Deep_Learning_Final_Project.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import pandas as pd\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Activation, Dense, Flatten, BatchNormalization, Conv2D, MaxPool2D, Dropout, AveragePooling2D\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.metrics import categorical_crossentropy, mean_squared_error\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "\n",
    "import itertools\n",
    "\n",
    "#libraries to help visualize data\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#libraries to shuffle image data\n",
    "import os #used to change directories and make directories\n",
    "import shutil #used to move image samples from directories\n",
    "import random #used to shuffle samples\n",
    "import glob\n",
    "\n",
    "#import another python file we created"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "type_1 = [\"Bug\", \"Dark\", \"Dragon\", \"Electric\", \"Fairy\", \"Fighting\", \"Fire\", \"Flying\", \"Ghost\", \"Grass\", \"Ground\", \"Ice\", \"Normal\", \"Poison\", \"Psychic\", \"Rock\", \"Steel\", \"Water\"]\n",
    "type_2 = [\"Bug\", \"Dark\", \"Dragon\", \"Electric\", \"Fairy\", \"Fighting\", \"Fire\", \"Flying\", \"Ghost\", \"Grass\", \"Ground\", \"Ice\", \"None\", \"Normal\", \"Poison\", \"Psychic\", \"Rock\", \"Steel\", \"Water\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#For the first type water is the largest with 803 files\n",
    "train_type1 = \"CSCI4931_DLFinal-main/Dataset/type1_sorted/train\"\n",
    "test_type1 = \"CSCI4931_DLFinal-main/Dataset/type1_sorted/test\"\n",
    "\n",
    "#For the second type none is the largest with 3,649 files\n",
    "train_type2 = \"CSCI4931_DLFinal-main/Dataset/type2_sorted/train\"\n",
    "test_type2 = \"CSCI4931_DLFinal-main/Dataset/type2_sorted/test\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# #read in csv file that has number of files in each folder\n",
    "num_files = pd.read_csv(\"Num_Types.csv\")\n",
    "# print(num_files[\"Type1\"].loc[0:5])\n",
    "type1_num = num_files[\"Type1\"]\n",
    "type2_num = num_files[\"Type2\"] \n",
    "\n",
    "# #threshold number to reach when oversampling (only works for 2 model multilabel classification, need to modify if using one model w/ sigmoid)\n",
    "type1_total = 850\n",
    "type2_total = 3700"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "#do image data augmentation to oversample?: https://www.marktechpost.com/2021/04/05/image-data-augmentation-in-keras/\n",
    "#oversample the data for each train folder\n",
    "#ONLY RUN THIS CELL ONCE\n",
    "os.mkdir(train_type1 + \"/temp/\")\n",
    "\n",
    "temp_path = train_type1 + \"/temp/\"\n",
    "\n",
    "#random oversampling of Pokemon's primary type \n",
    "#same process for each type where random sampling of images is generated and moved into specified folder\n",
    "#process needs to be done so that there are equal distribution of data\n",
    "\n",
    "#random oversampling for type Bug\n",
    "bug_needed = type1_total - type1_num.loc[0]\n",
    "for i in range(0, bug_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Bug/*\"), 1) #only chooses images from bug folder\n",
    "    parsedString = img[0].split('.')\n",
    "#     print(parsedString)\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "#     print(img_name)\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1] #want to save duplicates in separate folder, don't want random sample to choose out of bigger pool that includes duplicates\n",
    "    #also need to rename duplicates so that they can be saved in the same folder \n",
    "#     print(new_img)\n",
    "    shutil.copy(img[0], new_img) #copy duplicate to temp folder\n",
    "\n",
    "#https://www.geeksforgeeks.org/how-to-iterate-over-files-in-directory-using-python/\n",
    "#https://www.tutorialspoint.com/How-to-move-a-file-from-one-folder-to-another-using-Python\n",
    "dest_folder = train_type1 + \"/Bug/\"\n",
    "for dup in os.listdir(temp_path): #iterate through the list of files in the directory chosen\n",
    "    img_path = os.path.join(temp_path, dup) #combine file name and folder path to create absolute path to file\n",
    "    if os.path.isfile(img_path): #check if the path is a regular file (not a folder)\n",
    "#         print(img_path)\n",
    "        shutil.move(img_path, dest_folder) #move duplicates to designated class/type\n",
    "    \n",
    "#random oversampling for type Dark\n",
    "dark_needed = type1_total - type1_num.loc[1]\n",
    "for i in range(0, dark_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Dark/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Dark/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Dragon\n",
    "dragon_needed = type1_total - type1_num.loc[2]\n",
    "for i in range(0, dragon_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Dragon/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Dragon/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Electric\n",
    "electric_needed = type1_total - type1_num.loc[3]\n",
    "for i in range(0, electric_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Electric/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Electric/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Fairy\n",
    "fairy_needed = type1_total - type1_num.loc[4]\n",
    "for i in range(0, fairy_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Fairy/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Fairy/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "    \n",
    "#random oversampling for type Fighting\n",
    "fighting_needed = type1_total - type1_num.loc[5]\n",
    "for i in range(0, fighting_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Fighting/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Fighting/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Fire\n",
    "fire_needed = type1_total - type1_num.loc[6]\n",
    "for i in range(0, fire_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Fire/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Fire/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Flying\n",
    "flying_needed = type1_total - type1_num.loc[7]\n",
    "for i in range(0, flying_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Flying/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Flying/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ghost\n",
    "ghost_needed = type1_total - type1_num.loc[8]\n",
    "for i in range(0, ghost_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Ghost/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Ghost/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Grass\n",
    "grass_needed = type1_total - type1_num.loc[9]\n",
    "for i in range(0, grass_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Grass/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Grass/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ground\n",
    "ground_needed = type1_total - type1_num.loc[10]\n",
    "for i in range(0, ground_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Ground/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Ground/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ice\n",
    "ice_needed = type1_total - type1_num.loc[11]\n",
    "for i in range(0, ice_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Ice/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Ice/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Normal\n",
    "normal_needed = type1_total - type1_num.loc[13]\n",
    "for i in range(0, normal_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Normal/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Normal/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Poison\n",
    "poison_needed = type1_total - type1_num.loc[14]\n",
    "for i in range(0, poison_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Poison/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Poison/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Pyschic\n",
    "psychic_needed = type1_total - type1_num.loc[15]\n",
    "for i in range(0, psychic_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Psychic/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Psychic/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "    \n",
    "#random oversampling for type Rock\n",
    "rock_needed = type1_total - type1_num.loc[16]\n",
    "for i in range(0, rock_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Rock/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Rock/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Steel\n",
    "steel_needed = type1_total - type1_num.loc[17]\n",
    "for i in range(0, steel_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Steel/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Steel/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Water\n",
    "water_needed = type1_total - type1_num.loc[18]\n",
    "for i in range(0, water_needed):\n",
    "    img = random.sample(glob.glob(train_type1 + \"/Water/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type1 + \"/Water/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "os.rmdir(temp_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ONLY RUN THIS CELL ONCE\n",
    "#random oversampling for Pokemon's secondary type\n",
    "os.mkdir(train_type2 + \"/temp/\")\n",
    "\n",
    "temp_path = train_type2 + \"/temp/\"\n",
    "\n",
    "#random oversampling for type Bug\n",
    "bug_needed = type2_total - type2_num.loc[0]\n",
    "for i in range(0, bug_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Bug/*\"), 1) #only chooses images from bug folder\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1] #want to save duplicates in separate folder, don't want random sample to choose out of bigger pool that includes duplicates\n",
    "    #also need to rename duplicates so that they can be saved in the same folder \n",
    "    shutil.copy(img[0], new_img) #copy duplicate to temp folder\n",
    "\n",
    "dest_folder = train_type2 + \"/Bug/\"\n",
    "for dup in os.listdir(temp_path): #iterate through the list of files in the directory chosen\n",
    "    img_path = os.path.join(temp_path, dup) #combine file name and folder path to create absolute path to file\n",
    "    if os.path.isfile(img_path): #check if the path is a regular file (not a folder)\n",
    "        shutil.move(img_path, dest_folder) #move duplicates to designated class/type\n",
    "    \n",
    "#random oversampling for type Dark\n",
    "dark_needed = type2_total - type2_num.loc[1]\n",
    "for i in range(0, dark_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Dark/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Dark/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Dragon\n",
    "dragon_needed = type2_total - type2_num.loc[2]\n",
    "for i in range(0, dragon_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Dragon/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Dragon/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Electric\n",
    "electric_needed = type2_total - type2_num.loc[3]\n",
    "for i in range(0, electric_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Electric/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Electric/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Fairy\n",
    "fairy_needed = type2_total - type2_num.loc[4]\n",
    "for i in range(0, fairy_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Fairy/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Fairy/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "    \n",
    "#random oversampling for type Fighting\n",
    "fighting_needed = type2_total - type2_num.loc[5]\n",
    "for i in range(0, fighting_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Fighting/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Fighting/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Fire\n",
    "fire_needed = type2_total - type2_num.loc[6]\n",
    "for i in range(0, fire_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Fire/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Fire/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Flying\n",
    "flying_needed = type2_total - type2_num.loc[7]\n",
    "for i in range(0, flying_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Flying/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Flying/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ghost\n",
    "ghost_needed = type2_total - type2_num.loc[8]\n",
    "for i in range(0, ghost_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Ghost/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Ghost/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Grass\n",
    "grass_needed = type2_total - type2_num.loc[9]\n",
    "for i in range(0, grass_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Grass/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Grass/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ground\n",
    "ground_needed = type2_total - type2_num.loc[10]\n",
    "for i in range(0, ground_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Ground/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Ground/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Ice\n",
    "ice_needed = type2_total - type2_num.loc[11]\n",
    "for i in range(0, ice_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Ice/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Ice/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type None\n",
    "notype_needed = type2_total - type2_num.loc[12]\n",
    "for i in range(0, notype_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/None/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/None/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Normal\n",
    "normal_needed = type2_total - type2_num.loc[13]\n",
    "for i in range(0, normal_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Normal/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Normal/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Poison\n",
    "poison_needed = type2_total - type2_num.loc[14]\n",
    "for i in range(0, poison_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Poison/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Poison/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Pyschic\n",
    "psychic_needed = type2_total - type2_num.loc[15]\n",
    "for i in range(0, psychic_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Psychic/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Psychic/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "    \n",
    "#random oversampling for type Rock\n",
    "rock_needed = type2_total - type2_num.loc[16]\n",
    "for i in range(0, rock_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Rock/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path+ img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Rock/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Steel\n",
    "steel_needed = type2_total - type2_num.loc[17]\n",
    "for i in range(0, steel_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Steel/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Steel/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "#random oversampling for type Water\n",
    "water_needed = type2_total - type2_num.loc[18]\n",
    "for i in range(0, water_needed):\n",
    "    img = random.sample(glob.glob(train_type2 + \"/Water/*\"), 1)\n",
    "    parsedString = img[0].split('.')\n",
    "    img_name = parsedString[0].split(\"\\\\\")\n",
    "    new_img = temp_path + img_name[1] + \"(\" + str(i) + \").\" + parsedString[1]\n",
    "    shutil.copy(img[0], new_img)\n",
    "\n",
    "dest_folder = train_type2 + \"/Water/\"\n",
    "for dup in os.listdir(temp_path):\n",
    "    img_path = os.path.join(temp_path, dup)\n",
    "    if os.path.isfile(img_path):\n",
    "        shutil.move(img_path,dest_folder)\n",
    "\n",
    "os.rmdir(temp_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#data augmentation: https://www.analyticsvidhya.com/blog/2020/08/image-augmentation-on-the-fly-using-keras-imagedatagenerator/\n",
    "#ImageDataGenerator: https://keras.io/api/preprocessing/image/\n",
    "#added data augmentation so that the model does not memorize the data\n",
    "datagen_train = ImageDataGenerator(rotation_range=180,\n",
    "                             width_shift_range=0.2,\n",
    "                             height_shift_range=0.2,\n",
    "                             shear_range=0.2,\n",
    "                             zoom_range=0.4,\n",
    "                             horizontal_flip=True,\n",
    "                             vertical_flip=True,\n",
    "                             validation_split=0.2,\n",
    "                             preprocessing_function=tf.keras.applications.vgg16.preprocess_input\n",
    "                            )\n",
    "datagen_test = ImageDataGenerator(rotation_range=180,\n",
    "                             width_shift_range=0.2,\n",
    "                             height_shift_range=0.2,\n",
    "                             shear_range=0.2,\n",
    "                             zoom_range=0.4,\n",
    "                             horizontal_flip=True,\n",
    "                             vertical_flip=True,\n",
    "                             preprocessing_function=tf.keras.applications.vgg16.preprocess_input\n",
    "                            )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 15300 images belonging to 18 classes.\n",
      "Found 1475 images belonging to 18 classes.\n",
      "Found 70300 images belonging to 19 classes.\n",
      "Found 6783 images belonging to 19 classes.\n"
     ]
    }
   ],
   "source": [
    "train_batch = ImageDataGenerator(validation_split=0.2, preprocessing_function=tf.keras.applications.vgg16.preprocess_input)\\\n",
    "    .flow_from_directory(directory=train_type1, target_size=(64, 64), classes=type_1, batch_size=50, shuffle=True)\n",
    "test_batch = ImageDataGenerator(preprocessing_function=tf.keras.applications.vgg16.preprocess_input)\\\n",
    "    .flow_from_directory(directory=test_type1, target_size=(64, 64), classes=type_1, batch_size=10, shuffle=False)\n",
    "#This is for data augmentation. Foramt like this\n",
    "#train_batch = datagen_train.flow_from_directory(directory=train_type1, target_size=(64, 64), classes=type_1, batch_size=50, shuffle=True)\n",
    "#test_batch = datagen_test.flow_from_directory(directory=test_type1, target_size=(64, 64), classes=type_1, batch_size=10, shuffle=False)\n",
    "\n",
    "train_batch2 = ImageDataGenerator(validation_split=0.2, preprocessing_function=tf.keras.applications.vgg16.preprocess_input)\\\n",
    "    .flow_from_directory(directory=train_type2, target_size=(64, 64), classes=type_2, batch_size=50, shuffle=True)\n",
    "test_batch2 = ImageDataGenerator(preprocessing_function=tf.keras.applications.vgg16.preprocess_input)\\\n",
    "   .flow_from_directory(directory=test_type2, target_size=(64, 64), classes=type_2, batch_size=10, shuffle=False)\n",
    "\n",
    "#train_batch2 = datagen_train.flow_from_directory(directory=train_type2, target_size=(64, 64), classes=type_2, batch_size=50, shuffle=True)\n",
    "#test_batch2 = datagen_test.flow_from_directory(directory=test_type2, target_size=(64, 64), classes=type_2, batch_size=10, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      " [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      " [0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]]\n"
     ]
    }
   ],
   "source": [
    "#visualize data (function is from tensorflow's website)\n",
    "imgs, labels = next(train_batch)\n",
    "\n",
    "def plotImages(arr):\n",
    "    fig, axes = plt.subplots(1, 10, figsize=(10,20)) #will only show the first 10 images\n",
    "    axes = axes.flatten()\n",
    "    for img, label in zip(arr, axes): \n",
    "        label.imshow(img)\n",
    "        label.axis('off')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "plotImages(imgs)\n",
    "print(labels)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Brady\\Anaconda3\\lib\\site-packages\\PIL\\Image.py:989: UserWarning: Palette images with Transparency expressed in bytes should be converted to RGBA images\n",
      "  \"Palette images with Transparency expressed in bytes should be \"\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      " [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]]\n"
     ]
    }
   ],
   "source": [
    "#visualize data (function is from tensorflow's website)\n",
    "imgs2, labels2 = next(train_batch2)\n",
    "\n",
    "def plotImages(arr):\n",
    "    fig, axes = plt.subplots(1, 10, figsize=(10,20)) #will only show the first 10 images\n",
    "    axes = axes.flatten()\n",
    "    for img, label in zip(arr, axes): \n",
    "        label.imshow(img)\n",
    "        label.axis('off')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "plotImages(imgs2)\n",
    "print(labels2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential([\n",
    "    Conv2D(filters = 32, kernel_size = (3, 3), activation = 'relu', padding = 'same', input_shape = (64, 64, 3)),\n",
    "    Conv2D(filters = 32, kernel_size = (3, 3), activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (3,3), strides = 2),\n",
    "    Dropout(0.2),\n",
    "    Conv2D(filters = 64, kernel_size = (3, 3),activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 64, kernel_size = (3, 3), activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (3,3), strides = 2),\n",
    "    Dropout(0.2),\n",
    "    Conv2D(filters = 128, kernel_size = (3, 3),activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 256, kernel_size = (3, 3),activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 512, kernel_size = (3, 3),activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (3,3), strides = 2),\n",
    "    Dropout(0.2),\n",
    "    Flatten(),\n",
    "    Dense(units = 18, activation = 'softmax'),\n",
    "])\n",
    "#definitely need to extend, feels like data is being overfitted?\n",
    "#Try AveragePooling2d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer = Adam(learning_rate = 0.0001), loss = \"categorical_crossentropy\", metrics = [\"accuracy\",\"mean_squared_error\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 7.1406 - accuracy: 0.0762 - mean_squared_error: 0.0683\n",
      "Epoch 2/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 2.8081 - accuracy: 0.1131 - mean_squared_error: 0.0519\n",
      "Epoch 3/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 2.6342 - accuracy: 0.1749 - mean_squared_error: 0.0504\n",
      "Epoch 4/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 2.5222 - accuracy: 0.2138 - mean_squared_error: 0.0493\n",
      "Epoch 5/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 2.3566 - accuracy: 0.2930 - mean_squared_error: 0.0469\n",
      "Epoch 6/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 2.2519 - accuracy: 0.3214 - mean_squared_error: 0.0457\n",
      "Epoch 7/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 2.0865 - accuracy: 0.3599 - mean_squared_error: 0.0432\n",
      "Epoch 8/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 2.0218 - accuracy: 0.3791 - mean_squared_error: 0.0419\n",
      "Epoch 9/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 1.8448 - accuracy: 0.4217 - mean_squared_error: 0.0393\n",
      "Epoch 10/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.7550 - accuracy: 0.4615 - mean_squared_error: 0.0380\n",
      "Epoch 11/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.6870 - accuracy: 0.4896 - mean_squared_error: 0.0363\n",
      "Epoch 12/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.5872 - accuracy: 0.5278 - mean_squared_error: 0.0347\n",
      "Epoch 13/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 1.5109 - accuracy: 0.5379 - mean_squared_error: 0.0335\n",
      "Epoch 14/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 1.4762 - accuracy: 0.5442 - mean_squared_error: 0.0323\n",
      "Epoch 15/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 1.3597 - accuracy: 0.5817 - mean_squared_error: 0.0303\n",
      "Epoch 16/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 1.2983 - accuracy: 0.6097 - mean_squared_error: 0.0292\n",
      "Epoch 17/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.1739 - accuracy: 0.6496 - mean_squared_error: 0.0266\n",
      "Epoch 18/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.1143 - accuracy: 0.6519 - mean_squared_error: 0.0261\n",
      "Epoch 19/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.9863 - accuracy: 0.7111 - mean_squared_error: 0.0225\n",
      "Epoch 20/600\n",
      "50/50 [==============================] - 24s 470ms/step - loss: 0.9083 - accuracy: 0.7219 - mean_squared_error: 0.0209\n",
      "Epoch 21/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.8403 - accuracy: 0.7491 - mean_squared_error: 0.0195\n",
      "Epoch 22/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.8087 - accuracy: 0.7580 - mean_squared_error: 0.0188\n",
      "Epoch 23/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.7328 - accuracy: 0.7650 - mean_squared_error: 0.0179\n",
      "Epoch 24/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.6891 - accuracy: 0.8009 - mean_squared_error: 0.0164\n",
      "Epoch 25/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.6462 - accuracy: 0.8015 - mean_squared_error: 0.0157\n",
      "Epoch 26/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.6239 - accuracy: 0.8142 - mean_squared_error: 0.0149\n",
      "Epoch 27/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.6062 - accuracy: 0.8149 - mean_squared_error: 0.0145\n",
      "Epoch 28/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.5420 - accuracy: 0.8342 - mean_squared_error: 0.0133\n",
      "Epoch 29/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.4980 - accuracy: 0.8400 - mean_squared_error: 0.0126\n",
      "Epoch 30/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.4524 - accuracy: 0.8625 - mean_squared_error: 0.0110\n",
      "Epoch 31/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.4174 - accuracy: 0.8814 - mean_squared_error: 0.0101\n",
      "Epoch 32/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.3713 - accuracy: 0.8893 - mean_squared_error: 0.0092\n",
      "Epoch 33/600\n",
      "50/50 [==============================] - 24s 470ms/step - loss: 0.3619 - accuracy: 0.8908 - mean_squared_error: 0.0088\n",
      "Epoch 34/600\n",
      "50/50 [==============================] - 23s 470ms/step - loss: 0.3617 - accuracy: 0.9020 - mean_squared_error: 0.0085\n",
      "Epoch 35/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.3368 - accuracy: 0.9039 - mean_squared_error: 0.0082\n",
      "Epoch 36/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.2846 - accuracy: 0.9159 - mean_squared_error: 0.0071\n",
      "Epoch 37/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.2576 - accuracy: 0.9234 - mean_squared_error: 0.0066\n",
      "Epoch 38/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.2236 - accuracy: 0.9294 - mean_squared_error: 0.0057\n",
      "Epoch 39/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.2457 - accuracy: 0.9282 - mean_squared_error: 0.0060\n",
      "Epoch 40/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.3008 - accuracy: 0.9106 - mean_squared_error: 0.0074\n",
      "Epoch 41/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.2564 - accuracy: 0.9271 - mean_squared_error: 0.0063\n",
      "Epoch 42/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.2257 - accuracy: 0.9390 - mean_squared_error: 0.0054\n",
      "Epoch 43/600\n",
      "50/50 [==============================] - 24s 470ms/step - loss: 0.2501 - accuracy: 0.9300 - mean_squared_error: 0.0059\n",
      "Epoch 44/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.2421 - accuracy: 0.9214 - mean_squared_error: 0.0062\n",
      "Epoch 45/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.2109 - accuracy: 0.9402 - mean_squared_error: 0.0055\n",
      "Epoch 46/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.2037 - accuracy: 0.9387 - mean_squared_error: 0.0054\n",
      "Epoch 47/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.1616 - accuracy: 0.9461 - mean_squared_error: 0.0044\n",
      "Epoch 48/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.1408 - accuracy: 0.9559 - mean_squared_error: 0.0036\n",
      "Epoch 49/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.1637 - accuracy: 0.9524 - mean_squared_error: 0.0041\n",
      "Epoch 50/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.1610 - accuracy: 0.9505 - mean_squared_error: 0.0042\n",
      "Epoch 51/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.1774 - accuracy: 0.9443 - mean_squared_error: 0.0047\n",
      "Epoch 52/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.1649 - accuracy: 0.9529 - mean_squared_error: 0.0040\n",
      "Epoch 53/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.1259 - accuracy: 0.9666 - mean_squared_error: 0.0030\n",
      "Epoch 54/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.1100 - accuracy: 0.9644 - mean_squared_error: 0.0029\n",
      "Epoch 55/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.1198 - accuracy: 0.9685 - mean_squared_error: 0.0029\n",
      "Epoch 56/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.1052 - accuracy: 0.9675 - mean_squared_error: 0.0027\n",
      "Epoch 57/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0947 - accuracy: 0.9703 - mean_squared_error: 0.0025\n",
      "Epoch 58/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0893 - accuracy: 0.9707 - mean_squared_error: 0.0025\n",
      "Epoch 59/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.1539 - accuracy: 0.9485 - mean_squared_error: 0.0043\n",
      "Epoch 60/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.1115 - accuracy: 0.9644 - mean_squared_error: 0.0027\n",
      "Epoch 61/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0780 - accuracy: 0.9758 - mean_squared_error: 0.0021\n",
      "Epoch 62/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.1129 - accuracy: 0.9632 - mean_squared_error: 0.0031\n",
      "Epoch 63/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0985 - accuracy: 0.9690 - mean_squared_error: 0.0027\n",
      "Epoch 64/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.1167 - accuracy: 0.9633 - mean_squared_error: 0.0031\n",
      "Epoch 65/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0920 - accuracy: 0.9733 - mean_squared_error: 0.0024\n",
      "Epoch 66/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0911 - accuracy: 0.9673 - mean_squared_error: 0.0026\n",
      "Epoch 67/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0798 - accuracy: 0.9725 - mean_squared_error: 0.0023\n",
      "Epoch 68/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0862 - accuracy: 0.9713 - mean_squared_error: 0.0025\n",
      "Epoch 69/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0912 - accuracy: 0.9713 - mean_squared_error: 0.0025\n",
      "Epoch 70/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0714 - accuracy: 0.9787 - mean_squared_error: 0.0017\n",
      "Epoch 71/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0779 - accuracy: 0.9735 - mean_squared_error: 0.0022\n",
      "Epoch 72/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0896 - accuracy: 0.9686 - mean_squared_error: 0.0024\n",
      "Epoch 73/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.1047 - accuracy: 0.9722 - mean_squared_error: 0.0024\n",
      "Epoch 74/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0746 - accuracy: 0.9761 - mean_squared_error: 0.0020\n",
      "Epoch 75/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0778 - accuracy: 0.9753 - mean_squared_error: 0.0022\n",
      "Epoch 76/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0584 - accuracy: 0.9805 - mean_squared_error: 0.0016\n",
      "Epoch 77/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0654 - accuracy: 0.9802 - mean_squared_error: 0.0018\n",
      "Epoch 78/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0568 - accuracy: 0.9830 - mean_squared_error: 0.0015\n",
      "Epoch 79/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0667 - accuracy: 0.9781 - mean_squared_error: 0.0020\n",
      "Epoch 80/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0644 - accuracy: 0.9801 - mean_squared_error: 0.0017\n",
      "Epoch 81/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0663 - accuracy: 0.9829 - mean_squared_error: 0.0017\n",
      "Epoch 82/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0588 - accuracy: 0.9823 - mean_squared_error: 0.0016\n",
      "Epoch 83/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0681 - accuracy: 0.9814 - mean_squared_error: 0.0018\n",
      "Epoch 84/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0579 - accuracy: 0.9812 - mean_squared_error: 0.0016\n",
      "Epoch 85/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0832 - accuracy: 0.9723 - mean_squared_error: 0.0024\n",
      "Epoch 86/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0750 - accuracy: 0.9766 - mean_squared_error: 0.0022\n",
      "Epoch 87/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0539 - accuracy: 0.9836 - mean_squared_error: 0.0015\n",
      "Epoch 88/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0656 - accuracy: 0.9782 - mean_squared_error: 0.0019\n",
      "Epoch 89/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0756 - accuracy: 0.9758 - mean_squared_error: 0.0020\n",
      "Epoch 90/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0737 - accuracy: 0.9754 - mean_squared_error: 0.0021\n",
      "Epoch 91/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0625 - accuracy: 0.9843 - mean_squared_error: 0.0016\n",
      "Epoch 92/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0620 - accuracy: 0.9783 - mean_squared_error: 0.0018\n",
      "Epoch 93/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0451 - accuracy: 0.9875 - mean_squared_error: 0.0011\n",
      "Epoch 94/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0503 - accuracy: 0.9841 - mean_squared_error: 0.0013\n",
      "Epoch 95/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0301 - accuracy: 0.9909 - mean_squared_error: 7.8096e-04\n",
      "Epoch 96/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0440 - accuracy: 0.9861 - mean_squared_error: 0.0012\n",
      "Epoch 97/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0408 - accuracy: 0.9917 - mean_squared_error: 8.5377e-04\n",
      "Epoch 98/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0406 - accuracy: 0.9903 - mean_squared_error: 0.0011\n",
      "Epoch 99/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0435 - accuracy: 0.9882 - mean_squared_error: 0.0011\n",
      "Epoch 100/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0313 - accuracy: 0.9884 - mean_squared_error: 9.0369e-04\n",
      "Epoch 101/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0797 - accuracy: 0.9714 - mean_squared_error: 0.0023\n",
      "Epoch 102/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0422 - accuracy: 0.9867 - mean_squared_error: 0.0012\n",
      "Epoch 103/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0499 - accuracy: 0.9821 - mean_squared_error: 0.0015\n",
      "Epoch 104/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0480 - accuracy: 0.9869 - mean_squared_error: 0.0012\n",
      "Epoch 105/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0525 - accuracy: 0.9827 - mean_squared_error: 0.0014\n",
      "Epoch 106/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0669 - accuracy: 0.9795 - mean_squared_error: 0.0018\n",
      "Epoch 107/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0323 - accuracy: 0.9908 - mean_squared_error: 8.3855e-04\n",
      "Epoch 108/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0366 - accuracy: 0.9883 - mean_squared_error: 9.9352e-04\n",
      "Epoch 109/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0566 - accuracy: 0.9814 - mean_squared_error: 0.0016\n",
      "Epoch 110/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0340 - accuracy: 0.9878 - mean_squared_error: 9.6838e-04\n",
      "Epoch 111/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0326 - accuracy: 0.9917 - mean_squared_error: 8.3544e-04\n",
      "Epoch 112/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0444 - accuracy: 0.9878 - mean_squared_error: 0.0012\n",
      "Epoch 113/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0346 - accuracy: 0.9864 - mean_squared_error: 0.0011\n",
      "Epoch 114/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0374 - accuracy: 0.9871 - mean_squared_error: 0.0011\n",
      "Epoch 115/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0391 - accuracy: 0.9873 - mean_squared_error: 0.0010\n",
      "Epoch 116/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0406 - accuracy: 0.9865 - mean_squared_error: 0.0012\n",
      "Epoch 117/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0591 - accuracy: 0.9819 - mean_squared_error: 0.0015\n",
      "Epoch 118/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0394 - accuracy: 0.9873 - mean_squared_error: 0.0011\n",
      "Epoch 119/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0471 - accuracy: 0.9864 - mean_squared_error: 0.0011\n",
      "Epoch 120/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0313 - accuracy: 0.9898 - mean_squared_error: 8.5398e-04\n",
      "Epoch 121/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0244 - accuracy: 0.9915 - mean_squared_error: 7.0149e-04\n",
      "Epoch 122/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0469 - accuracy: 0.9878 - mean_squared_error: 0.0012\n",
      "Epoch 123/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0394 - accuracy: 0.9841 - mean_squared_error: 0.0012\n",
      "Epoch 124/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0410 - accuracy: 0.9846 - mean_squared_error: 0.0012\n",
      "Epoch 125/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0597 - accuracy: 0.9816 - mean_squared_error: 0.0016\n",
      "Epoch 126/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0503 - accuracy: 0.9876 - mean_squared_error: 0.0013\n",
      "Epoch 127/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0750 - accuracy: 0.9791 - mean_squared_error: 0.0019\n",
      "Epoch 128/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0531 - accuracy: 0.9807 - mean_squared_error: 0.0016\n",
      "Epoch 129/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0443 - accuracy: 0.9835 - mean_squared_error: 0.0013\n",
      "Epoch 130/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0341 - accuracy: 0.9887 - mean_squared_error: 9.9562e-04\n",
      "Epoch 131/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0352 - accuracy: 0.9878 - mean_squared_error: 9.6180e-04\n",
      "Epoch 132/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0485 - accuracy: 0.9824 - mean_squared_error: 0.0014\n",
      "Epoch 133/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0274 - accuracy: 0.9900 - mean_squared_error: 7.6973e-04\n",
      "Epoch 134/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0257 - accuracy: 0.9935 - mean_squared_error: 6.8201e-04\n",
      "Epoch 135/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0210 - accuracy: 0.9924 - mean_squared_error: 5.9847e-04\n",
      "Epoch 136/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0258 - accuracy: 0.9929 - mean_squared_error: 5.7881e-04\n",
      "Epoch 137/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0328 - accuracy: 0.9872 - mean_squared_error: 0.0010\n",
      "Epoch 138/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0267 - accuracy: 0.9885 - mean_squared_error: 8.2845e-04\n",
      "Epoch 139/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0323 - accuracy: 0.9899 - mean_squared_error: 9.1553e-04\n",
      "Epoch 140/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0320 - accuracy: 0.9896 - mean_squared_error: 9.8066e-04\n",
      "Epoch 141/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0483 - accuracy: 0.9822 - mean_squared_error: 0.0014\n",
      "Epoch 142/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0300 - accuracy: 0.9896 - mean_squared_error: 7.8793e-04\n",
      "Epoch 143/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0196 - accuracy: 0.9942 - mean_squared_error: 5.7472e-04\n",
      "Epoch 144/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0266 - accuracy: 0.9909 - mean_squared_error: 7.9259e-04\n",
      "Epoch 145/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0520 - accuracy: 0.9829 - mean_squared_error: 0.0015\n",
      "Epoch 146/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0337 - accuracy: 0.9916 - mean_squared_error: 8.7623e-04\n",
      "Epoch 147/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0412 - accuracy: 0.9892 - mean_squared_error: 0.0010\n",
      "Epoch 148/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0293 - accuracy: 0.9915 - mean_squared_error: 7.8096e-04\n",
      "Epoch 149/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0355 - accuracy: 0.9889 - mean_squared_error: 9.2713e-04\n",
      "Epoch 150/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0422 - accuracy: 0.9871 - mean_squared_error: 0.0011\n",
      "Epoch 151/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0367 - accuracy: 0.9871 - mean_squared_error: 0.0010\n",
      "Epoch 152/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0432 - accuracy: 0.9878 - mean_squared_error: 0.0010\n",
      "Epoch 153/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0280 - accuracy: 0.9895 - mean_squared_error: 8.1190e-04\n",
      "Epoch 154/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0129 - accuracy: 0.9953 - mean_squared_error: 3.5623e-04\n",
      "Epoch 155/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0121 - accuracy: 0.9960 - mean_squared_error: 3.0571e-04\n",
      "Epoch 156/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0084 - accuracy: 0.9982 - mean_squared_error: 2.0148e-04\n",
      "Epoch 157/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0185 - accuracy: 0.9949 - mean_squared_error: 4.5883e-04\n",
      "Epoch 158/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0279 - accuracy: 0.9901 - mean_squared_error: 7.7795e-04\n",
      "Epoch 159/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0301 - accuracy: 0.9910 - mean_squared_error: 8.2445e-04\n",
      "Epoch 160/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0210 - accuracy: 0.9896 - mean_squared_error: 6.6878e-04\n",
      "Epoch 161/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0247 - accuracy: 0.9914 - mean_squared_error: 7.0148e-04\n",
      "Epoch 162/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0450 - accuracy: 0.9871 - mean_squared_error: 0.0012\n",
      "Epoch 163/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0402 - accuracy: 0.9888 - mean_squared_error: 0.0010\n",
      "Epoch 164/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0361 - accuracy: 0.9872 - mean_squared_error: 0.0010\n",
      "Epoch 165/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0281 - accuracy: 0.9936 - mean_squared_error: 6.7956e-04\n",
      "Epoch 166/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0456 - accuracy: 0.9850 - mean_squared_error: 0.0012\n",
      "Epoch 167/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0238 - accuracy: 0.9917 - mean_squared_error: 6.6909e-04\n",
      "Epoch 168/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0255 - accuracy: 0.9917 - mean_squared_error: 7.2757e-04\n",
      "Epoch 169/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0304 - accuracy: 0.9919 - mean_squared_error: 7.6365e-04\n",
      "Epoch 170/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0510 - accuracy: 0.9812 - mean_squared_error: 0.0016\n",
      "Epoch 171/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0222 - accuracy: 0.9935 - mean_squared_error: 6.0535e-04\n",
      "Epoch 172/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0315 - accuracy: 0.9910 - mean_squared_error: 8.5247e-04\n",
      "Epoch 173/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0218 - accuracy: 0.9931 - mean_squared_error: 6.0243e-04\n",
      "Epoch 174/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0185 - accuracy: 0.9931 - mean_squared_error: 5.5747e-04\n",
      "Epoch 175/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0362 - accuracy: 0.9906 - mean_squared_error: 8.4143e-04\n",
      "Epoch 176/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0344 - accuracy: 0.9886 - mean_squared_error: 0.0010\n",
      "Epoch 177/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0303 - accuracy: 0.9919 - mean_squared_error: 8.4817e-04\n",
      "Epoch 178/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0169 - accuracy: 0.9939 - mean_squared_error: 4.7583e-04\n",
      "Epoch 179/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0165 - accuracy: 0.9952 - mean_squared_error: 4.6715e-04\n",
      "Epoch 180/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0277 - accuracy: 0.9907 - mean_squared_error: 8.1212e-04\n",
      "Epoch 181/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0238 - accuracy: 0.9949 - mean_squared_error: 4.7702e-04\n",
      "Epoch 182/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0171 - accuracy: 0.9937 - mean_squared_error: 5.0527e-04\n",
      "Epoch 183/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0222 - accuracy: 0.9949 - mean_squared_error: 4.9345e-04\n",
      "Epoch 184/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0227 - accuracy: 0.9923 - mean_squared_error: 6.2595e-04\n",
      "Epoch 185/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0144 - accuracy: 0.9961 - mean_squared_error: 4.3568e-04\n",
      "Epoch 186/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0093 - accuracy: 0.9964 - mean_squared_error: 2.6008e-04\n",
      "Epoch 187/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0215 - accuracy: 0.9930 - mean_squared_error: 5.0890e-04\n",
      "Epoch 188/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0166 - accuracy: 0.9955 - mean_squared_error: 4.3710e-04\n",
      "Epoch 189/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0381 - accuracy: 0.9870 - mean_squared_error: 0.0011\n",
      "Epoch 190/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0460 - accuracy: 0.9869 - mean_squared_error: 0.0012\n",
      "Epoch 191/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0212 - accuracy: 0.9939 - mean_squared_error: 6.2488e-04\n",
      "Epoch 192/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0288 - accuracy: 0.9901 - mean_squared_error: 7.7540e-04\n",
      "Epoch 193/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0181 - accuracy: 0.9939 - mean_squared_error: 5.4837e-04\n",
      "Epoch 194/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0230 - accuracy: 0.9917 - mean_squared_error: 6.3327e-04\n",
      "Epoch 195/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0167 - accuracy: 0.9953 - mean_squared_error: 4.3033e-04\n",
      "Epoch 196/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0190 - accuracy: 0.9940 - mean_squared_error: 4.9926e-04\n",
      "Epoch 197/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0348 - accuracy: 0.9919 - mean_squared_error: 7.9687e-04\n",
      "Epoch 198/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0286 - accuracy: 0.9931 - mean_squared_error: 7.1860e-04\n",
      "Epoch 199/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0134 - accuracy: 0.9965 - mean_squared_error: 4.1708e-04\n",
      "Epoch 200/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0230 - accuracy: 0.9944 - mean_squared_error: 5.4128e-04\n",
      "Epoch 201/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0316 - accuracy: 0.9887 - mean_squared_error: 0.0010\n",
      "Epoch 202/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0293 - accuracy: 0.9912 - mean_squared_error: 7.6709e-04\n",
      "Epoch 203/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0267 - accuracy: 0.9910 - mean_squared_error: 7.3446e-04\n",
      "Epoch 204/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0296 - accuracy: 0.9917 - mean_squared_error: 7.3783e-04\n",
      "Epoch 205/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0220 - accuracy: 0.9913 - mean_squared_error: 6.1298e-04\n",
      "Epoch 206/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0194 - accuracy: 0.9938 - mean_squared_error: 5.6539e-04\n",
      "Epoch 207/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0260 - accuracy: 0.9921 - mean_squared_error: 7.5178e-04\n",
      "Epoch 208/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0106 - accuracy: 0.9952 - mean_squared_error: 2.9813e-04\n",
      "Epoch 209/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0201 - accuracy: 0.9911 - mean_squared_error: 6.9317e-04\n",
      "Epoch 210/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0269 - accuracy: 0.9888 - mean_squared_error: 8.0414e-04\n",
      "Epoch 211/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0277 - accuracy: 0.9904 - mean_squared_error: 7.9519e-04\n",
      "Epoch 212/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0434 - accuracy: 0.9868 - mean_squared_error: 0.0011\n",
      "Epoch 213/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0227 - accuracy: 0.9913 - mean_squared_error: 6.4044e-04\n",
      "Epoch 214/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0379 - accuracy: 0.9875 - mean_squared_error: 0.0011\n",
      "Epoch 215/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0209 - accuracy: 0.9919 - mean_squared_error: 6.4278e-04\n",
      "Epoch 216/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0233 - accuracy: 0.9941 - mean_squared_error: 6.5514e-04\n",
      "Epoch 217/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0292 - accuracy: 0.9904 - mean_squared_error: 8.0130e-04\n",
      "Epoch 218/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0070 - accuracy: 0.9973 - mean_squared_error: 1.9771e-04\n",
      "Epoch 219/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0130 - accuracy: 0.9960 - mean_squared_error: 3.8554e-04\n",
      "Epoch 220/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0343 - accuracy: 0.9928 - mean_squared_error: 7.5592e-04\n",
      "Epoch 221/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0108 - accuracy: 0.9971 - mean_squared_error: 2.8976e-04\n",
      "Epoch 222/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0104 - accuracy: 0.9945 - mean_squared_error: 3.3178e-04\n",
      "Epoch 223/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0229 - accuracy: 0.9928 - mean_squared_error: 6.1194e-04\n",
      "Epoch 224/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0084 - accuracy: 0.9982 - mean_squared_error: 2.1969e-04\n",
      "Epoch 225/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0172 - accuracy: 0.9952 - mean_squared_error: 4.6149e-04\n",
      "Epoch 226/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0245 - accuracy: 0.9948 - mean_squared_error: 5.0720e-04\n",
      "Epoch 227/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0231 - accuracy: 0.9941 - mean_squared_error: 5.6935e-04\n",
      "Epoch 228/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0132 - accuracy: 0.9958 - mean_squared_error: 3.3261e-04\n",
      "Epoch 229/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0235 - accuracy: 0.9941 - mean_squared_error: 5.2093e-04\n",
      "Epoch 230/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0211 - accuracy: 0.9926 - mean_squared_error: 5.9074e-04\n",
      "Epoch 231/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0144 - accuracy: 0.9957 - mean_squared_error: 3.4183e-04\n",
      "Epoch 232/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0108 - accuracy: 0.9969 - mean_squared_error: 2.6463e-04\n",
      "Epoch 233/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0337 - accuracy: 0.9921 - mean_squared_error: 7.7239e-04\n",
      "Epoch 234/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0144 - accuracy: 0.9959 - mean_squared_error: 3.2184e-04\n",
      "Epoch 235/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0081 - accuracy: 0.9974 - mean_squared_error: 2.2658e-04\n",
      "Epoch 236/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0187 - accuracy: 0.9947 - mean_squared_error: 5.1242e-04\n",
      "Epoch 237/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0304 - accuracy: 0.9926 - mean_squared_error: 8.4277e-04\n",
      "Epoch 238/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0176 - accuracy: 0.9930 - mean_squared_error: 5.5929e-04\n",
      "Epoch 239/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0079 - accuracy: 0.9975 - mean_squared_error: 2.0298e-04\n",
      "Epoch 240/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0100 - accuracy: 0.9960 - mean_squared_error: 3.0660e-04\n",
      "Epoch 241/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0124 - accuracy: 0.9962 - mean_squared_error: 3.5592e-04\n",
      "Epoch 242/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0248 - accuracy: 0.9946 - mean_squared_error: 5.8120e-04\n",
      "Epoch 243/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0081 - accuracy: 0.9967 - mean_squared_error: 2.4172e-04\n",
      "Epoch 244/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0266 - accuracy: 0.9917 - mean_squared_error: 7.4611e-04\n",
      "Epoch 245/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0348 - accuracy: 0.9910 - mean_squared_error: 8.0342e-04\n",
      "Epoch 246/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0312 - accuracy: 0.9911 - mean_squared_error: 8.3765e-04\n",
      "Epoch 247/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0227 - accuracy: 0.9940 - mean_squared_error: 6.1574e-04\n",
      "Epoch 248/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0394 - accuracy: 0.9871 - mean_squared_error: 0.0011\n",
      "Epoch 249/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0209 - accuracy: 0.9912 - mean_squared_error: 6.4930e-04\n",
      "Epoch 250/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0172 - accuracy: 0.9944 - mean_squared_error: 4.3256e-04\n",
      "Epoch 251/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0162 - accuracy: 0.9971 - mean_squared_error: 2.6421e-04\n",
      "Epoch 252/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0175 - accuracy: 0.9946 - mean_squared_error: 4.4533e-04\n",
      "Epoch 253/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0194 - accuracy: 0.9949 - mean_squared_error: 5.1260e-04\n",
      "Epoch 254/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0147 - accuracy: 0.9970 - mean_squared_error: 3.3757e-04\n",
      "Epoch 255/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0383 - accuracy: 0.9893 - mean_squared_error: 9.3189e-04\n",
      "Epoch 256/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0164 - accuracy: 0.9959 - mean_squared_error: 4.2228e-04\n",
      "Epoch 257/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0277 - accuracy: 0.9922 - mean_squared_error: 7.4658e-04\n",
      "Epoch 258/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0170 - accuracy: 0.9969 - mean_squared_error: 3.9100e-04\n",
      "Epoch 259/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0147 - accuracy: 0.9930 - mean_squared_error: 4.8570e-04\n",
      "Epoch 260/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0427 - accuracy: 0.9885 - mean_squared_error: 0.0011\n",
      "Epoch 261/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0196 - accuracy: 0.9923 - mean_squared_error: 6.4374e-04\n",
      "Epoch 262/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0224 - accuracy: 0.9934 - mean_squared_error: 5.9523e-04\n",
      "Epoch 263/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0132 - accuracy: 0.9969 - mean_squared_error: 3.1376e-04\n",
      "Epoch 264/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0077 - accuracy: 0.9969 - mean_squared_error: 2.1474e-04\n",
      "Epoch 265/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0150 - accuracy: 0.9953 - mean_squared_error: 4.4880e-04\n",
      "Epoch 266/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0167 - accuracy: 0.9945 - mean_squared_error: 5.3895e-04\n",
      "Epoch 267/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0324 - accuracy: 0.9923 - mean_squared_error: 7.0572e-04\n",
      "Epoch 268/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0185 - accuracy: 0.9923 - mean_squared_error: 5.8152e-04\n",
      "Epoch 269/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0237 - accuracy: 0.9928 - mean_squared_error: 6.4578e-04\n",
      "Epoch 270/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0151 - accuracy: 0.9968 - mean_squared_error: 3.5373e-04\n",
      "Epoch 271/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0239 - accuracy: 0.9912 - mean_squared_error: 6.7172e-04\n",
      "Epoch 272/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0230 - accuracy: 0.9944 - mean_squared_error: 6.3656e-04\n",
      "Epoch 273/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0158 - accuracy: 0.9947 - mean_squared_error: 4.6695e-04\n",
      "Epoch 274/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0387 - accuracy: 0.9885 - mean_squared_error: 9.5859e-04\n",
      "Epoch 275/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0171 - accuracy: 0.9941 - mean_squared_error: 4.7504e-04\n",
      "Epoch 276/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0294 - accuracy: 0.9885 - mean_squared_error: 9.4867e-04\n",
      "Epoch 277/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0109 - accuracy: 0.9976 - mean_squared_error: 3.1084e-04\n",
      "Epoch 278/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0080 - accuracy: 0.9984 - mean_squared_error: 1.9453e-04\n",
      "Epoch 279/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0194 - accuracy: 0.9959 - mean_squared_error: 4.2060e-04\n",
      "Epoch 280/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0163 - accuracy: 0.9940 - mean_squared_error: 4.7978e-04\n",
      "Epoch 281/600\n",
      "50/50 [==============================] - 24s 481ms/step - loss: 0.0062 - accuracy: 0.9985 - mean_squared_error: 1.8849e-04\n",
      "Epoch 282/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0062 - accuracy: 0.9983 - mean_squared_error: 1.3941e-04\n",
      "Epoch 283/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0015 - accuracy: 1.0000 - mean_squared_error: 2.3516e-05\n",
      "Epoch 284/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0036 - accuracy: 0.9997 - mean_squared_error: 5.6651e-05\n",
      "Epoch 285/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0134 - accuracy: 0.9964 - mean_squared_error: 3.2125e-04\n",
      "Epoch 286/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0055 - accuracy: 0.9986 - mean_squared_error: 1.5547e-04\n",
      "Epoch 287/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0139 - accuracy: 0.9969 - mean_squared_error: 3.2934e-04\n",
      "Epoch 288/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0076 - accuracy: 0.9964 - mean_squared_error: 2.4466e-04\n",
      "Epoch 289/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0152 - accuracy: 0.9952 - mean_squared_error: 4.3315e-04\n",
      "Epoch 290/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0211 - accuracy: 0.9927 - mean_squared_error: 6.6753e-04\n",
      "Epoch 291/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0085 - accuracy: 0.9978 - mean_squared_error: 2.3837e-04\n",
      "Epoch 292/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0169 - accuracy: 0.9934 - mean_squared_error: 4.7431e-04\n",
      "Epoch 293/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0169 - accuracy: 0.9945 - mean_squared_error: 4.5376e-04\n",
      "Epoch 294/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0185 - accuracy: 0.9942 - mean_squared_error: 5.1393e-04\n",
      "Epoch 295/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0291 - accuracy: 0.9906 - mean_squared_error: 8.1503e-04\n",
      "Epoch 296/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0184 - accuracy: 0.9932 - mean_squared_error: 5.2235e-04\n",
      "Epoch 297/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0258 - accuracy: 0.9923 - mean_squared_error: 6.5580e-04\n",
      "Epoch 298/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0407 - accuracy: 0.9903 - mean_squared_error: 8.6712e-04\n",
      "Epoch 299/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0313 - accuracy: 0.9919 - mean_squared_error: 7.5892e-04\n",
      "Epoch 300/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0278 - accuracy: 0.9924 - mean_squared_error: 7.6232e-04\n",
      "Epoch 301/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0278 - accuracy: 0.9906 - mean_squared_error: 8.0635e-04\n",
      "Epoch 302/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0181 - accuracy: 0.9947 - mean_squared_error: 5.0021e-04\n",
      "Epoch 303/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0088 - accuracy: 0.9969 - mean_squared_error: 2.6303e-04\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 304/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0168 - accuracy: 0.9944 - mean_squared_error: 4.2308e-04\n",
      "Epoch 305/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0077 - accuracy: 0.9978 - mean_squared_error: 2.2078e-04\n",
      "Epoch 306/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0136 - accuracy: 0.9947 - mean_squared_error: 4.4611e-04\n",
      "Epoch 307/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0247 - accuracy: 0.9943 - mean_squared_error: 6.1602e-04\n",
      "Epoch 308/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0214 - accuracy: 0.9923 - mean_squared_error: 6.3211e-04\n",
      "Epoch 309/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0250 - accuracy: 0.9948 - mean_squared_error: 5.4205e-04\n",
      "Epoch 310/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0116 - accuracy: 0.9967 - mean_squared_error: 3.1877e-04\n",
      "Epoch 311/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0137 - accuracy: 0.9948 - mean_squared_error: 3.7477e-04\n",
      "Epoch 312/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0146 - accuracy: 0.9947 - mean_squared_error: 4.1639e-04\n",
      "Epoch 313/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0095 - accuracy: 0.9968 - mean_squared_error: 2.8778e-04\n",
      "Epoch 314/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0059 - accuracy: 0.9986 - mean_squared_error: 1.6212e-04\n",
      "Epoch 315/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0061 - accuracy: 0.9983 - mean_squared_error: 1.7915e-04\n",
      "Epoch 316/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0063 - accuracy: 0.9979 - mean_squared_error: 1.5348e-04\n",
      "Epoch 317/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0089 - accuracy: 0.9961 - mean_squared_error: 2.8518e-04\n",
      "Epoch 318/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0192 - accuracy: 0.9934 - mean_squared_error: 5.7100e-04\n",
      "Epoch 319/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0073 - accuracy: 0.9983 - mean_squared_error: 1.4276e-04\n",
      "Epoch 320/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0052 - accuracy: 0.9986 - mean_squared_error: 1.5549e-04\n",
      "Epoch 321/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0150 - accuracy: 0.9941 - mean_squared_error: 4.7891e-04\n",
      "Epoch 322/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0077 - accuracy: 0.9979 - mean_squared_error: 1.8353e-04\n",
      "Epoch 323/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0116 - accuracy: 0.9965 - mean_squared_error: 2.9245e-04\n",
      "Epoch 324/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0127 - accuracy: 0.9963 - mean_squared_error: 3.5640e-04\n",
      "Epoch 325/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0182 - accuracy: 0.9967 - mean_squared_error: 3.8195e-04\n",
      "Epoch 326/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0149 - accuracy: 0.9938 - mean_squared_error: 4.9697e-04\n",
      "Epoch 327/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0279 - accuracy: 0.9903 - mean_squared_error: 7.7139e-04\n",
      "Epoch 328/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0085 - accuracy: 0.9979 - mean_squared_error: 1.5473e-04\n",
      "Epoch 329/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0043 - accuracy: 0.9987 - mean_squared_error: 1.2610e-04\n",
      "Epoch 330/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0127 - accuracy: 0.9939 - mean_squared_error: 4.0604e-04\n",
      "Epoch 331/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0179 - accuracy: 0.9941 - mean_squared_error: 4.9091e-04\n",
      "Epoch 332/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0237 - accuracy: 0.9952 - mean_squared_error: 4.8998e-04\n",
      "Epoch 333/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0146 - accuracy: 0.9964 - mean_squared_error: 3.2222e-04\n",
      "Epoch 334/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0110 - accuracy: 0.9971 - mean_squared_error: 2.4932e-04\n",
      "Epoch 335/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0260 - accuracy: 0.9916 - mean_squared_error: 7.4463e-04\n",
      "Epoch 336/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0312 - accuracy: 0.9902 - mean_squared_error: 8.8179e-04\n",
      "Epoch 337/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0116 - accuracy: 0.9967 - mean_squared_error: 3.0692e-04\n",
      "Epoch 338/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0089 - accuracy: 0.9973 - mean_squared_error: 2.4112e-04\n",
      "Epoch 339/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0269 - accuracy: 0.9883 - mean_squared_error: 8.7278e-04\n",
      "Epoch 340/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0137 - accuracy: 0.9971 - mean_squared_error: 1.9579e-04\n",
      "Epoch 341/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0058 - accuracy: 0.9974 - mean_squared_error: 2.0120e-04\n",
      "Epoch 342/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0022 - accuracy: 0.9998 - mean_squared_error: 5.2016e-05\n",
      "Epoch 343/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0037 - accuracy: 0.9980 - mean_squared_error: 1.1809e-04\n",
      "Epoch 344/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0146 - accuracy: 0.9955 - mean_squared_error: 4.1544e-04\n",
      "Epoch 345/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0214 - accuracy: 0.9946 - mean_squared_error: 5.3208e-04\n",
      "Epoch 346/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0172 - accuracy: 0.9948 - mean_squared_error: 4.7009e-04\n",
      "Epoch 347/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0110 - accuracy: 0.9951 - mean_squared_error: 2.9172e-04\n",
      "Epoch 348/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0049 - accuracy: 0.9979 - mean_squared_error: 1.4278e-04\n",
      "Epoch 349/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0246 - accuracy: 0.9925 - mean_squared_error: 6.2807e-04\n",
      "Epoch 350/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0240 - accuracy: 0.9929 - mean_squared_error: 6.1908e-04\n",
      "Epoch 351/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0178 - accuracy: 0.9940 - mean_squared_error: 5.5314e-04\n",
      "Epoch 352/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0170 - accuracy: 0.9943 - mean_squared_error: 4.6756e-04\n",
      "Epoch 353/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0306 - accuracy: 0.9894 - mean_squared_error: 9.5331e-04\n",
      "Epoch 354/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0291 - accuracy: 0.9909 - mean_squared_error: 8.1221e-04\n",
      "Epoch 355/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0187 - accuracy: 0.9924 - mean_squared_error: 6.3639e-04\n",
      "Epoch 356/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0159 - accuracy: 0.9941 - mean_squared_error: 4.4553e-04\n",
      "Epoch 357/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0106 - accuracy: 0.9965 - mean_squared_error: 3.2362e-04\n",
      "Epoch 358/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0052 - accuracy: 0.9982 - mean_squared_error: 1.3840e-04\n",
      "Epoch 359/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0020 - accuracy: 0.9998 - mean_squared_error: 3.4204e-05\n",
      "Epoch 360/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0016 - accuracy: 1.0000 - mean_squared_error: 3.0116e-05\n",
      "Epoch 361/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0029 - accuracy: 0.9995 - mean_squared_error: 7.8891e-05\n",
      "Epoch 362/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0086 - accuracy: 0.9958 - mean_squared_error: 2.8326e-04\n",
      "Epoch 363/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0043 - accuracy: 0.9985 - mean_squared_error: 1.1756e-04\n",
      "Epoch 364/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0066 - accuracy: 0.9972 - mean_squared_error: 2.1156e-04\n",
      "Epoch 365/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0123 - accuracy: 0.9986 - mean_squared_error: 1.8576e-04\n",
      "Epoch 366/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0049 - accuracy: 0.9986 - mean_squared_error: 1.3030e-04\n",
      "Epoch 367/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0073 - accuracy: 0.9976 - mean_squared_error: 2.3202e-04\n",
      "Epoch 368/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0155 - accuracy: 0.9949 - mean_squared_error: 4.3356e-04\n",
      "Epoch 369/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0148 - accuracy: 0.9956 - mean_squared_error: 3.8401e-04\n",
      "Epoch 370/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0111 - accuracy: 0.9972 - mean_squared_error: 2.7432e-04\n",
      "Epoch 371/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0106 - accuracy: 0.9975 - mean_squared_error: 2.4417e-04\n",
      "Epoch 372/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0046 - accuracy: 0.9979 - mean_squared_error: 1.4583e-04\n",
      "Epoch 373/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0096 - accuracy: 0.9958 - mean_squared_error: 2.9424e-04\n",
      "Epoch 374/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0148 - accuracy: 0.9950 - mean_squared_error: 4.3391e-04\n",
      "Epoch 375/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0423 - accuracy: 0.9939 - mean_squared_error: 5.8645e-04\n",
      "Epoch 376/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0153 - accuracy: 0.9977 - mean_squared_error: 3.0345e-04\n",
      "Epoch 377/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0092 - accuracy: 0.9961 - mean_squared_error: 2.8055e-04\n",
      "Epoch 378/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0125 - accuracy: 0.9970 - mean_squared_error: 3.4467e-04\n",
      "Epoch 379/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0306 - accuracy: 0.9906 - mean_squared_error: 8.0101e-04\n",
      "Epoch 380/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0223 - accuracy: 0.9942 - mean_squared_error: 5.8371e-04\n",
      "Epoch 381/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0149 - accuracy: 0.9948 - mean_squared_error: 4.7195e-04\n",
      "Epoch 382/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0226 - accuracy: 0.9940 - mean_squared_error: 5.7238e-04\n",
      "Epoch 383/600\n",
      "50/50 [==============================] - 24s 484ms/step - loss: 0.0070 - accuracy: 0.9970 - mean_squared_error: 2.3227e-04\n",
      "Epoch 384/600\n",
      "50/50 [==============================] - 25s 490ms/step - loss: 0.0060 - accuracy: 0.9976 - mean_squared_error: 1.6307e-04\n",
      "Epoch 385/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0060 - accuracy: 0.9978 - mean_squared_error: 1.8301e-04\n",
      "Epoch 386/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0096 - accuracy: 0.9966 - mean_squared_error: 2.7710e-04\n",
      "Epoch 387/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0096 - accuracy: 0.9968 - mean_squared_error: 2.8432e-04\n",
      "Epoch 388/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0117 - accuracy: 0.9961 - mean_squared_error: 3.3407e-04\n",
      "Epoch 389/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0170 - accuracy: 0.9939 - mean_squared_error: 5.4196e-04\n",
      "Epoch 390/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0095 - accuracy: 0.9953 - mean_squared_error: 3.5609e-04\n",
      "Epoch 391/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0212 - accuracy: 0.9929 - mean_squared_error: 5.5549e-04\n",
      "Epoch 392/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0240 - accuracy: 0.9947 - mean_squared_error: 5.3602e-04\n",
      "Epoch 393/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0084 - accuracy: 0.9973 - mean_squared_error: 2.5374e-04\n",
      "Epoch 394/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0095 - accuracy: 0.9963 - mean_squared_error: 3.1912e-04\n",
      "Epoch 395/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0119 - accuracy: 0.9947 - mean_squared_error: 4.2407e-04\n",
      "Epoch 396/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0085 - accuracy: 0.9977 - mean_squared_error: 2.1735e-04\n",
      "Epoch 397/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0186 - accuracy: 0.9956 - mean_squared_error: 4.3626e-04\n",
      "Epoch 398/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0335 - accuracy: 0.9940 - mean_squared_error: 6.0259e-04\n",
      "Epoch 399/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0063 - accuracy: 0.9978 - mean_squared_error: 1.7689e-04\n",
      "Epoch 400/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0110 - accuracy: 0.9976 - mean_squared_error: 3.0920e-04\n",
      "Epoch 401/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0079 - accuracy: 0.9972 - mean_squared_error: 1.9493e-04\n",
      "Epoch 402/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0112 - accuracy: 0.9952 - mean_squared_error: 3.5324e-04\n",
      "Epoch 403/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0173 - accuracy: 0.9943 - mean_squared_error: 4.5419e-04\n",
      "Epoch 404/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0084 - accuracy: 0.9964 - mean_squared_error: 2.5475e-04\n",
      "Epoch 405/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0023 - accuracy: 0.9995 - mean_squared_error: 5.6645e-05\n",
      "Epoch 406/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0016 - accuracy: 0.9997 - mean_squared_error: 3.0778e-05\n",
      "Epoch 407/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0121 - accuracy: 0.9959 - mean_squared_error: 3.5994e-04\n",
      "Epoch 408/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0059 - accuracy: 0.9985 - mean_squared_error: 1.6387e-04\n",
      "Epoch 409/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0142 - accuracy: 0.9962 - mean_squared_error: 3.4698e-04\n",
      "Epoch 410/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0076 - accuracy: 0.9984 - mean_squared_error: 1.4966e-04\n",
      "Epoch 411/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0218 - accuracy: 0.9924 - mean_squared_error: 5.8355e-04\n",
      "Epoch 412/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0318 - accuracy: 0.9936 - mean_squared_error: 6.9578e-04\n",
      "Epoch 413/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0133 - accuracy: 0.9967 - mean_squared_error: 3.4133e-04\n",
      "Epoch 414/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0076 - accuracy: 0.9973 - mean_squared_error: 2.2721e-04\n",
      "Epoch 415/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0031 - accuracy: 0.9994 - mean_squared_error: 8.6412e-05\n",
      "Epoch 416/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0047 - accuracy: 0.9982 - mean_squared_error: 1.2321e-04\n",
      "Epoch 417/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0070 - accuracy: 0.9980 - mean_squared_error: 2.1537e-04\n",
      "Epoch 418/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0197 - accuracy: 0.9946 - mean_squared_error: 5.3772e-04\n",
      "Epoch 419/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0224 - accuracy: 0.9921 - mean_squared_error: 6.4916e-04\n",
      "Epoch 420/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0164 - accuracy: 0.9942 - mean_squared_error: 5.0838e-04\n",
      "Epoch 421/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0179 - accuracy: 0.9963 - mean_squared_error: 3.8419e-04\n",
      "Epoch 422/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0215 - accuracy: 0.9917 - mean_squared_error: 6.8925e-04\n",
      "Epoch 423/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0183 - accuracy: 0.9933 - mean_squared_error: 5.4181e-04\n",
      "Epoch 424/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0132 - accuracy: 0.9962 - mean_squared_error: 3.4019e-04\n",
      "Epoch 425/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0163 - accuracy: 0.9958 - mean_squared_error: 3.6542e-04\n",
      "Epoch 426/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0108 - accuracy: 0.9951 - mean_squared_error: 3.6122e-04\n",
      "Epoch 427/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0169 - accuracy: 0.9961 - mean_squared_error: 2.5584e-04\n",
      "Epoch 428/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0061 - accuracy: 0.9982 - mean_squared_error: 1.7218e-04\n",
      "Epoch 429/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0012 - accuracy: 0.9998 - mean_squared_error: 1.9090e-05\n",
      "Epoch 430/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0065 - accuracy: 0.9987 - mean_squared_error: 1.5438e-04\n",
      "Epoch 431/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0089 - accuracy: 0.9977 - mean_squared_error: 2.4690e-04\n",
      "Epoch 432/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0084 - accuracy: 0.9986 - mean_squared_error: 1.7457e-04\n",
      "Epoch 433/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0075 - accuracy: 0.9990 - mean_squared_error: 1.5553e-04\n",
      "Epoch 434/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0048 - accuracy: 0.9980 - mean_squared_error: 1.5822e-04\n",
      "Epoch 435/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0030 - accuracy: 0.9996 - mean_squared_error: 8.8913e-05\n",
      "Epoch 436/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0012 - accuracy: 0.9996 - mean_squared_error: 3.2561e-05\n",
      "Epoch 437/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0015 - accuracy: 0.9995 - mean_squared_error: 4.4679e-05\n",
      "Epoch 438/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0089 - accuracy: 0.9960 - mean_squared_error: 3.0379e-04\n",
      "Epoch 439/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0102 - accuracy: 0.9975 - mean_squared_error: 2.6102e-04\n",
      "Epoch 440/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0027 - accuracy: 0.9997 - mean_squared_error: 4.4549e-05\n",
      "Epoch 441/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0128 - accuracy: 0.9974 - mean_squared_error: 2.4131e-04\n",
      "Epoch 442/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0231 - accuracy: 0.9927 - mean_squared_error: 5.7932e-04\n",
      "Epoch 443/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0163 - accuracy: 0.9950 - mean_squared_error: 4.3053e-04\n",
      "Epoch 444/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0258 - accuracy: 0.9918 - mean_squared_error: 7.1735e-04\n",
      "Epoch 445/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0155 - accuracy: 0.9966 - mean_squared_error: 3.6140e-04\n",
      "Epoch 446/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0167 - accuracy: 0.9934 - mean_squared_error: 5.1911e-04\n",
      "Epoch 447/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0490 - accuracy: 0.9910 - mean_squared_error: 8.3561e-04\n",
      "Epoch 448/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0102 - accuracy: 0.9966 - mean_squared_error: 2.4304e-04\n",
      "Epoch 449/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0223 - accuracy: 0.9947 - mean_squared_error: 5.3213e-04\n",
      "Epoch 450/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0118 - accuracy: 0.9967 - mean_squared_error: 3.3057e-04\n",
      "Epoch 451/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0073 - accuracy: 0.9967 - mean_squared_error: 1.9655e-04\n",
      "Epoch 452/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0127 - accuracy: 0.9969 - mean_squared_error: 2.9559e-04\n",
      "Epoch 453/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0072 - accuracy: 0.9983 - mean_squared_error: 1.9465e-04\n",
      "Epoch 454/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0101 - accuracy: 0.9967 - mean_squared_error: 2.6355e-04\n",
      "Epoch 455/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0100 - accuracy: 0.9970 - mean_squared_error: 3.0362e-04\n",
      "Epoch 456/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0139 - accuracy: 0.9964 - mean_squared_error: 3.5016e-04\n",
      "Epoch 457/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0143 - accuracy: 0.9968 - mean_squared_error: 3.3062e-04\n",
      "Epoch 458/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0336 - accuracy: 0.9911 - mean_squared_error: 7.3128e-04\n",
      "Epoch 459/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0102 - accuracy: 0.9975 - mean_squared_error: 2.5572e-04\n",
      "Epoch 460/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0073 - accuracy: 0.9980 - mean_squared_error: 2.1189e-04\n",
      "Epoch 461/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0068 - accuracy: 0.9994 - mean_squared_error: 1.1249e-04\n",
      "Epoch 462/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0037 - accuracy: 0.9978 - mean_squared_error: 1.3103e-04\n",
      "Epoch 463/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0062 - accuracy: 0.9972 - mean_squared_error: 2.0676e-04\n",
      "Epoch 464/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0127 - accuracy: 0.9958 - mean_squared_error: 3.7977e-04\n",
      "Epoch 465/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0099 - accuracy: 0.9974 - mean_squared_error: 2.6600e-04\n",
      "Epoch 466/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0031 - accuracy: 0.9988 - mean_squared_error: 9.6899e-05\n",
      "Epoch 467/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0046 - accuracy: 0.9980 - mean_squared_error: 1.3773e-04\n",
      "Epoch 468/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0010 - accuracy: 0.9993 - mean_squared_error: 3.3156e-05\n",
      "Epoch 469/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0067 - accuracy: 0.9979 - mean_squared_error: 1.7071e-04\n",
      "Epoch 470/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0028 - accuracy: 0.9985 - mean_squared_error: 9.8331e-05\n",
      "Epoch 471/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0045 - accuracy: 0.9992 - mean_squared_error: 1.2129e-04\n",
      "Epoch 472/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 5.1717e-04 - accuracy: 1.0000 - mean_squared_error: 6.5665e-06\n",
      "Epoch 473/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 1.9238e-04 - accuracy: 1.0000 - mean_squared_error: 1.1114e-06\n",
      "Epoch 474/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0035 - accuracy: 0.9988 - mean_squared_error: 1.1830e-04\n",
      "Epoch 475/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0052 - accuracy: 0.9984 - mean_squared_error: 1.5838e-04\n",
      "Epoch 476/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0168 - accuracy: 0.9946 - mean_squared_error: 3.9500e-04\n",
      "Epoch 477/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0179 - accuracy: 0.9934 - mean_squared_error: 5.0332e-04\n",
      "Epoch 478/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0267 - accuracy: 0.9943 - mean_squared_error: 5.3617e-04\n",
      "Epoch 479/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0228 - accuracy: 0.9941 - mean_squared_error: 4.5079e-04\n",
      "Epoch 480/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0156 - accuracy: 0.9936 - mean_squared_error: 5.3217e-04\n",
      "Epoch 481/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0208 - accuracy: 0.9943 - mean_squared_error: 5.6511e-04\n",
      "Epoch 482/600\n",
      "50/50 [==============================] - 24s 481ms/step - loss: 0.0187 - accuracy: 0.9940 - mean_squared_error: 4.9618e-04\n",
      "Epoch 483/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0055 - accuracy: 0.9984 - mean_squared_error: 1.5955e-04\n",
      "Epoch 484/600\n",
      "50/50 [==============================] - 24s 481ms/step - loss: 0.0077 - accuracy: 0.9978 - mean_squared_error: 2.3559e-04\n",
      "Epoch 485/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0125 - accuracy: 0.9958 - mean_squared_error: 3.6702e-04\n",
      "Epoch 486/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0144 - accuracy: 0.9958 - mean_squared_error: 3.8450e-04\n",
      "Epoch 487/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0073 - accuracy: 0.9969 - mean_squared_error: 2.3285e-04\n",
      "Epoch 488/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0044 - accuracy: 0.9996 - mean_squared_error: 8.1780e-05\n",
      "Epoch 489/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0024 - accuracy: 0.9993 - mean_squared_error: 7.2140e-05\n",
      "Epoch 490/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0028 - accuracy: 0.9990 - mean_squared_error: 8.9415e-05\n",
      "Epoch 491/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0073 - accuracy: 0.9974 - mean_squared_error: 2.3716e-04\n",
      "Epoch 492/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0065 - accuracy: 0.9983 - mean_squared_error: 1.7288e-04\n",
      "Epoch 493/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0181 - accuracy: 0.9937 - mean_squared_error: 4.5943e-04\n",
      "Epoch 494/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0077 - accuracy: 0.9955 - mean_squared_error: 2.8061e-04\n",
      "Epoch 495/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0089 - accuracy: 0.9960 - mean_squared_error: 2.8139e-04\n",
      "Epoch 496/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0201 - accuracy: 0.9916 - mean_squared_error: 6.0929e-04\n",
      "Epoch 497/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0108 - accuracy: 0.9982 - mean_squared_error: 2.2385e-04\n",
      "Epoch 498/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0079 - accuracy: 0.9965 - mean_squared_error: 2.5169e-04\n",
      "Epoch 499/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0186 - accuracy: 0.9952 - mean_squared_error: 4.5534e-04\n",
      "Epoch 500/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0129 - accuracy: 0.9973 - mean_squared_error: 2.9382e-04\n",
      "Epoch 501/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0090 - accuracy: 0.9985 - mean_squared_error: 1.7065e-04\n",
      "Epoch 502/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0026 - accuracy: 0.9990 - mean_squared_error: 7.1098e-05\n",
      "Epoch 503/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0250 - accuracy: 0.9951 - mean_squared_error: 4.3574e-04\n",
      "Epoch 504/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0142 - accuracy: 0.9969 - mean_squared_error: 3.4003e-04\n",
      "Epoch 505/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0159 - accuracy: 0.9935 - mean_squared_error: 5.2971e-04\n",
      "Epoch 506/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0212 - accuracy: 0.9928 - mean_squared_error: 6.6923e-04\n",
      "Epoch 507/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0036 - accuracy: 0.9993 - mean_squared_error: 8.5329e-05\n",
      "Epoch 508/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0140 - accuracy: 0.9932 - mean_squared_error: 5.2868e-04\n",
      "Epoch 509/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0044 - accuracy: 0.9987 - mean_squared_error: 1.3622e-04\n",
      "Epoch 510/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0063 - accuracy: 0.9963 - mean_squared_error: 2.4434e-04\n",
      "Epoch 511/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0040 - accuracy: 0.9989 - mean_squared_error: 1.2811e-04\n",
      "Epoch 512/600\n",
      "50/50 [==============================] - 24s 471ms/step - loss: 0.0022 - accuracy: 0.9997 - mean_squared_error: 5.8455e-05\n",
      "Epoch 513/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0026 - accuracy: 0.9987 - mean_squared_error: 9.1547e-05\n",
      "Epoch 514/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0089 - accuracy: 0.9977 - mean_squared_error: 2.1303e-04\n",
      "Epoch 515/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0104 - accuracy: 0.9971 - mean_squared_error: 2.9505e-04\n",
      "Epoch 516/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0094 - accuracy: 0.9967 - mean_squared_error: 2.4077e-04\n",
      "Epoch 517/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0179 - accuracy: 0.9950 - mean_squared_error: 4.9248e-04\n",
      "Epoch 518/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0148 - accuracy: 0.9952 - mean_squared_error: 4.3235e-04\n",
      "Epoch 519/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0106 - accuracy: 0.9957 - mean_squared_error: 3.2623e-04\n",
      "Epoch 520/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0109 - accuracy: 0.9955 - mean_squared_error: 3.2751e-04\n",
      "Epoch 521/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0033 - accuracy: 0.9997 - mean_squared_error: 8.3736e-05\n",
      "Epoch 522/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0027 - accuracy: 0.9983 - mean_squared_error: 8.5335e-05\n",
      "Epoch 523/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0096 - accuracy: 0.9961 - mean_squared_error: 3.2094e-04\n",
      "Epoch 524/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0331 - accuracy: 0.9923 - mean_squared_error: 7.5063e-04\n",
      "Epoch 525/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0346 - accuracy: 0.9917 - mean_squared_error: 8.0577e-04\n",
      "Epoch 526/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0206 - accuracy: 0.9948 - mean_squared_error: 4.9257e-04\n",
      "Epoch 527/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0179 - accuracy: 0.9969 - mean_squared_error: 3.5352e-04\n",
      "Epoch 528/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0124 - accuracy: 0.9954 - mean_squared_error: 4.0602e-04\n",
      "Epoch 529/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0107 - accuracy: 0.9961 - mean_squared_error: 3.5107e-04\n",
      "Epoch 530/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0032 - accuracy: 0.9990 - mean_squared_error: 1.0200e-04\n",
      "Epoch 531/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0050 - accuracy: 0.9989 - mean_squared_error: 1.1383e-04\n",
      "Epoch 532/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0121 - accuracy: 0.9957 - mean_squared_error: 2.8979e-04\n",
      "Epoch 533/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0014 - accuracy: 0.9998 - mean_squared_error: 3.1076e-05\n",
      "Epoch 534/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0027 - accuracy: 0.9995 - mean_squared_error: 7.9741e-05\n",
      "Epoch 535/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0010 - accuracy: 1.0000 - mean_squared_error: 2.5963e-05\n",
      "Epoch 536/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 5.8969e-04 - accuracy: 1.0000 - mean_squared_error: 9.4908e-06\n",
      "Epoch 537/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 6.2743e-04 - accuracy: 0.9997 - mean_squared_error: 1.7665e-05\n",
      "Epoch 538/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0033 - accuracy: 0.9988 - mean_squared_error: 1.1652e-04\n",
      "Epoch 539/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0159 - accuracy: 0.9973 - mean_squared_error: 3.2353e-04\n",
      "Epoch 540/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0071 - accuracy: 0.9961 - mean_squared_error: 2.4156e-04\n",
      "Epoch 541/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0077 - accuracy: 0.9985 - mean_squared_error: 1.6428e-04\n",
      "Epoch 542/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0020 - accuracy: 0.9994 - mean_squared_error: 5.5252e-05\n",
      "Epoch 543/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0023 - accuracy: 0.9989 - mean_squared_error: 6.0603e-05\n",
      "Epoch 544/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 7.7837e-04 - accuracy: 0.9999 - mean_squared_error: 1.4506e-05\n",
      "Epoch 545/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0058 - accuracy: 0.9987 - mean_squared_error: 1.5389e-04\n",
      "Epoch 546/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0173 - accuracy: 0.9961 - mean_squared_error: 4.1580e-04\n",
      "Epoch 547/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0041 - accuracy: 0.9990 - mean_squared_error: 1.2591e-04\n",
      "Epoch 548/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0056 - accuracy: 0.9982 - mean_squared_error: 1.6581e-04\n",
      "Epoch 549/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0163 - accuracy: 0.9941 - mean_squared_error: 5.2142e-04\n",
      "Epoch 550/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0149 - accuracy: 0.9942 - mean_squared_error: 4.1520e-04\n",
      "Epoch 551/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0047 - accuracy: 0.9984 - mean_squared_error: 1.4635e-04\n",
      "Epoch 552/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0205 - accuracy: 0.9925 - mean_squared_error: 6.8643e-04\n",
      "Epoch 553/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0124 - accuracy: 0.9950 - mean_squared_error: 3.9194e-04\n",
      "Epoch 554/600\n",
      "50/50 [==============================] - 24s 478ms/step - loss: 0.0291 - accuracy: 0.9925 - mean_squared_error: 7.3799e-04\n",
      "Epoch 555/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0232 - accuracy: 0.9944 - mean_squared_error: 5.5996e-04\n",
      "Epoch 556/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0142 - accuracy: 0.9957 - mean_squared_error: 3.7933e-04\n",
      "Epoch 557/600\n",
      "50/50 [==============================] - 24s 480ms/step - loss: 0.0115 - accuracy: 0.9958 - mean_squared_error: 3.8419e-04\n",
      "Epoch 558/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0059 - accuracy: 0.9981 - mean_squared_error: 1.4519e-04\n",
      "Epoch 559/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0070 - accuracy: 0.9989 - mean_squared_error: 1.2069e-04\n",
      "Epoch 560/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0053 - accuracy: 0.9979 - mean_squared_error: 1.6713e-04\n",
      "Epoch 561/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0029 - accuracy: 0.9992 - mean_squared_error: 8.5604e-05\n",
      "Epoch 562/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0187 - accuracy: 0.9956 - mean_squared_error: 3.6040e-04\n",
      "Epoch 563/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0062 - accuracy: 0.9978 - mean_squared_error: 1.8687e-04\n",
      "Epoch 564/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0022 - accuracy: 0.9988 - mean_squared_error: 7.1955e-05\n",
      "Epoch 565/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0067 - accuracy: 0.9988 - mean_squared_error: 1.3334e-04\n",
      "Epoch 566/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0022 - accuracy: 0.9988 - mean_squared_error: 6.8470e-05\n",
      "Epoch 567/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0090 - accuracy: 0.9967 - mean_squared_error: 2.6367e-04\n",
      "Epoch 568/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0119 - accuracy: 0.9953 - mean_squared_error: 3.8417e-04\n",
      "Epoch 569/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0116 - accuracy: 0.9974 - mean_squared_error: 3.1023e-04\n",
      "Epoch 570/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0318 - accuracy: 0.9947 - mean_squared_error: 4.8015e-04\n",
      "Epoch 571/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0201 - accuracy: 0.9954 - mean_squared_error: 4.1249e-04\n",
      "Epoch 572/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0173 - accuracy: 0.9930 - mean_squared_error: 5.0485e-04\n",
      "Epoch 573/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0328 - accuracy: 0.9910 - mean_squared_error: 8.1227e-04\n",
      "Epoch 574/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0118 - accuracy: 0.9967 - mean_squared_error: 2.6426e-04\n",
      "Epoch 575/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0107 - accuracy: 0.9962 - mean_squared_error: 2.8709e-04\n",
      "Epoch 576/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0102 - accuracy: 0.9970 - mean_squared_error: 1.8520e-04\n",
      "Epoch 577/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0053 - accuracy: 0.9974 - mean_squared_error: 1.6835e-04\n",
      "Epoch 578/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0177 - accuracy: 0.9942 - mean_squared_error: 4.6271e-04\n",
      "Epoch 579/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0096 - accuracy: 0.9969 - mean_squared_error: 3.0774e-04\n",
      "Epoch 580/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0017 - accuracy: 0.9992 - mean_squared_error: 5.7236e-05\n",
      "Epoch 581/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0033 - accuracy: 0.9989 - mean_squared_error: 9.4209e-05\n",
      "Epoch 582/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0091 - accuracy: 0.9966 - mean_squared_error: 2.3264e-04\n",
      "Epoch 583/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0157 - accuracy: 0.9967 - mean_squared_error: 3.2599e-04\n",
      "Epoch 584/600\n",
      "50/50 [==============================] - 24s 479ms/step - loss: 0.0014 - accuracy: 1.0000 - mean_squared_error: 2.5585e-05\n",
      "Epoch 585/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0076 - accuracy: 0.9977 - mean_squared_error: 2.0860e-04\n",
      "Epoch 586/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0076 - accuracy: 0.9974 - mean_squared_error: 2.5171e-04\n",
      "Epoch 587/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0318 - accuracy: 0.9949 - mean_squared_error: 4.7506e-04\n",
      "Epoch 588/600\n",
      "50/50 [==============================] - 24s 473ms/step - loss: 0.0104 - accuracy: 0.9963 - mean_squared_error: 3.3152e-04\n",
      "Epoch 589/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0104 - accuracy: 0.9967 - mean_squared_error: 3.0815e-04\n",
      "Epoch 590/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0027 - accuracy: 0.9997 - mean_squared_error: 5.3939e-05\n",
      "Epoch 591/600\n",
      "50/50 [==============================] - 24s 472ms/step - loss: 0.0060 - accuracy: 0.9966 - mean_squared_error: 2.1031e-04\n",
      "Epoch 592/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0033 - accuracy: 0.9993 - mean_squared_error: 7.3097e-05\n",
      "Epoch 593/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0114 - accuracy: 0.9972 - mean_squared_error: 3.0265e-04\n",
      "Epoch 594/600\n",
      "50/50 [==============================] - 24s 477ms/step - loss: 0.0025 - accuracy: 0.9998 - mean_squared_error: 3.6282e-05\n",
      "Epoch 595/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0033 - accuracy: 0.9989 - mean_squared_error: 1.0480e-04\n",
      "Epoch 596/600\n",
      "50/50 [==============================] - 24s 476ms/step - loss: 0.0039 - accuracy: 0.9986 - mean_squared_error: 1.3613e-04\n",
      "Epoch 597/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0057 - accuracy: 0.9982 - mean_squared_error: 1.1720e-04\n",
      "Epoch 598/600\n",
      "50/50 [==============================] - 24s 474ms/step - loss: 0.0180 - accuracy: 0.9952 - mean_squared_error: 3.7344e-04\n",
      "Epoch 599/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0142 - accuracy: 0.9944 - mean_squared_error: 4.2214e-04\n",
      "Epoch 600/600\n",
      "50/50 [==============================] - 24s 475ms/step - loss: 0.0126 - accuracy: 0.9967 - mean_squared_error: 3.1222e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x21c55a51988>"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x=train_batch, steps_per_epoch = 50, epochs = 600, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2 = Sequential([\n",
    "    Conv2D(filters = 32, kernel_size = (4, 4), activation = 'relu', padding = 'same', input_shape = (64, 64, 3)),\n",
    "    Conv2D(filters = 32, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (2,2), strides = 2),\n",
    "    Dropout(0.5),\n",
    "    Conv2D(filters = 64, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 64, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (2,2), strides = 2),\n",
    "    Dropout(0.5),\n",
    "    Conv2D(filters = 128, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 256, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    Conv2D(filters = 512, kernel_size = (4, 4), activation = 'relu', padding = 'same'),\n",
    "    MaxPool2D(pool_size = (2,2), strides = 2),\n",
    "    Dropout(0.5),\n",
    "    Flatten(),\n",
    "    Dense(units = 19, activation = 'softmax'),\n",
    "])\n",
    "#definitely need to extend, feels like data is being overfitted?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2.compile(optimizer = Adam(learning_rate = 0.0001), loss = \"categorical_crossentropy\", metrics = [\"accuracy\",\"mean_squared_error\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/600\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Brady\\Anaconda3\\lib\\site-packages\\PIL\\Image.py:989: UserWarning: Palette images with Transparency expressed in bytes should be converted to RGBA images\n",
      "  \"Palette images with Transparency expressed in bytes should be \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 835ms/step - loss: 11.3596 - accuracy: 0.0529 - mean_squared_error: 0.0722\n",
      "Epoch 2/600\n",
      "50/50 [==============================] - 42s 832ms/step - loss: 3.0184 - accuracy: 0.0567 - mean_squared_error: 0.0503\n",
      "Epoch 3/600\n",
      "50/50 [==============================] - 42s 833ms/step - loss: 2.9931 - accuracy: 0.0564 - mean_squared_error: 0.0502\n",
      "Epoch 4/600\n",
      "50/50 [==============================] - 42s 832ms/step - loss: 2.9607 - accuracy: 0.0619 - mean_squared_error: 0.0500\n",
      "Epoch 5/600\n",
      "50/50 [==============================] - 42s 835ms/step - loss: 2.9438 - accuracy: 0.0703 - mean_squared_error: 0.0499\n",
      "Epoch 6/600\n",
      "50/50 [==============================] - 42s 836ms/step - loss: 2.9330 - accuracy: 0.0630 - mean_squared_error: 0.0498\n",
      "Epoch 7/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 2.8856 - accuracy: 0.1075 - mean_squared_error: 0.0495\n",
      "Epoch 8/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 2.8478 - accuracy: 0.1190 - mean_squared_error: 0.0491\n",
      "Epoch 9/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 2.7784 - accuracy: 0.1562 - mean_squared_error: 0.0484\n",
      "Epoch 10/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 2.6710 - accuracy: 0.1830 - mean_squared_error: 0.0477\n",
      "Epoch 11/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 2.5877 - accuracy: 0.2066 - mean_squared_error: 0.0465\n",
      "Epoch 12/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 2.5276 - accuracy: 0.2221 - mean_squared_error: 0.0460\n",
      "Epoch 13/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 2.3896 - accuracy: 0.2814 - mean_squared_error: 0.0444\n",
      "Epoch 14/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 2.3288 - accuracy: 0.3148 - mean_squared_error: 0.0434\n",
      "Epoch 15/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 2.2767 - accuracy: 0.2971 - mean_squared_error: 0.0430\n",
      "Epoch 16/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 2.1988 - accuracy: 0.3309 - mean_squared_error: 0.0418\n",
      "Epoch 17/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 2.0965 - accuracy: 0.3586 - mean_squared_error: 0.0402\n",
      "Epoch 18/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 2.0034 - accuracy: 0.3837 - mean_squared_error: 0.0392\n",
      "Epoch 19/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 1.8901 - accuracy: 0.4204 - mean_squared_error: 0.0376\n",
      "Epoch 20/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 1.8493 - accuracy: 0.4256 - mean_squared_error: 0.0371\n",
      "Epoch 21/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 1.7659 - accuracy: 0.4676 - mean_squared_error: 0.0354\n",
      "Epoch 22/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 1.6285 - accuracy: 0.5059 - mean_squared_error: 0.0330\n",
      "Epoch 23/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 1.6186 - accuracy: 0.5153 - mean_squared_error: 0.0327\n",
      "Epoch 24/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 1.4847 - accuracy: 0.5624 - mean_squared_error: 0.0303\n",
      "Epoch 25/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 1.3731 - accuracy: 0.5798 - mean_squared_error: 0.0288\n",
      "Epoch 26/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 1.3683 - accuracy: 0.5926 - mean_squared_error: 0.0289\n",
      "Epoch 27/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 1.2458 - accuracy: 0.6262 - mean_squared_error: 0.0262\n",
      "Epoch 28/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 1.1999 - accuracy: 0.6485 - mean_squared_error: 0.0253\n",
      "Epoch 29/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 1.0765 - accuracy: 0.6818 - mean_squared_error: 0.0234\n",
      "Epoch 30/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 1.1192 - accuracy: 0.6649 - mean_squared_error: 0.0239\n",
      "Epoch 31/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 1.0124 - accuracy: 0.7005 - mean_squared_error: 0.0219\n",
      "Epoch 32/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.8988 - accuracy: 0.7287 - mean_squared_error: 0.0197\n",
      "Epoch 33/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.8713 - accuracy: 0.7417 - mean_squared_error: 0.0190\n",
      "Epoch 34/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.8879 - accuracy: 0.7292 - mean_squared_error: 0.0196\n",
      "Epoch 35/600\n",
      "50/50 [==============================] - 42s 835ms/step - loss: 0.8167 - accuracy: 0.7521 - mean_squared_error: 0.0182\n",
      "Epoch 36/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.8012 - accuracy: 0.7564 - mean_squared_error: 0.0176\n",
      "Epoch 37/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.6987 - accuracy: 0.8028 - mean_squared_error: 0.0152\n",
      "Epoch 38/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.6956 - accuracy: 0.7845 - mean_squared_error: 0.0151\n",
      "Epoch 39/600\n",
      "50/50 [==============================] - 42s 836ms/step - loss: 0.6486 - accuracy: 0.8102 - mean_squared_error: 0.0147\n",
      "Epoch 40/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.6558 - accuracy: 0.8027 - mean_squared_error: 0.0148\n",
      "Epoch 41/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.5918 - accuracy: 0.8194 - mean_squared_error: 0.0135\n",
      "Epoch 42/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.5512 - accuracy: 0.8245 - mean_squared_error: 0.0127\n",
      "Epoch 43/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.5047 - accuracy: 0.8421 - mean_squared_error: 0.0115\n",
      "Epoch 44/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.5203 - accuracy: 0.8415 - mean_squared_error: 0.0121\n",
      "Epoch 45/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.4725 - accuracy: 0.8557 - mean_squared_error: 0.0108\n",
      "Epoch 46/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.3784 - accuracy: 0.8820 - mean_squared_error: 0.0087\n",
      "Epoch 47/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.4743 - accuracy: 0.8493 - mean_squared_error: 0.0110\n",
      "Epoch 48/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.4509 - accuracy: 0.8683 - mean_squared_error: 0.0105\n",
      "Epoch 49/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.4366 - accuracy: 0.8676 - mean_squared_error: 0.0104\n",
      "Epoch 50/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.4384 - accuracy: 0.8601 - mean_squared_error: 0.0102\n",
      "Epoch 51/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.4395 - accuracy: 0.8609 - mean_squared_error: 0.0104\n",
      "Epoch 52/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.3568 - accuracy: 0.8850 - mean_squared_error: 0.0085\n",
      "Epoch 53/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.3651 - accuracy: 0.8886 - mean_squared_error: 0.0086\n",
      "Epoch 54/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.3081 - accuracy: 0.9066 - mean_squared_error: 0.0073\n",
      "Epoch 55/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.3111 - accuracy: 0.9050 - mean_squared_error: 0.0074\n",
      "Epoch 56/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.3250 - accuracy: 0.8948 - mean_squared_error: 0.0077\n",
      "Epoch 57/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.3547 - accuracy: 0.8971 - mean_squared_error: 0.0080\n",
      "Epoch 58/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.2973 - accuracy: 0.9020 - mean_squared_error: 0.0072\n",
      "Epoch 59/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.2577 - accuracy: 0.9212 - mean_squared_error: 0.0062\n",
      "Epoch 60/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.3116 - accuracy: 0.9041 - mean_squared_error: 0.0077\n",
      "Epoch 61/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.3340 - accuracy: 0.8961 - mean_squared_error: 0.0078\n",
      "Epoch 62/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.2606 - accuracy: 0.9200 - mean_squared_error: 0.0065\n",
      "Epoch 63/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 841ms/step - loss: 0.2861 - accuracy: 0.9110 - mean_squared_error: 0.0069\n",
      "Epoch 64/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.2456 - accuracy: 0.9256 - mean_squared_error: 0.0059\n",
      "Epoch 65/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.2528 - accuracy: 0.9203 - mean_squared_error: 0.0060\n",
      "Epoch 66/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.2479 - accuracy: 0.9286 - mean_squared_error: 0.0059\n",
      "Epoch 67/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.2386 - accuracy: 0.9272 - mean_squared_error: 0.0057\n",
      "Epoch 68/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.2841 - accuracy: 0.9177 - mean_squared_error: 0.0066\n",
      "Epoch 69/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.2013 - accuracy: 0.9358 - mean_squared_error: 0.0050\n",
      "Epoch 70/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.2265 - accuracy: 0.9332 - mean_squared_error: 0.0053\n",
      "Epoch 71/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.2583 - accuracy: 0.9317 - mean_squared_error: 0.0058\n",
      "Epoch 72/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.2647 - accuracy: 0.9210 - mean_squared_error: 0.0062\n",
      "Epoch 73/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.2191 - accuracy: 0.9299 - mean_squared_error: 0.0054\n",
      "Epoch 74/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.2112 - accuracy: 0.9314 - mean_squared_error: 0.0053\n",
      "Epoch 75/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.1733 - accuracy: 0.9412 - mean_squared_error: 0.0043\n",
      "Epoch 76/600\n",
      "50/50 [==============================] - 42s 836ms/step - loss: 0.1999 - accuracy: 0.9369 - mean_squared_error: 0.0049\n",
      "Epoch 77/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.2098 - accuracy: 0.9366 - mean_squared_error: 0.0051\n",
      "Epoch 78/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.2404 - accuracy: 0.9215 - mean_squared_error: 0.0060\n",
      "Epoch 79/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1624 - accuracy: 0.9557 - mean_squared_error: 0.0040\n",
      "Epoch 80/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1591 - accuracy: 0.9555 - mean_squared_error: 0.0039\n",
      "Epoch 81/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.1675 - accuracy: 0.9509 - mean_squared_error: 0.0038\n",
      "Epoch 82/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1529 - accuracy: 0.9489 - mean_squared_error: 0.0040\n",
      "Epoch 83/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1899 - accuracy: 0.9395 - mean_squared_error: 0.0046\n",
      "Epoch 84/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1505 - accuracy: 0.9487 - mean_squared_error: 0.0038\n",
      "Epoch 85/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1504 - accuracy: 0.9444 - mean_squared_error: 0.0040\n",
      "Epoch 86/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1768 - accuracy: 0.9471 - mean_squared_error: 0.0043\n",
      "Epoch 87/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.2025 - accuracy: 0.9451 - mean_squared_error: 0.0044\n",
      "Epoch 88/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.1303 - accuracy: 0.9540 - mean_squared_error: 0.0035\n",
      "Epoch 89/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1352 - accuracy: 0.9610 - mean_squared_error: 0.0033\n",
      "Epoch 90/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1764 - accuracy: 0.9453 - mean_squared_error: 0.0044\n",
      "Epoch 91/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.1411 - accuracy: 0.9583 - mean_squared_error: 0.0033\n",
      "Epoch 92/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.1721 - accuracy: 0.9524 - mean_squared_error: 0.0041\n",
      "Epoch 93/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.1838 - accuracy: 0.9442 - mean_squared_error: 0.0045\n",
      "Epoch 94/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1724 - accuracy: 0.9459 - mean_squared_error: 0.0042\n",
      "Epoch 95/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.1714 - accuracy: 0.9479 - mean_squared_error: 0.0042\n",
      "Epoch 96/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.1406 - accuracy: 0.9509 - mean_squared_error: 0.0037\n",
      "Epoch 97/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1595 - accuracy: 0.9478 - mean_squared_error: 0.0040\n",
      "Epoch 98/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1759 - accuracy: 0.9458 - mean_squared_error: 0.0042\n",
      "Epoch 99/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.1183 - accuracy: 0.9696 - mean_squared_error: 0.0028\n",
      "Epoch 100/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.1386 - accuracy: 0.9560 - mean_squared_error: 0.0034\n",
      "Epoch 101/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1160 - accuracy: 0.9604 - mean_squared_error: 0.0031\n",
      "Epoch 102/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.1433 - accuracy: 0.9496 - mean_squared_error: 0.0036\n",
      "Epoch 103/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1177 - accuracy: 0.9624 - mean_squared_error: 0.0029\n",
      "Epoch 104/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.1551 - accuracy: 0.9558 - mean_squared_error: 0.0036\n",
      "Epoch 105/600\n",
      "50/50 [==============================] - 43s 853ms/step - loss: 0.1373 - accuracy: 0.9524 - mean_squared_error: 0.0036\n",
      "Epoch 106/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0950 - accuracy: 0.9729 - mean_squared_error: 0.0023\n",
      "Epoch 107/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.1157 - accuracy: 0.9642 - mean_squared_error: 0.0026\n",
      "Epoch 108/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.1193 - accuracy: 0.9674 - mean_squared_error: 0.0028\n",
      "Epoch 109/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.1295 - accuracy: 0.9541 - mean_squared_error: 0.0035\n",
      "Epoch 110/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0920 - accuracy: 0.9756 - mean_squared_error: 0.0022\n",
      "Epoch 111/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0898 - accuracy: 0.9749 - mean_squared_error: 0.0023\n",
      "Epoch 112/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.1171 - accuracy: 0.9644 - mean_squared_error: 0.0028\n",
      "Epoch 113/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1236 - accuracy: 0.9636 - mean_squared_error: 0.0030\n",
      "Epoch 114/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.1178 - accuracy: 0.9664 - mean_squared_error: 0.0028\n",
      "Epoch 115/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0863 - accuracy: 0.9749 - mean_squared_error: 0.0022\n",
      "Epoch 116/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1038 - accuracy: 0.9664 - mean_squared_error: 0.0026\n",
      "Epoch 117/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1276 - accuracy: 0.9579 - mean_squared_error: 0.0032\n",
      "Epoch 118/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0867 - accuracy: 0.9678 - mean_squared_error: 0.0024\n",
      "Epoch 119/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0929 - accuracy: 0.9695 - mean_squared_error: 0.0025\n",
      "Epoch 120/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1152 - accuracy: 0.9575 - mean_squared_error: 0.0032\n",
      "Epoch 121/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0956 - accuracy: 0.9639 - mean_squared_error: 0.0026\n",
      "Epoch 122/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.1113 - accuracy: 0.9659 - mean_squared_error: 0.0026\n",
      "Epoch 123/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0932 - accuracy: 0.9701 - mean_squared_error: 0.0023\n",
      "Epoch 124/600\n",
      "50/50 [==============================] - 43s 852ms/step - loss: 0.1008 - accuracy: 0.9683 - mean_squared_error: 0.0024\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 125/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0885 - accuracy: 0.9704 - mean_squared_error: 0.0022\n",
      "Epoch 126/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0697 - accuracy: 0.9756 - mean_squared_error: 0.0018\n",
      "Epoch 127/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.1085 - accuracy: 0.9648 - mean_squared_error: 0.0028\n",
      "Epoch 128/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0879 - accuracy: 0.9704 - mean_squared_error: 0.0023\n",
      "Epoch 129/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0847 - accuracy: 0.9701 - mean_squared_error: 0.0022\n",
      "Epoch 130/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0971 - accuracy: 0.9661 - mean_squared_error: 0.0026\n",
      "Epoch 131/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0927 - accuracy: 0.9713 - mean_squared_error: 0.0023\n",
      "Epoch 132/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0776 - accuracy: 0.9776 - mean_squared_error: 0.0019\n",
      "Epoch 133/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0801 - accuracy: 0.9737 - mean_squared_error: 0.0021\n",
      "Epoch 134/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0887 - accuracy: 0.9757 - mean_squared_error: 0.0021\n",
      "Epoch 135/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.1207 - accuracy: 0.9633 - mean_squared_error: 0.0030\n",
      "Epoch 136/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0866 - accuracy: 0.9738 - mean_squared_error: 0.0022\n",
      "Epoch 137/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0701 - accuracy: 0.9764 - mean_squared_error: 0.0019\n",
      "Epoch 138/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0791 - accuracy: 0.9754 - mean_squared_error: 0.0020\n",
      "Epoch 139/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0943 - accuracy: 0.9729 - mean_squared_error: 0.0023\n",
      "Epoch 140/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0929 - accuracy: 0.9683 - mean_squared_error: 0.0026\n",
      "Epoch 141/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0842 - accuracy: 0.9741 - mean_squared_error: 0.0021\n",
      "Epoch 142/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0732 - accuracy: 0.9776 - mean_squared_error: 0.0019\n",
      "Epoch 143/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0919 - accuracy: 0.9711 - mean_squared_error: 0.0025\n",
      "Epoch 144/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0858 - accuracy: 0.9707 - mean_squared_error: 0.0023\n",
      "Epoch 145/600\n",
      "50/50 [==============================] - 43s 851ms/step - loss: 0.0902 - accuracy: 0.9696 - mean_squared_error: 0.0023\n",
      "Epoch 146/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0816 - accuracy: 0.9746 - mean_squared_error: 0.0022\n",
      "Epoch 147/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.1020 - accuracy: 0.9670 - mean_squared_error: 0.0026\n",
      "Epoch 148/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0861 - accuracy: 0.9734 - mean_squared_error: 0.0022\n",
      "Epoch 149/600\n",
      "50/50 [==============================] - 42s 849ms/step - loss: 0.0633 - accuracy: 0.9797 - mean_squared_error: 0.0017\n",
      "Epoch 150/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0611 - accuracy: 0.9824 - mean_squared_error: 0.0015\n",
      "Epoch 151/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0743 - accuracy: 0.9720 - mean_squared_error: 0.0020\n",
      "Epoch 152/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0880 - accuracy: 0.9722 - mean_squared_error: 0.0022\n",
      "Epoch 153/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0747 - accuracy: 0.9718 - mean_squared_error: 0.0021\n",
      "Epoch 154/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0752 - accuracy: 0.9778 - mean_squared_error: 0.0019\n",
      "Epoch 155/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0909 - accuracy: 0.9709 - mean_squared_error: 0.0023\n",
      "Epoch 156/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0759 - accuracy: 0.9774 - mean_squared_error: 0.0019\n",
      "Epoch 157/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0674 - accuracy: 0.9813 - mean_squared_error: 0.0017\n",
      "Epoch 158/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0740 - accuracy: 0.9774 - mean_squared_error: 0.0020\n",
      "Epoch 159/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.1072 - accuracy: 0.9721 - mean_squared_error: 0.0023\n",
      "Epoch 160/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0820 - accuracy: 0.9721 - mean_squared_error: 0.0021\n",
      "Epoch 161/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0655 - accuracy: 0.9797 - mean_squared_error: 0.0016\n",
      "Epoch 162/600\n",
      "50/50 [==============================] - 42s 849ms/step - loss: 0.0811 - accuracy: 0.9768 - mean_squared_error: 0.0020\n",
      "Epoch 163/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0701 - accuracy: 0.9795 - mean_squared_error: 0.0019\n",
      "Epoch 164/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0727 - accuracy: 0.9773 - mean_squared_error: 0.0018\n",
      "Epoch 165/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0498 - accuracy: 0.9817 - mean_squared_error: 0.0014\n",
      "Epoch 166/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0618 - accuracy: 0.9806 - mean_squared_error: 0.0016\n",
      "Epoch 167/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0562 - accuracy: 0.9798 - mean_squared_error: 0.0015\n",
      "Epoch 168/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0568 - accuracy: 0.9803 - mean_squared_error: 0.0014\n",
      "Epoch 169/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0706 - accuracy: 0.9763 - mean_squared_error: 0.0020\n",
      "Epoch 170/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0805 - accuracy: 0.9721 - mean_squared_error: 0.0019\n",
      "Epoch 171/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0728 - accuracy: 0.9806 - mean_squared_error: 0.0017\n",
      "Epoch 172/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0539 - accuracy: 0.9833 - mean_squared_error: 0.0013\n",
      "Epoch 173/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0523 - accuracy: 0.9850 - mean_squared_error: 0.0012\n",
      "Epoch 174/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0816 - accuracy: 0.9691 - mean_squared_error: 0.0022\n",
      "Epoch 175/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0527 - accuracy: 0.9826 - mean_squared_error: 0.0014\n",
      "Epoch 176/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0717 - accuracy: 0.9775 - mean_squared_error: 0.0019\n",
      "Epoch 177/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0511 - accuracy: 0.9810 - mean_squared_error: 0.0015\n",
      "Epoch 178/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0630 - accuracy: 0.9791 - mean_squared_error: 0.0016\n",
      "Epoch 179/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0653 - accuracy: 0.9753 - mean_squared_error: 0.0018\n",
      "Epoch 180/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0692 - accuracy: 0.9801 - mean_squared_error: 0.0017\n",
      "Epoch 181/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0677 - accuracy: 0.9771 - mean_squared_error: 0.0019\n",
      "Epoch 182/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0531 - accuracy: 0.9854 - mean_squared_error: 0.0014\n",
      "Epoch 183/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0576 - accuracy: 0.9805 - mean_squared_error: 0.0015\n",
      "Epoch 184/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0894 - accuracy: 0.9750 - mean_squared_error: 0.0021\n",
      "Epoch 185/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0730 - accuracy: 0.9762 - mean_squared_error: 0.0021\n",
      "Epoch 186/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0594 - accuracy: 0.9839 - mean_squared_error: 0.0014\n",
      "Epoch 187/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0731 - accuracy: 0.9801 - mean_squared_error: 0.0018\n",
      "Epoch 188/600\n",
      "50/50 [==============================] - 43s 850ms/step - loss: 0.0609 - accuracy: 0.9864 - mean_squared_error: 0.0014\n",
      "Epoch 189/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0557 - accuracy: 0.9823 - mean_squared_error: 0.0014\n",
      "Epoch 190/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0546 - accuracy: 0.9825 - mean_squared_error: 0.0015\n",
      "Epoch 191/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0594 - accuracy: 0.9818 - mean_squared_error: 0.0015\n",
      "Epoch 192/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0622 - accuracy: 0.9780 - mean_squared_error: 0.0017\n",
      "Epoch 193/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0448 - accuracy: 0.9861 - mean_squared_error: 0.0012\n",
      "Epoch 194/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0628 - accuracy: 0.9829 - mean_squared_error: 0.0016\n",
      "Epoch 195/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0405 - accuracy: 0.9858 - mean_squared_error: 0.0011\n",
      "Epoch 196/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0642 - accuracy: 0.9777 - mean_squared_error: 0.0017\n",
      "Epoch 197/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0459 - accuracy: 0.9838 - mean_squared_error: 0.0011\n",
      "Epoch 198/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0470 - accuracy: 0.9836 - mean_squared_error: 0.0012\n",
      "Epoch 199/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0781 - accuracy: 0.9751 - mean_squared_error: 0.0020\n",
      "Epoch 200/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0452 - accuracy: 0.9853 - mean_squared_error: 0.0012\n",
      "Epoch 201/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0725 - accuracy: 0.9833 - mean_squared_error: 0.0015\n",
      "Epoch 202/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0576 - accuracy: 0.9800 - mean_squared_error: 0.0015\n",
      "Epoch 203/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0796 - accuracy: 0.9805 - mean_squared_error: 0.0019\n",
      "Epoch 204/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0901 - accuracy: 0.9791 - mean_squared_error: 0.0019\n",
      "Epoch 205/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0379 - accuracy: 0.9876 - mean_squared_error: 0.0010\n",
      "Epoch 206/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0355 - accuracy: 0.9876 - mean_squared_error: 9.8294e-04\n",
      "Epoch 207/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0505 - accuracy: 0.9845 - mean_squared_error: 0.0012\n",
      "Epoch 208/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0563 - accuracy: 0.9826 - mean_squared_error: 0.0014\n",
      "Epoch 209/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0416 - accuracy: 0.9848 - mean_squared_error: 0.0011\n",
      "Epoch 210/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0537 - accuracy: 0.9815 - mean_squared_error: 0.0014\n",
      "Epoch 211/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0455 - accuracy: 0.9845 - mean_squared_error: 0.0012\n",
      "Epoch 212/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0567 - accuracy: 0.9811 - mean_squared_error: 0.0015\n",
      "Epoch 213/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0504 - accuracy: 0.9864 - mean_squared_error: 0.0013\n",
      "Epoch 214/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0501 - accuracy: 0.9841 - mean_squared_error: 0.0013\n",
      "Epoch 215/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0580 - accuracy: 0.9816 - mean_squared_error: 0.0016\n",
      "Epoch 216/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0468 - accuracy: 0.9863 - mean_squared_error: 0.0011\n",
      "Epoch 217/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0476 - accuracy: 0.9847 - mean_squared_error: 0.0011\n",
      "Epoch 218/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0398 - accuracy: 0.9897 - mean_squared_error: 9.7317e-04\n",
      "Epoch 219/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0553 - accuracy: 0.9824 - mean_squared_error: 0.0015\n",
      "Epoch 220/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0491 - accuracy: 0.9809 - mean_squared_error: 0.0014\n",
      "Epoch 221/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0535 - accuracy: 0.9848 - mean_squared_error: 0.0013\n",
      "Epoch 222/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0581 - accuracy: 0.9817 - mean_squared_error: 0.0014\n",
      "Epoch 223/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0575 - accuracy: 0.9816 - mean_squared_error: 0.0015\n",
      "Epoch 224/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0677 - accuracy: 0.9816 - mean_squared_error: 0.0015\n",
      "Epoch 225/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0349 - accuracy: 0.9885 - mean_squared_error: 9.1864e-04\n",
      "Epoch 226/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0512 - accuracy: 0.9851 - mean_squared_error: 0.0013\n",
      "Epoch 227/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0395 - accuracy: 0.9883 - mean_squared_error: 0.0010\n",
      "Epoch 228/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0380 - accuracy: 0.9860 - mean_squared_error: 9.4943e-04\n",
      "Epoch 229/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0381 - accuracy: 0.9886 - mean_squared_error: 9.8444e-04\n",
      "Epoch 230/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0375 - accuracy: 0.9857 - mean_squared_error: 0.0011\n",
      "Epoch 231/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0631 - accuracy: 0.9817 - mean_squared_error: 0.0015\n",
      "Epoch 232/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0429 - accuracy: 0.9872 - mean_squared_error: 0.0011\n",
      "Epoch 233/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0341 - accuracy: 0.9888 - mean_squared_error: 9.1913e-04\n",
      "Epoch 234/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0416 - accuracy: 0.9888 - mean_squared_error: 0.0010\n",
      "Epoch 235/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0492 - accuracy: 0.9864 - mean_squared_error: 0.0011\n",
      "Epoch 236/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0722 - accuracy: 0.9746 - mean_squared_error: 0.0020\n",
      "Epoch 237/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0605 - accuracy: 0.9837 - mean_squared_error: 0.0014\n",
      "Epoch 238/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0563 - accuracy: 0.9791 - mean_squared_error: 0.0016\n",
      "Epoch 239/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0606 - accuracy: 0.9834 - mean_squared_error: 0.0015\n",
      "Epoch 240/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0444 - accuracy: 0.9874 - mean_squared_error: 0.0011\n",
      "Epoch 241/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0340 - accuracy: 0.9898 - mean_squared_error: 7.5471e-04\n",
      "Epoch 242/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0393 - accuracy: 0.9857 - mean_squared_error: 0.0011\n",
      "Epoch 243/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0396 - accuracy: 0.9899 - mean_squared_error: 9.4515e-04\n",
      "Epoch 244/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0465 - accuracy: 0.9848 - mean_squared_error: 0.0013\n",
      "Epoch 245/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0394 - accuracy: 0.9876 - mean_squared_error: 0.0010\n",
      "Epoch 246/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0363 - accuracy: 0.9887 - mean_squared_error: 9.3899e-04\n",
      "Epoch 247/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 849ms/step - loss: 0.0239 - accuracy: 0.9922 - mean_squared_error: 5.9670e-04\n",
      "Epoch 248/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0484 - accuracy: 0.9872 - mean_squared_error: 0.0011\n",
      "Epoch 249/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0458 - accuracy: 0.9837 - mean_squared_error: 0.0013\n",
      "Epoch 250/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0311 - accuracy: 0.9925 - mean_squared_error: 7.5097e-04\n",
      "Epoch 251/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0560 - accuracy: 0.9832 - mean_squared_error: 0.0015\n",
      "Epoch 252/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0353 - accuracy: 0.9875 - mean_squared_error: 9.2798e-04\n",
      "Epoch 253/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0690 - accuracy: 0.9879 - mean_squared_error: 0.0013\n",
      "Epoch 254/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0496 - accuracy: 0.9880 - mean_squared_error: 0.0012\n",
      "Epoch 255/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0266 - accuracy: 0.9909 - mean_squared_error: 7.1361e-04\n",
      "Epoch 256/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0256 - accuracy: 0.9921 - mean_squared_error: 6.6089e-04\n",
      "Epoch 257/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0415 - accuracy: 0.9865 - mean_squared_error: 0.0011\n",
      "Epoch 258/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0272 - accuracy: 0.9926 - mean_squared_error: 6.4701e-04\n",
      "Epoch 259/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0505 - accuracy: 0.9855 - mean_squared_error: 0.0011\n",
      "Epoch 260/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0427 - accuracy: 0.9879 - mean_squared_error: 9.7161e-04\n",
      "Epoch 261/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0398 - accuracy: 0.9895 - mean_squared_error: 0.0010\n",
      "Epoch 262/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0379 - accuracy: 0.9882 - mean_squared_error: 9.5978e-04\n",
      "Epoch 263/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0280 - accuracy: 0.9920 - mean_squared_error: 6.8762e-04\n",
      "Epoch 264/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0383 - accuracy: 0.9865 - mean_squared_error: 0.0011\n",
      "Epoch 265/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0357 - accuracy: 0.9902 - mean_squared_error: 8.3255e-04\n",
      "Epoch 266/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0444 - accuracy: 0.9876 - mean_squared_error: 0.0010\n",
      "Epoch 267/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0601 - accuracy: 0.9859 - mean_squared_error: 0.0013\n",
      "Epoch 268/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0339 - accuracy: 0.9874 - mean_squared_error: 9.7929e-04\n",
      "Epoch 269/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0451 - accuracy: 0.9881 - mean_squared_error: 0.0010\n",
      "Epoch 270/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0438 - accuracy: 0.9809 - mean_squared_error: 0.0013\n",
      "Epoch 271/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0322 - accuracy: 0.9913 - mean_squared_error: 7.8822e-04\n",
      "Epoch 272/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0370 - accuracy: 0.9894 - mean_squared_error: 9.6216e-04\n",
      "Epoch 273/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0383 - accuracy: 0.9863 - mean_squared_error: 0.0010\n",
      "Epoch 274/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0333 - accuracy: 0.9908 - mean_squared_error: 7.9643e-04\n",
      "Epoch 275/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0389 - accuracy: 0.9902 - mean_squared_error: 8.7440e-04\n",
      "Epoch 276/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0280 - accuracy: 0.9905 - mean_squared_error: 7.8918e-04\n",
      "Epoch 277/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0338 - accuracy: 0.9897 - mean_squared_error: 8.6830e-04\n",
      "Epoch 278/600\n",
      "50/50 [==============================] - 42s 835ms/step - loss: 0.0475 - accuracy: 0.9849 - mean_squared_error: 0.0013\n",
      "Epoch 279/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0457 - accuracy: 0.9861 - mean_squared_error: 0.0011\n",
      "Epoch 280/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0314 - accuracy: 0.9914 - mean_squared_error: 7.8052e-04\n",
      "Epoch 281/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0509 - accuracy: 0.9875 - mean_squared_error: 0.0012\n",
      "Epoch 282/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0199 - accuracy: 0.9958 - mean_squared_error: 4.5453e-04\n",
      "Epoch 283/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0264 - accuracy: 0.9926 - mean_squared_error: 6.7540e-04\n",
      "Epoch 284/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0379 - accuracy: 0.9878 - mean_squared_error: 9.5976e-04\n",
      "Epoch 285/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0400 - accuracy: 0.9862 - mean_squared_error: 0.0010\n",
      "Epoch 286/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0523 - accuracy: 0.9808 - mean_squared_error: 0.0015\n",
      "Epoch 287/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0427 - accuracy: 0.9850 - mean_squared_error: 0.0012\n",
      "Epoch 288/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0286 - accuracy: 0.9917 - mean_squared_error: 7.3138e-04\n",
      "Epoch 289/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0470 - accuracy: 0.9872 - mean_squared_error: 0.0011\n",
      "Epoch 290/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0251 - accuracy: 0.9946 - mean_squared_error: 5.9120e-04\n",
      "Epoch 291/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0292 - accuracy: 0.9911 - mean_squared_error: 7.0314e-04\n",
      "Epoch 292/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0383 - accuracy: 0.9885 - mean_squared_error: 0.0010\n",
      "Epoch 293/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0303 - accuracy: 0.9900 - mean_squared_error: 8.1810e-04\n",
      "Epoch 294/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0454 - accuracy: 0.9838 - mean_squared_error: 0.0013\n",
      "Epoch 295/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0390 - accuracy: 0.9894 - mean_squared_error: 9.8589e-04\n",
      "Epoch 296/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0401 - accuracy: 0.9895 - mean_squared_error: 0.0010\n",
      "Epoch 297/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0363 - accuracy: 0.9884 - mean_squared_error: 9.4397e-04\n",
      "Epoch 298/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0384 - accuracy: 0.9894 - mean_squared_error: 9.8405e-04\n",
      "Epoch 299/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0445 - accuracy: 0.9898 - mean_squared_error: 9.1984e-04\n",
      "Epoch 300/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0552 - accuracy: 0.9855 - mean_squared_error: 0.0012\n",
      "Epoch 301/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0211 - accuracy: 0.9907 - mean_squared_error: 6.0651e-04\n",
      "Epoch 302/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0325 - accuracy: 0.9888 - mean_squared_error: 9.3147e-04\n",
      "Epoch 303/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0350 - accuracy: 0.9859 - mean_squared_error: 9.4992e-04\n",
      "Epoch 304/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0380 - accuracy: 0.9877 - mean_squared_error: 0.0010\n",
      "Epoch 305/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0362 - accuracy: 0.9873 - mean_squared_error: 9.5050e-04\n",
      "Epoch 306/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0384 - accuracy: 0.9877 - mean_squared_error: 0.0010\n",
      "Epoch 307/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0384 - accuracy: 0.9868 - mean_squared_error: 0.0011\n",
      "Epoch 308/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0481 - accuracy: 0.9862 - mean_squared_error: 0.0011\n",
      "Epoch 309/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0196 - accuracy: 0.9927 - mean_squared_error: 5.3980e-04\n",
      "Epoch 310/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0357 - accuracy: 0.9931 - mean_squared_error: 7.7550e-04\n",
      "Epoch 311/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0285 - accuracy: 0.9898 - mean_squared_error: 7.1991e-04\n",
      "Epoch 312/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0341 - accuracy: 0.9864 - mean_squared_error: 0.0010\n",
      "Epoch 313/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0316 - accuracy: 0.9857 - mean_squared_error: 0.0010\n",
      "Epoch 314/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0458 - accuracy: 0.9862 - mean_squared_error: 0.0012\n",
      "Epoch 315/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0330 - accuracy: 0.9908 - mean_squared_error: 8.1830e-04\n",
      "Epoch 316/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0271 - accuracy: 0.9891 - mean_squared_error: 7.9361e-04\n",
      "Epoch 317/600\n",
      "50/50 [==============================] - 42s 849ms/step - loss: 0.0242 - accuracy: 0.9916 - mean_squared_error: 7.0097e-04\n",
      "Epoch 318/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0344 - accuracy: 0.9905 - mean_squared_error: 9.0251e-04\n",
      "Epoch 319/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0361 - accuracy: 0.9884 - mean_squared_error: 8.4923e-04\n",
      "Epoch 320/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0397 - accuracy: 0.9869 - mean_squared_error: 0.0010\n",
      "Epoch 321/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0337 - accuracy: 0.9902 - mean_squared_error: 9.3244e-04\n",
      "Epoch 322/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0307 - accuracy: 0.9904 - mean_squared_error: 7.6876e-04\n",
      "Epoch 323/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0368 - accuracy: 0.9895 - mean_squared_error: 9.2213e-04\n",
      "Epoch 324/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0396 - accuracy: 0.9847 - mean_squared_error: 0.0012\n",
      "Epoch 325/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0249 - accuracy: 0.9904 - mean_squared_error: 6.9763e-04\n",
      "Epoch 326/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0346 - accuracy: 0.9899 - mean_squared_error: 8.2542e-04\n",
      "Epoch 327/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0313 - accuracy: 0.9877 - mean_squared_error: 9.1428e-04\n",
      "Epoch 328/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0379 - accuracy: 0.9892 - mean_squared_error: 9.5971e-04\n",
      "Epoch 329/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0395 - accuracy: 0.9885 - mean_squared_error: 9.3814e-04\n",
      "Epoch 330/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0334 - accuracy: 0.9899 - mean_squared_error: 8.0529e-04\n",
      "Epoch 331/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0223 - accuracy: 0.9930 - mean_squared_error: 6.1741e-04\n",
      "Epoch 332/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0235 - accuracy: 0.9914 - mean_squared_error: 6.9500e-04\n",
      "Epoch 333/600\n",
      "50/50 [==============================] - 42s 849ms/step - loss: 0.0312 - accuracy: 0.9884 - mean_squared_error: 9.2134e-04\n",
      "Epoch 334/600\n",
      "50/50 [==============================] - 43s 861ms/step - loss: 0.0200 - accuracy: 0.9929 - mean_squared_error: 5.3588e-04\n",
      "Epoch 335/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0338 - accuracy: 0.9888 - mean_squared_error: 9.9971e-04\n",
      "Epoch 336/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0400 - accuracy: 0.9890 - mean_squared_error: 9.8830e-04\n",
      "Epoch 337/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0352 - accuracy: 0.9889 - mean_squared_error: 0.0010\n",
      "Epoch 338/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0335 - accuracy: 0.9902 - mean_squared_error: 7.5947e-04\n",
      "Epoch 339/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0272 - accuracy: 0.9917 - mean_squared_error: 7.7099e-04\n",
      "Epoch 340/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0289 - accuracy: 0.9895 - mean_squared_error: 8.2170e-04\n",
      "Epoch 341/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0397 - accuracy: 0.9854 - mean_squared_error: 0.0011\n",
      "Epoch 342/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0308 - accuracy: 0.9907 - mean_squared_error: 7.8981e-04\n",
      "Epoch 343/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0272 - accuracy: 0.9933 - mean_squared_error: 6.1717e-04\n",
      "Epoch 344/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0280 - accuracy: 0.9893 - mean_squared_error: 7.9005e-04\n",
      "Epoch 345/600\n",
      "50/50 [==============================] - 43s 858ms/step - loss: 0.0104 - accuracy: 0.9959 - mean_squared_error: 2.9482e-04\n",
      "Epoch 346/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0196 - accuracy: 0.9942 - mean_squared_error: 5.2458e-04\n",
      "Epoch 347/600\n",
      "50/50 [==============================] - 43s 858ms/step - loss: 0.0318 - accuracy: 0.9888 - mean_squared_error: 8.7134e-04\n",
      "Epoch 348/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0313 - accuracy: 0.9908 - mean_squared_error: 8.1470e-04\n",
      "Epoch 349/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0155 - accuracy: 0.9932 - mean_squared_error: 4.5532e-04\n",
      "Epoch 350/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0304 - accuracy: 0.9888 - mean_squared_error: 8.4190e-04\n",
      "Epoch 351/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0348 - accuracy: 0.9901 - mean_squared_error: 8.6588e-04\n",
      "Epoch 352/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0321 - accuracy: 0.9912 - mean_squared_error: 7.9542e-04\n",
      "Epoch 353/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0338 - accuracy: 0.9939 - mean_squared_error: 6.8152e-04\n",
      "Epoch 354/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0289 - accuracy: 0.9887 - mean_squared_error: 8.0792e-04\n",
      "Epoch 355/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0197 - accuracy: 0.9947 - mean_squared_error: 4.1940e-04\n",
      "Epoch 356/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0126 - accuracy: 0.9959 - mean_squared_error: 3.3376e-04\n",
      "Epoch 357/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0201 - accuracy: 0.9920 - mean_squared_error: 5.9928e-04\n",
      "Epoch 358/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0214 - accuracy: 0.9928 - mean_squared_error: 5.8942e-04\n",
      "Epoch 359/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0268 - accuracy: 0.9913 - mean_squared_error: 7.4047e-04\n",
      "Epoch 360/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0238 - accuracy: 0.9917 - mean_squared_error: 7.1101e-04\n",
      "Epoch 361/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0298 - accuracy: 0.9904 - mean_squared_error: 7.7701e-04\n",
      "Epoch 362/600\n",
      "50/50 [==============================] - 43s 850ms/step - loss: 0.0339 - accuracy: 0.9888 - mean_squared_error: 8.4157e-04\n",
      "Epoch 363/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0383 - accuracy: 0.9874 - mean_squared_error: 0.0010\n",
      "Epoch 364/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0180 - accuracy: 0.9929 - mean_squared_error: 4.2099e-04\n",
      "Epoch 365/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0334 - accuracy: 0.9872 - mean_squared_error: 9.9280e-04\n",
      "Epoch 366/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0130 - accuracy: 0.9961 - mean_squared_error: 3.5097e-04\n",
      "Epoch 367/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 43s 852ms/step - loss: 0.0198 - accuracy: 0.9943 - mean_squared_error: 5.3336e-04\n",
      "Epoch 368/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0231 - accuracy: 0.9896 - mean_squared_error: 7.0791e-04\n",
      "Epoch 369/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0252 - accuracy: 0.9917 - mean_squared_error: 6.9156e-04\n",
      "Epoch 370/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0414 - accuracy: 0.9864 - mean_squared_error: 0.0011\n",
      "Epoch 371/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0190 - accuracy: 0.9935 - mean_squared_error: 5.0065e-04\n",
      "Epoch 372/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0106 - accuracy: 0.9964 - mean_squared_error: 2.7578e-04\n",
      "Epoch 373/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0320 - accuracy: 0.9915 - mean_squared_error: 8.1436e-04\n",
      "Epoch 374/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0227 - accuracy: 0.9928 - mean_squared_error: 5.7650e-04\n",
      "Epoch 375/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0190 - accuracy: 0.9931 - mean_squared_error: 5.2692e-04\n",
      "Epoch 376/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0250 - accuracy: 0.9914 - mean_squared_error: 7.1919e-04\n",
      "Epoch 377/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0308 - accuracy: 0.9909 - mean_squared_error: 7.5853e-04\n",
      "Epoch 378/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0580 - accuracy: 0.9866 - mean_squared_error: 0.0012\n",
      "Epoch 379/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0446 - accuracy: 0.9873 - mean_squared_error: 0.0013\n",
      "Epoch 380/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0175 - accuracy: 0.9933 - mean_squared_error: 5.0081e-04\n",
      "Epoch 381/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0207 - accuracy: 0.9946 - mean_squared_error: 4.7321e-04\n",
      "Epoch 382/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0369 - accuracy: 0.9841 - mean_squared_error: 0.0011\n",
      "Epoch 383/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0172 - accuracy: 0.9938 - mean_squared_error: 4.8853e-04\n",
      "Epoch 384/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0151 - accuracy: 0.9940 - mean_squared_error: 4.7649e-04\n",
      "Epoch 385/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0294 - accuracy: 0.9901 - mean_squared_error: 7.6798e-04\n",
      "Epoch 386/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0235 - accuracy: 0.9937 - mean_squared_error: 6.0449e-04\n",
      "Epoch 387/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0271 - accuracy: 0.9938 - mean_squared_error: 6.1059e-04\n",
      "Epoch 388/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0246 - accuracy: 0.9899 - mean_squared_error: 7.3212e-04\n",
      "Epoch 389/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0260 - accuracy: 0.9916 - mean_squared_error: 7.1249e-04\n",
      "Epoch 390/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0300 - accuracy: 0.9917 - mean_squared_error: 7.8488e-04\n",
      "Epoch 391/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0290 - accuracy: 0.9908 - mean_squared_error: 7.7678e-04\n",
      "Epoch 392/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0272 - accuracy: 0.9921 - mean_squared_error: 7.3617e-04\n",
      "Epoch 393/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0270 - accuracy: 0.9897 - mean_squared_error: 7.3139e-04\n",
      "Epoch 394/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0238 - accuracy: 0.9916 - mean_squared_error: 6.6301e-04\n",
      "Epoch 395/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0269 - accuracy: 0.9937 - mean_squared_error: 6.0643e-04\n",
      "Epoch 396/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0166 - accuracy: 0.9945 - mean_squared_error: 4.7798e-04\n",
      "Epoch 397/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0326 - accuracy: 0.9906 - mean_squared_error: 9.2994e-04\n",
      "Epoch 398/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0335 - accuracy: 0.9885 - mean_squared_error: 8.0010e-04\n",
      "Epoch 399/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0356 - accuracy: 0.9885 - mean_squared_error: 8.9813e-04\n",
      "Epoch 400/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0366 - accuracy: 0.9941 - mean_squared_error: 6.0591e-04\n",
      "Epoch 401/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0332 - accuracy: 0.9888 - mean_squared_error: 8.9415e-04\n",
      "Epoch 402/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0358 - accuracy: 0.9853 - mean_squared_error: 9.9245e-04\n",
      "Epoch 403/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0360 - accuracy: 0.9896 - mean_squared_error: 8.2979e-04\n",
      "Epoch 404/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0191 - accuracy: 0.9922 - mean_squared_error: 5.2332e-04\n",
      "Epoch 405/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0155 - accuracy: 0.9913 - mean_squared_error: 4.6681e-04\n",
      "Epoch 406/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0216 - accuracy: 0.9921 - mean_squared_error: 5.7483e-04\n",
      "Epoch 407/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0282 - accuracy: 0.9912 - mean_squared_error: 6.3032e-04\n",
      "Epoch 408/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0250 - accuracy: 0.9893 - mean_squared_error: 7.4417e-04\n",
      "Epoch 409/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0134 - accuracy: 0.9958 - mean_squared_error: 3.6859e-04\n",
      "Epoch 410/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0149 - accuracy: 0.9952 - mean_squared_error: 3.8191e-04\n",
      "Epoch 411/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0204 - accuracy: 0.9915 - mean_squared_error: 5.9531e-04\n",
      "Epoch 412/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0167 - accuracy: 0.9958 - mean_squared_error: 4.3501e-04\n",
      "Epoch 413/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0267 - accuracy: 0.9907 - mean_squared_error: 7.3920e-04\n",
      "Epoch 414/600\n",
      "50/50 [==============================] - 42s 836ms/step - loss: 0.0294 - accuracy: 0.9901 - mean_squared_error: 7.2581e-04\n",
      "Epoch 415/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0210 - accuracy: 0.9923 - mean_squared_error: 6.3150e-04\n",
      "Epoch 416/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0223 - accuracy: 0.9945 - mean_squared_error: 4.8307e-04\n",
      "Epoch 417/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0180 - accuracy: 0.9921 - mean_squared_error: 5.2029e-04\n",
      "Epoch 418/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0135 - accuracy: 0.9961 - mean_squared_error: 3.5977e-04\n",
      "Epoch 419/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0231 - accuracy: 0.9941 - mean_squared_error: 5.7272e-04\n",
      "Epoch 420/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0234 - accuracy: 0.9925 - mean_squared_error: 6.9283e-04\n",
      "Epoch 421/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0128 - accuracy: 0.9962 - mean_squared_error: 3.4657e-04\n",
      "Epoch 422/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0326 - accuracy: 0.9912 - mean_squared_error: 7.5975e-04\n",
      "Epoch 423/600\n",
      "50/50 [==============================] - 43s 857ms/step - loss: 0.0273 - accuracy: 0.9911 - mean_squared_error: 7.2253e-04\n",
      "Epoch 424/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0206 - accuracy: 0.9930 - mean_squared_error: 5.5509e-04\n",
      "Epoch 425/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0199 - accuracy: 0.9940 - mean_squared_error: 5.0849e-04\n",
      "Epoch 426/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0183 - accuracy: 0.9967 - mean_squared_error: 3.6854e-04\n",
      "Epoch 427/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0137 - accuracy: 0.9940 - mean_squared_error: 3.9842e-04\n",
      "Epoch 428/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0298 - accuracy: 0.9910 - mean_squared_error: 7.8546e-04\n",
      "Epoch 429/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0287 - accuracy: 0.9922 - mean_squared_error: 7.4051e-04\n",
      "Epoch 430/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0154 - accuracy: 0.9939 - mean_squared_error: 4.8804e-04\n",
      "Epoch 431/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0163 - accuracy: 0.9929 - mean_squared_error: 5.0261e-04\n",
      "Epoch 432/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0194 - accuracy: 0.9926 - mean_squared_error: 5.6884e-04\n",
      "Epoch 433/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0417 - accuracy: 0.9891 - mean_squared_error: 9.5130e-04\n",
      "Epoch 434/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0284 - accuracy: 0.9910 - mean_squared_error: 7.3746e-04\n",
      "Epoch 435/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0303 - accuracy: 0.9916 - mean_squared_error: 6.6743e-04\n",
      "Epoch 436/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0154 - accuracy: 0.9934 - mean_squared_error: 4.7312e-04\n",
      "Epoch 437/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0157 - accuracy: 0.9946 - mean_squared_error: 4.3319e-04\n",
      "Epoch 438/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0255 - accuracy: 0.9919 - mean_squared_error: 6.0462e-04\n",
      "Epoch 439/600\n",
      "50/50 [==============================] - 42s 837ms/step - loss: 0.0193 - accuracy: 0.9927 - mean_squared_error: 5.3115e-04\n",
      "Epoch 440/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0320 - accuracy: 0.9903 - mean_squared_error: 7.1812e-04\n",
      "Epoch 441/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0363 - accuracy: 0.9885 - mean_squared_error: 8.5952e-04\n",
      "Epoch 442/600\n",
      "50/50 [==============================] - 42s 836ms/step - loss: 0.0168 - accuracy: 0.9948 - mean_squared_error: 4.7379e-04\n",
      "Epoch 443/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0282 - accuracy: 0.9906 - mean_squared_error: 8.0688e-04\n",
      "Epoch 444/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0254 - accuracy: 0.9930 - mean_squared_error: 6.3169e-04\n",
      "Epoch 445/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0186 - accuracy: 0.9945 - mean_squared_error: 4.8032e-04\n",
      "Epoch 446/600\n",
      "50/50 [==============================] - 43s 852ms/step - loss: 0.0255 - accuracy: 0.9928 - mean_squared_error: 6.7730e-04\n",
      "Epoch 447/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0102 - accuracy: 0.9964 - mean_squared_error: 2.6340e-04\n",
      "Epoch 448/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0263 - accuracy: 0.9908 - mean_squared_error: 6.9002e-04\n",
      "Epoch 449/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0232 - accuracy: 0.9898 - mean_squared_error: 7.1293e-04\n",
      "Epoch 450/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0258 - accuracy: 0.9916 - mean_squared_error: 6.5556e-04\n",
      "Epoch 451/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0219 - accuracy: 0.9947 - mean_squared_error: 5.0476e-04\n",
      "Epoch 452/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0217 - accuracy: 0.9954 - mean_squared_error: 4.8672e-04\n",
      "Epoch 453/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0380 - accuracy: 0.9898 - mean_squared_error: 9.0961e-04\n",
      "Epoch 454/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0269 - accuracy: 0.9914 - mean_squared_error: 7.5755e-04\n",
      "Epoch 455/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0264 - accuracy: 0.9926 - mean_squared_error: 6.1295e-04\n",
      "Epoch 456/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0214 - accuracy: 0.9933 - mean_squared_error: 5.1178e-04\n",
      "Epoch 457/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0186 - accuracy: 0.9938 - mean_squared_error: 5.3733e-04\n",
      "Epoch 458/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0163 - accuracy: 0.9953 - mean_squared_error: 4.4612e-04\n",
      "Epoch 459/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0205 - accuracy: 0.9933 - mean_squared_error: 6.0045e-04\n",
      "Epoch 460/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0230 - accuracy: 0.9926 - mean_squared_error: 6.0378e-04\n",
      "Epoch 461/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0205 - accuracy: 0.9916 - mean_squared_error: 6.5286e-04\n",
      "Epoch 462/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0173 - accuracy: 0.9922 - mean_squared_error: 5.2501e-04\n",
      "Epoch 463/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0204 - accuracy: 0.9941 - mean_squared_error: 5.5240e-04\n",
      "Epoch 464/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0209 - accuracy: 0.9956 - mean_squared_error: 3.8570e-04\n",
      "Epoch 465/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0194 - accuracy: 0.9951 - mean_squared_error: 5.2393e-04\n",
      "Epoch 466/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0116 - accuracy: 0.9962 - mean_squared_error: 2.9739e-04\n",
      "Epoch 467/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0237 - accuracy: 0.9934 - mean_squared_error: 5.7148e-04\n",
      "Epoch 468/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0188 - accuracy: 0.9936 - mean_squared_error: 4.9288e-04\n",
      "Epoch 469/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0111 - accuracy: 0.9949 - mean_squared_error: 3.0948e-04\n",
      "Epoch 470/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0205 - accuracy: 0.9943 - mean_squared_error: 4.3364e-04\n",
      "Epoch 471/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0205 - accuracy: 0.9902 - mean_squared_error: 6.3088e-04\n",
      "Epoch 472/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0160 - accuracy: 0.9964 - mean_squared_error: 3.9723e-04\n",
      "Epoch 473/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0214 - accuracy: 0.9939 - mean_squared_error: 5.5237e-04\n",
      "Epoch 474/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0208 - accuracy: 0.9915 - mean_squared_error: 6.4883e-04\n",
      "Epoch 475/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0295 - accuracy: 0.9881 - mean_squared_error: 8.3596e-04\n",
      "Epoch 476/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0257 - accuracy: 0.9917 - mean_squared_error: 6.5724e-04\n",
      "Epoch 477/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0243 - accuracy: 0.9935 - mean_squared_error: 6.9116e-04\n",
      "Epoch 478/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0385 - accuracy: 0.9853 - mean_squared_error: 0.0011\n",
      "Epoch 479/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0165 - accuracy: 0.9946 - mean_squared_error: 4.5049e-04\n",
      "Epoch 480/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0129 - accuracy: 0.9940 - mean_squared_error: 3.9009e-04\n",
      "Epoch 481/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0193 - accuracy: 0.9937 - mean_squared_error: 5.1157e-04\n",
      "Epoch 482/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0206 - accuracy: 0.9901 - mean_squared_error: 6.1660e-04\n",
      "Epoch 483/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0246 - accuracy: 0.9904 - mean_squared_error: 5.7423e-04\n",
      "Epoch 484/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0278 - accuracy: 0.9915 - mean_squared_error: 7.2647e-04\n",
      "Epoch 485/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0279 - accuracy: 0.9901 - mean_squared_error: 7.4041e-04\n",
      "Epoch 486/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0326 - accuracy: 0.9914 - mean_squared_error: 8.1669e-04\n",
      "Epoch 487/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0213 - accuracy: 0.9907 - mean_squared_error: 6.0783e-04\n",
      "Epoch 488/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0133 - accuracy: 0.9944 - mean_squared_error: 3.9310e-04\n",
      "Epoch 489/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0248 - accuracy: 0.9923 - mean_squared_error: 5.9310e-04\n",
      "Epoch 490/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0275 - accuracy: 0.9905 - mean_squared_error: 7.9694e-04\n",
      "Epoch 491/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0255 - accuracy: 0.9912 - mean_squared_error: 6.9634e-04\n",
      "Epoch 492/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0418 - accuracy: 0.9916 - mean_squared_error: 9.2715e-04\n",
      "Epoch 493/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0234 - accuracy: 0.9920 - mean_squared_error: 6.3655e-04\n",
      "Epoch 494/600\n",
      "50/50 [==============================] - 42s 850ms/step - loss: 0.0193 - accuracy: 0.9933 - mean_squared_error: 5.7968e-04\n",
      "Epoch 495/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0117 - accuracy: 0.9966 - mean_squared_error: 3.0095e-04\n",
      "Epoch 496/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0195 - accuracy: 0.9946 - mean_squared_error: 3.9248e-04\n",
      "Epoch 497/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0295 - accuracy: 0.9917 - mean_squared_error: 7.5904e-04\n",
      "Epoch 498/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0230 - accuracy: 0.9919 - mean_squared_error: 7.1775e-04\n",
      "Epoch 499/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0094 - accuracy: 0.9975 - mean_squared_error: 2.5059e-04\n",
      "Epoch 500/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0300 - accuracy: 0.9932 - mean_squared_error: 6.3205e-04\n",
      "Epoch 501/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0204 - accuracy: 0.9931 - mean_squared_error: 6.1381e-04\n",
      "Epoch 502/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0192 - accuracy: 0.9917 - mean_squared_error: 5.4948e-04\n",
      "Epoch 503/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0302 - accuracy: 0.9926 - mean_squared_error: 6.7660e-04\n",
      "Epoch 504/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0279 - accuracy: 0.9939 - mean_squared_error: 5.6428e-04\n",
      "Epoch 505/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0199 - accuracy: 0.9950 - mean_squared_error: 4.8843e-04\n",
      "Epoch 506/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0096 - accuracy: 0.9981 - mean_squared_error: 2.2206e-04\n",
      "Epoch 507/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0121 - accuracy: 0.9970 - mean_squared_error: 3.1396e-04\n",
      "Epoch 508/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0302 - accuracy: 0.9918 - mean_squared_error: 6.8130e-04\n",
      "Epoch 509/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0180 - accuracy: 0.9941 - mean_squared_error: 4.3416e-04\n",
      "Epoch 510/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0227 - accuracy: 0.9939 - mean_squared_error: 5.3010e-04\n",
      "Epoch 511/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0374 - accuracy: 0.9881 - mean_squared_error: 8.2371e-04\n",
      "Epoch 512/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0078 - accuracy: 0.9982 - mean_squared_error: 1.9692e-04\n",
      "Epoch 513/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0181 - accuracy: 0.9950 - mean_squared_error: 4.4542e-04\n",
      "Epoch 514/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0166 - accuracy: 0.9948 - mean_squared_error: 4.5105e-04\n",
      "Epoch 515/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0258 - accuracy: 0.9942 - mean_squared_error: 5.0420e-04\n",
      "Epoch 516/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0203 - accuracy: 0.9955 - mean_squared_error: 4.5797e-04\n",
      "Epoch 517/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0194 - accuracy: 0.9936 - mean_squared_error: 5.4110e-04\n",
      "Epoch 518/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0254 - accuracy: 0.9937 - mean_squared_error: 5.9367e-04\n",
      "Epoch 519/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0191 - accuracy: 0.9950 - mean_squared_error: 4.0891e-04\n",
      "Epoch 520/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0202 - accuracy: 0.9933 - mean_squared_error: 5.6465e-04\n",
      "Epoch 521/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0093 - accuracy: 0.9961 - mean_squared_error: 2.7637e-04\n",
      "Epoch 522/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0342 - accuracy: 0.9909 - mean_squared_error: 8.1638e-04\n",
      "Epoch 523/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0170 - accuracy: 0.9929 - mean_squared_error: 4.7583e-04\n",
      "Epoch 524/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0257 - accuracy: 0.9944 - mean_squared_error: 4.6986e-04\n",
      "Epoch 525/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0184 - accuracy: 0.9951 - mean_squared_error: 4.7004e-04\n",
      "Epoch 526/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0210 - accuracy: 0.9935 - mean_squared_error: 5.1125e-04\n",
      "Epoch 527/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0170 - accuracy: 0.9955 - mean_squared_error: 3.9197e-04\n",
      "Epoch 528/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0125 - accuracy: 0.9957 - mean_squared_error: 3.6434e-04\n",
      "Epoch 529/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0163 - accuracy: 0.9952 - mean_squared_error: 3.7368e-04\n",
      "Epoch 530/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0135 - accuracy: 0.9954 - mean_squared_error: 3.9438e-04\n",
      "Epoch 531/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0234 - accuracy: 0.9938 - mean_squared_error: 4.8190e-04\n",
      "Epoch 532/600\n",
      "50/50 [==============================] - 43s 851ms/step - loss: 0.0381 - accuracy: 0.9863 - mean_squared_error: 9.5602e-04\n",
      "Epoch 533/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0179 - accuracy: 0.9932 - mean_squared_error: 4.7796e-04\n",
      "Epoch 534/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0165 - accuracy: 0.9946 - mean_squared_error: 4.0064e-04\n",
      "Epoch 535/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0156 - accuracy: 0.9943 - mean_squared_error: 4.8483e-04\n",
      "Epoch 536/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0159 - accuracy: 0.9932 - mean_squared_error: 4.9723e-04\n",
      "Epoch 537/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0108 - accuracy: 0.9972 - mean_squared_error: 2.7305e-04\n",
      "Epoch 538/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0155 - accuracy: 0.9974 - mean_squared_error: 3.2686e-04\n",
      "Epoch 539/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0179 - accuracy: 0.9944 - mean_squared_error: 4.0565e-04\n",
      "Epoch 540/600\n",
      "50/50 [==============================] - 42s 847ms/step - loss: 0.0201 - accuracy: 0.9942 - mean_squared_error: 4.1710e-04\n",
      "Epoch 541/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0122 - accuracy: 0.9962 - mean_squared_error: 3.5606e-04\n",
      "Epoch 542/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0440 - accuracy: 0.9865 - mean_squared_error: 0.0012\n",
      "Epoch 543/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0383 - accuracy: 0.9878 - mean_squared_error: 9.5528e-04\n",
      "Epoch 544/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0179 - accuracy: 0.9906 - mean_squared_error: 5.4979e-04\n",
      "Epoch 545/600\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0097 - accuracy: 0.9958 - mean_squared_error: 3.0321e-04\n",
      "Epoch 546/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0135 - accuracy: 0.9953 - mean_squared_error: 3.6385e-04\n",
      "Epoch 547/600\n",
      "50/50 [==============================] - 42s 839ms/step - loss: 0.0178 - accuracy: 0.9949 - mean_squared_error: 3.9557e-04\n",
      "Epoch 548/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0142 - accuracy: 0.9970 - mean_squared_error: 3.2010e-04\n",
      "Epoch 549/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0147 - accuracy: 0.9948 - mean_squared_error: 4.2009e-04\n",
      "Epoch 550/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0186 - accuracy: 0.9924 - mean_squared_error: 5.3283e-04\n",
      "Epoch 551/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0092 - accuracy: 0.9971 - mean_squared_error: 2.4982e-04\n",
      "Epoch 552/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0345 - accuracy: 0.9898 - mean_squared_error: 7.6830e-04\n",
      "Epoch 553/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0231 - accuracy: 0.9915 - mean_squared_error: 6.2862e-04\n",
      "Epoch 554/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0097 - accuracy: 0.9967 - mean_squared_error: 2.8412e-04\n",
      "Epoch 555/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0210 - accuracy: 0.9933 - mean_squared_error: 6.1333e-04\n",
      "Epoch 556/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0235 - accuracy: 0.9946 - mean_squared_error: 4.9826e-04\n",
      "Epoch 557/600\n",
      "50/50 [==============================] - 42s 838ms/step - loss: 0.0161 - accuracy: 0.9937 - mean_squared_error: 4.9416e-04\n",
      "Epoch 558/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0152 - accuracy: 0.9958 - mean_squared_error: 4.3108e-04\n",
      "Epoch 559/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0154 - accuracy: 0.9949 - mean_squared_error: 4.2705e-04\n",
      "Epoch 560/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0272 - accuracy: 0.9905 - mean_squared_error: 7.2874e-04\n",
      "Epoch 561/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0444 - accuracy: 0.9880 - mean_squared_error: 0.0011\n",
      "Epoch 562/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0203 - accuracy: 0.9921 - mean_squared_error: 5.5345e-04\n",
      "Epoch 563/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0325 - accuracy: 0.9906 - mean_squared_error: 7.7295e-04\n",
      "Epoch 564/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0285 - accuracy: 0.9923 - mean_squared_error: 6.7256e-04\n",
      "Epoch 565/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0220 - accuracy: 0.9933 - mean_squared_error: 5.8758e-04\n",
      "Epoch 566/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0227 - accuracy: 0.9928 - mean_squared_error: 6.3388e-04\n",
      "Epoch 567/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0180 - accuracy: 0.9941 - mean_squared_error: 5.1310e-04\n",
      "Epoch 568/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0247 - accuracy: 0.9917 - mean_squared_error: 6.2839e-04\n",
      "Epoch 569/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0154 - accuracy: 0.9962 - mean_squared_error: 3.7914e-04\n",
      "Epoch 570/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0193 - accuracy: 0.9941 - mean_squared_error: 4.2468e-04\n",
      "Epoch 571/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0160 - accuracy: 0.9956 - mean_squared_error: 3.4277e-04\n",
      "Epoch 572/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0158 - accuracy: 0.9959 - mean_squared_error: 4.1987e-04\n",
      "Epoch 573/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0155 - accuracy: 0.9950 - mean_squared_error: 3.6718e-04\n",
      "Epoch 574/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0157 - accuracy: 0.9944 - mean_squared_error: 4.6991e-04\n",
      "Epoch 575/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0103 - accuracy: 0.9961 - mean_squared_error: 3.2480e-04\n",
      "Epoch 576/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0202 - accuracy: 0.9936 - mean_squared_error: 5.2639e-04\n",
      "Epoch 577/600\n",
      "50/50 [==============================] - 43s 850ms/step - loss: 0.0111 - accuracy: 0.9969 - mean_squared_error: 2.4273e-04\n",
      "Epoch 578/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0140 - accuracy: 0.9950 - mean_squared_error: 4.2091e-04\n",
      "Epoch 579/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0079 - accuracy: 0.9974 - mean_squared_error: 2.0767e-04\n",
      "Epoch 580/600\n",
      "50/50 [==============================] - 42s 840ms/step - loss: 0.0106 - accuracy: 0.9963 - mean_squared_error: 3.0370e-04\n",
      "Epoch 581/600\n",
      "50/50 [==============================] - 42s 845ms/step - loss: 0.0129 - accuracy: 0.9965 - mean_squared_error: 3.4351e-04\n",
      "Epoch 582/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0128 - accuracy: 0.9955 - mean_squared_error: 3.3391e-04\n",
      "Epoch 583/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0265 - accuracy: 0.9930 - mean_squared_error: 5.7198e-04\n",
      "Epoch 584/600\n",
      "50/50 [==============================] - 42s 841ms/step - loss: 0.0231 - accuracy: 0.9938 - mean_squared_error: 4.8210e-04\n",
      "Epoch 585/600\n",
      "50/50 [==============================] - 42s 846ms/step - loss: 0.0216 - accuracy: 0.9935 - mean_squared_error: 5.5006e-04\n",
      "Epoch 586/600\n",
      "50/50 [==============================] - 43s 849ms/step - loss: 0.0360 - accuracy: 0.9917 - mean_squared_error: 7.2195e-04\n",
      "Epoch 587/600\n",
      "50/50 [==============================] - 42s 842ms/step - loss: 0.0244 - accuracy: 0.9923 - mean_squared_error: 6.9431e-04\n",
      "Epoch 588/600\n",
      "50/50 [==============================] - 42s 848ms/step - loss: 0.0126 - accuracy: 0.9937 - mean_squared_error: 3.6705e-04\n",
      "Epoch 589/600\n",
      "50/50 [==============================] - 43s 851ms/step - loss: 0.0087 - accuracy: 0.9956 - mean_squared_error: 2.3593e-04\n",
      "Epoch 590/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0328 - accuracy: 0.9905 - mean_squared_error: 7.9312e-04\n",
      "Epoch 591/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0118 - accuracy: 0.9945 - mean_squared_error: 3.1726e-04\n",
      "Epoch 592/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0130 - accuracy: 0.9965 - mean_squared_error: 3.1370e-04\n",
      "Epoch 593/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0091 - accuracy: 0.9979 - mean_squared_error: 2.1780e-04\n",
      "Epoch 594/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0120 - accuracy: 0.9953 - mean_squared_error: 3.3074e-04\n",
      "Epoch 595/600\n",
      "50/50 [==============================] - 43s 854ms/step - loss: 0.0091 - accuracy: 0.9973 - mean_squared_error: 2.5083e-04\n",
      "Epoch 596/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0252 - accuracy: 0.9941 - mean_squared_error: 6.1062e-04\n",
      "Epoch 597/600\n",
      "50/50 [==============================] - 42s 844ms/step - loss: 0.0097 - accuracy: 0.9957 - mean_squared_error: 3.1387e-04\n",
      "Epoch 598/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0075 - accuracy: 0.9986 - mean_squared_error: 1.7117e-04\n",
      "Epoch 599/600\n",
      "50/50 [==============================] - 42s 843ms/step - loss: 0.0071 - accuracy: 0.9979 - mean_squared_error: 1.6966e-04\n",
      "Epoch 600/600\n",
      "50/50 [==============================] - 43s 850ms/step - loss: 0.0044 - accuracy: 0.9986 - mean_squared_error: 1.2311e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x21c55a35188>"
      ]
     },
     "execution_count": 220,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.fit(x=train_batch2, steps_per_epoch = 50, epochs=600, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[3.35e-01, 3.63e-08, 2.60e-07, ..., 5.11e-07, 3.74e-10, 6.07e-01],\n",
       "       [3.61e-03, 8.82e-06, 2.06e-10, ..., 7.70e-01, 1.66e-10, 2.77e-09],\n",
       "       [1.13e-02, 7.44e-07, 9.71e-09, ..., 1.58e-05, 5.00e-02, 7.49e-01],\n",
       "       ...,\n",
       "       [4.02e-18, 4.17e-15, 2.84e-18, ..., 5.90e-14, 1.74e-08, 1.00e+00],\n",
       "       [1.04e-08, 2.81e-11, 2.72e-08, ..., 1.14e-15, 9.40e-11, 1.00e+00],\n",
       "       [5.22e-07, 1.28e-12, 1.26e-14, ..., 5.04e-07, 7.94e-13, 1.38e-07]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output = model.predict(test_batch)\n",
    "output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.22e-06, 2.04e-05, 7.42e-06, ..., 9.55e-07, 9.40e-01, 5.93e-12],\n",
       "       [9.91e-01, 6.56e-09, 2.92e-09, ..., 3.51e-09, 1.51e-12, 1.71e-11],\n",
       "       [1.00e+00, 4.27e-14, 1.42e-13, ..., 2.64e-13, 1.89e-12, 3.54e-15],\n",
       "       ...,\n",
       "       [6.47e-21, 4.52e-16, 3.67e-15, ..., 7.91e-15, 1.39e-17, 1.00e+00],\n",
       "       [9.12e-20, 2.75e-10, 9.71e-12, ..., 1.03e-08, 1.83e-10, 1.00e+00],\n",
       "       [1.38e-21, 5.35e-14, 9.37e-17, ..., 1.20e-13, 4.55e-12, 1.00e+00]],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output2 = model2.predict(test_batch2)\n",
    "output2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction = []\n",
    "prediction2 = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getType(pokemon):\n",
    "    return np.argmax(pokemon)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [],
   "source": [
    "for x in range(0,len(output)):\n",
    "    prediction.append(getType(output[x]))\n",
    "    \n",
    "for x in range(0, len(output2)):\n",
    "    prediction2.append(getType(output2[x]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Type 1 accuracy:\n",
      "65.5593220338983\n",
      "Type 2 accuracy:\n",
      "88.55963438006782\n"
     ]
    }
   ],
   "source": [
    "def test_accuracy(prediction, real):\n",
    "    correct = 0\n",
    "    if len(prediction) != len(real):\n",
    "        print(\"Error\")\n",
    "        return 0\n",
    "    for x in range(0, len(prediction)):\n",
    "        if prediction[x] == real[x]:\n",
    "            correct += 1\n",
    "    print(100 * (correct / len(prediction)))\n",
    "\n",
    "print(\"Type 1 accuracy:\")\n",
    "test_accuracy(prediction,test_batch.labels)\n",
    "print(\"Type 2 accuracy:\")\n",
    "test_accuracy(prediction2,test_batch2.labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [],
   "source": [
    "cm = confusion_matrix(test_batch.labels,prediction)\n",
    "np.set_printoptions(precision = 2)\n",
    "cm2 = confusion_matrix(test_batch2.labels,prediction2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(cm, classes, normalize=False, title='Confusion matrix', cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\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",
    "    fmt = '.2f' if normalize else 'd'\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, format(cm[i, j], fmt),\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": 229,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = [13,10])\n",
    "plot_confusion_matrix(cm, classes = type_1, title = \"Type 1 Confusion Matrix\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = [13,10])\n",
    "plot_confusion_matrix(cm2, classes = type_2, title = \"Type 2 Confusion Matrix\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}