.8327\n",
      "Epoch 421/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4603 - accuracy: 0.8338\n",
      "Epoch 422/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4581 - accuracy: 0.8335\n",
      "Epoch 423/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4620 - accuracy: 0.8322\n",
      "Epoch 424/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4591 - accuracy: 0.8339\n",
      "Epoch 425/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4591 - accuracy: 0.8329\n",
      "Epoch 426/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4603 - accuracy: 0.8336\n",
      "Epoch 427/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4605 - accuracy: 0.8326\n",
      "Epoch 428/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4605 - accuracy: 0.8318\n",
      "Epoch 429/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4600 - accuracy: 0.8336\n",
      "Epoch 430/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4589 - accuracy: 0.8332\n",
      "Epoch 431/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4604 - accuracy: 0.8321\n",
      "Epoch 432/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4583 - accuracy: 0.8327\n",
      "Epoch 433/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4591 - accuracy: 0.8322\n",
      "Epoch 434/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4589 - accuracy: 0.8330\n",
      "Epoch 435/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4582 - accuracy: 0.8326\n",
      "Epoch 436/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4594 - accuracy: 0.8330\n",
      "Epoch 437/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4577 - accuracy: 0.8341\n",
      "Epoch 438/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4553 - accuracy: 0.8337\n",
      "Epoch 439/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4579 - accuracy: 0.8335\n",
      "Epoch 440/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4569 - accuracy: 0.8339\n",
      "Epoch 441/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4567 - accuracy: 0.8337\n",
      "Epoch 442/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4576 - accuracy: 0.8331\n",
      "Epoch 443/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4564 - accuracy: 0.8343\n",
      "Epoch 444/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4562 - accuracy: 0.8335\n",
      "Epoch 445/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4539 - accuracy: 0.8346\n",
      "Epoch 446/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4555 - accuracy: 0.8349\n",
      "Epoch 447/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4567 - accuracy: 0.8334\n",
      "Epoch 448/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4552 - accuracy: 0.8341\n",
      "Epoch 449/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4540 - accuracy: 0.8350\n",
      "Epoch 450/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4530 - accuracy: 0.8357\n",
      "Epoch 451/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4547 - accuracy: 0.8331\n",
      "Epoch 452/500\n",
      "1875/1875 [==============================] - 11s 6ms/step - loss: 0.4518 - accuracy: 0.8359\n",
      "Epoch 453/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4539 - accuracy: 0.8353\n",
      "Epoch 454/500\n",
      "1875/1875 [==============================] - 1896s 1s/step - loss: 0.4530 - accuracy: 0.8354\n",
      "Epoch 455/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4520 - accuracy: 0.8355\n",
      "Epoch 456/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4534 - accuracy: 0.8350\n",
      "Epoch 457/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.4537 - accuracy: 0.8349\n",
      "Epoch 458/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4518 - accuracy: 0.8361\n",
      "Epoch 459/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4527 - accuracy: 0.8351\n",
      "Epoch 460/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4514 - accuracy: 0.8339\n",
      "Epoch 461/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4522 - accuracy: 0.8356\n",
      "Epoch 462/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4496 - accuracy: 0.8366\n",
      "Epoch 463/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4507 - accuracy: 0.8358\n",
      "Epoch 464/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4496 - accuracy: 0.8363\n",
      "Epoch 465/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4516 - accuracy: 0.8362\n",
      "Epoch 466/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4489 - accuracy: 0.8370\n",
      "Epoch 467/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4517 - accuracy: 0.8349\n",
      "Epoch 468/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4504 - accuracy: 0.8353\n",
      "Epoch 469/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4513 - accuracy: 0.8359\n",
      "Epoch 470/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4490 - accuracy: 0.8357\n",
      "Epoch 471/500\n",
      "1875/1875 [==============================] - 12s 7ms/step - loss: 0.4482 - accuracy: 0.8352\n",
      "Epoch 472/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4508 - accuracy: 0.8353\n",
      "Epoch 473/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4489 - accuracy: 0.8365\n",
      "Epoch 474/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4495 - accuracy: 0.8360\n",
      "Epoch 475/500\n",
      "1875/1875 [==============================] - 12s 6ms/step - loss: 0.4468 - accuracy: 0.8371\n",
      "Epoch 476/500\n",
      "1875/1875 [==============================] - 3180s 2s/step - loss: 0.4493 - accuracy: 0.8359\n",
      "Epoch 477/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4479 - accuracy: 0.8341\n",
      "Epoch 478/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4470 - accuracy: 0.8362\n",
      "Epoch 479/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4473 - accuracy: 0.8366\n",
      "Epoch 480/500\n",
      "1875/1875 [==============================] - 14s 8ms/step - loss: 0.4502 - accuracy: 0.8360\n",
      "Epoch 481/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.4481 - accuracy: 0.8362\n",
      "Epoch 482/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4469 - accuracy: 0.8374\n",
      "Epoch 483/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4472 - accuracy: 0.8364\n",
      "Epoch 484/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4465 - accuracy: 0.8383\n",
      "Epoch 485/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4462 - accuracy: 0.8374\n",
      "Epoch 486/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4457 - accuracy: 0.8367\n",
      "Epoch 487/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4451 - accuracy: 0.8373\n",
      "Epoch 488/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4444 - accuracy: 0.8372\n",
      "Epoch 489/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4463 - accuracy: 0.8383\n",
      "Epoch 490/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4460 - accuracy: 0.8379\n",
      "Epoch 491/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4453 - accuracy: 0.8381\n",
      "Epoch 492/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4447 - accuracy: 0.8386\n",
      "Epoch 493/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4450 - accuracy: 0.8375\n",
      "Epoch 494/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4461 - accuracy: 0.8371\n",
      "Epoch 495/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4462 - accuracy: 0.8384\n",
      "Epoch 496/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4448 - accuracy: 0.8372\n",
      "Epoch 497/500\n",
      "1875/1875 [==============================] - 15s 8ms/step - loss: 0.4451 - accuracy: 0.8373\n",
      "Epoch 498/500\n",
      "1875/1875 [==============================] - 13s 7ms/step - loss: 0.4442 - accuracy: 0.8382\n",
      "Epoch 499/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4436 - accuracy: 0.8394\n",
      "Epoch 500/500\n",
      "1875/1875 [==============================] - 14s 7ms/step - loss: 0.4426 - accuracy: 0.8389\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x17126b2b0>"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model=keras.Sequential([\n",
    "    keras.layers.Flatten(input_shape=(28,28)),\n",
    "    keras.layers.Dense(1000,activation='relu'),\n",
    "    keras.layers.Dense(100,activation='relu'),\n",
    "    keras.layers.Dense(len(Class),activation='sigmoid')\n",
    "])\n",
    "\n",
    "model.compile(optimizer='SGD',\n",
    "                  loss='categorical_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "\n",
    "model.fit(X_train_scaled,y_train_categorical,epochs=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "237efa43",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Ankle boot'"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "eda5649e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step - loss: 37.8605 - accuracy: 0.7407\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[37.8604736328125, 0.7407000064849854]"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(X_test_scaled,y_test_categorical)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "d3dd0961",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'Sandal'"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model.predict(X_test_scaled)[0])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "3eb76323",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Ankle boot'"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "6b00de67",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 5ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'T-shirt/top'"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model.predict(X_test_scaled)[10])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "28993eeb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Coat'"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[10]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "6cb03aab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'T-shirt/top'"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model.predict(X_test_scaled)[256])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "cecea536",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Bag'"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[256]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "14bb9250",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"Model needs to be trained further for better accuracy. Let us change the optimizer and check for results\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "fea4d57f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1875/1875 [==============================] - 26s 13ms/step - loss: 0.4722 - accuracy: 0.8300\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 25s 14ms/step - loss: 0.3565 - accuracy: 0.8690\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 26s 14ms/step - loss: 0.3217 - accuracy: 0.8814\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 26s 14ms/step - loss: 0.2987 - accuracy: 0.8888\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 26s 14ms/step - loss: 0.2788 - accuracy: 0.8949\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1595ae700>"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1=keras.Sequential([\n",
    "    keras.layers.Flatten(input_shape=(28,28)),\n",
    "    keras.layers.Dense(1000,activation='relu'),\n",
    "    keras.layers.Dense(100,activation='relu'),\n",
    "    keras.layers.Dense(len(Class),activation='sigmoid')\n",
    "])\n",
    "\n",
    "model1.compile(optimizer='adam',\n",
    "                  loss='categorical_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "\n",
    "model1.fit(X_train_scaled,y_train_categorical,epochs=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "5af7e2f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'Bag'"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model1.predict(X_test_scaled)[256])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "230a80b9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Bag'"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[256]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "88cd2706",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 3ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'Sneaker'"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model1.predict(X_test_scaled)[22])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "02c483c1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Sneaker'"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[22]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "356a0e92",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 3ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'T-shirt/top'"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[np.argmax(model1.predict(X_test_scaled)[19])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "946c002c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'T-shirt/top'"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Class[y_test[19]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58f9a655",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}