Endsem.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6fbd9fb7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hello\n"
     ]
    }
   ],
   "source": [
    "print ()'Hello')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0b16d175",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: tensorflow==2.4.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (2.4.1)\n",
      "Requirement already satisfied: tensorflow-gpu==2.4.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (2.4.1)\n",
      "Requirement already satisfied: opencv-python in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (3.4.2.17)\n",
      "Note: you may need to restart the kernel to use updated packages.Requirement already satisfied: mediapipe in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (0.8.3)\n",
      "\n",
      "Requirement already satisfied: sklearn in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (0.0)\n",
      "Requirement already satisfied: matplotlib in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (3.2.2)\n",
      "Requirement already satisfied: opt-einsum~=3.3.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (3.3.0)\n",
      "Requirement already satisfied: wheel~=0.35 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (0.37.1)\n",
      "Requirement already satisfied: grpcio~=1.32.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.32.0)\n",
      "Requirement already satisfied: typing-extensions~=3.7.4 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (3.7.4.3)\n",
      "Requirement already satisfied: google-pasta~=0.2 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (0.2.0)\n",
      "Requirement already satisfied: absl-py~=0.10 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (0.15.0)\n",
      "Requirement already satisfied: keras-preprocessing~=1.1.2 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.1.2)\n",
      "Requirement already satisfied: wrapt~=1.12.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.12.1)\n",
      "Requirement already satisfied: protobuf>=3.9.2 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (3.19.4)\n",
      "Requirement already satisfied: tensorflow-estimator<2.5.0,>=2.4.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (2.4.0)\n",
      "Requirement already satisfied: astunparse~=1.6.3 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.6.3)\n",
      "Requirement already satisfied: flatbuffers~=1.12.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.12)\n",
      "Requirement already satisfied: termcolor~=1.1.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.1.0)\n",
      "Requirement already satisfied: tensorboard~=2.4 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (2.8.0)\n",
      "Requirement already satisfied: numpy~=1.19.2 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.19.3)\n",
      "Requirement already satisfied: gast==0.3.3 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (0.3.3)\n",
      "Requirement already satisfied: six~=1.15.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (1.15.0)\n",
      "Requirement already satisfied: h5py~=2.10.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorflow==2.4.1) (2.10.0)\n",
      "Requirement already satisfied: attrs in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from mediapipe) (21.4.0)\n",
      "Requirement already satisfied: dataclasses in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from mediapipe) (0.8)\n",
      "Requirement already satisfied: scikit-learn in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from sklearn) (0.24.2)\n",
      "Requirement already satisfied: cycler>=0.10 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from matplotlib) (0.11.0)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from matplotlib) (3.0.7)\n",
      "Requirement already satisfied: python-dateutil>=2.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from matplotlib) (2.8.2)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from matplotlib) (1.3.1)\n",
      "Requirement already satisfied: markdown>=2.6.8 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (3.3.6)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (59.6.0)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (1.8.1)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (0.4.6)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (2.0.2)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (2.27.1)\n",
      "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (0.6.1)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from tensorboard~=2.4->tensorflow==2.4.1) (2.6.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from scikit-learn->sklearn) (3.1.0)\n",
      "Requirement already satisfied: joblib>=0.11 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from scikit-learn->sklearn) (1.1.0)\n",
      "Requirement already satisfied: scipy>=0.19.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from scikit-learn->sklearn) (1.2.2)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.1) (4.8)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.1) (0.2.8)\n",
      "Requirement already satisfied: cachetools<6.0,>=2.0.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.1) (4.2.4)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.4->tensorflow==2.4.1) (1.3.1)\n",
      "Requirement already satisfied: importlib-metadata>=4.4 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from markdown>=2.6.8->tensorboard~=2.4->tensorflow==2.4.1) (4.8.3)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.1) (2021.10.8)\n",
      "Requirement already satisfied: charset-normalizer~=2.0.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.1) (2.0.11)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.1) (1.26.8)\n",
      "Requirement already satisfied: idna<4,>=2.5 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.4->tensorflow==2.4.1) (3.3)\n",
      "Requirement already satisfied: zipp>=0.5 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard~=2.4->tensorflow==2.4.1) (3.6.0)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard~=2.4->tensorflow==2.4.1) (0.4.8)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in c:\\users\\aditya\\appdata\\local\\programs\\python\\python36\\lib\\site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.4->tensorflow==2.4.1) (3.2.0)\n"
     ]
    }
   ],
   "source": [
    "pip install tensorflow==2.4.1 tensorflow-gpu==2.4.1 opencv-python mediapipe sklearn matplotlib"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "299d5228",
   "metadata": {},
   "source": [
    "# Import and Install Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6e319a9d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import numpy as np\n",
    "import os\n",
    "from matplotlib import pyplot as plt\n",
    "import time\n",
    "import mediapipe as mp"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92224d46",
   "metadata": {},
   "source": [
    "# Keypoints using MP Holistic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "2c1120e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "mp_holistic = mp.solutions.holistic # Holistic model\n",
    "mp_drawing = mp.solutions.drawing_utils # Drawing utilities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "29da6251",
   "metadata": {},
   "outputs": [],
   "source": [
    "def mediapipe_detection(image, model):\n",
    "    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) # COLOR CONVERSION BGR 2 RGB\n",
    "    image.flags.writeable = False                  # Image is no longer writeable\n",
    "    results = model.process(image)                 # Make prediction\n",
    "    image.flags.writeable = True                   # Image is now writeable \n",
    "    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR) # COLOR COVERSION RGB 2 BGR\n",
    "    return image, results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "62f5d1c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_landmarks(image, results):\n",
    "    mp_drawing.draw_landmarks(image, results.face_landmarks, mp_holistic.FACE_CONNECTIONS) # Draw face connections\n",
    "    mp_drawing.draw_landmarks(image, results.pose_landmarks, mp_holistic.POSE_CONNECTIONS) # Draw pose connections\n",
    "    mp_drawing.draw_landmarks(image, results.left_hand_landmarks, mp_holistic.HAND_CONNECTIONS) # Draw left hand connections\n",
    "    mp_drawing.draw_landmarks(image, results.right_hand_landmarks, mp_holistic.HAND_CONNECTIONS) # Draw right hand connections"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6b9c0b57",
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_styled_landmarks(image, results):\n",
    "    # Draw face connections\n",
    "    mp_drawing.draw_landmarks(image, results.face_landmarks, mp_holistic.FACE_CONNECTIONS, \n",
    "                             mp_drawing.DrawingSpec(color=(80,110,10), thickness=1, circle_radius=1), \n",
    "                             mp_drawing.DrawingSpec(color=(80,256,121), thickness=1, circle_radius=1)\n",
    "                             ) \n",
    "    # Draw pose connections\n",
    "    mp_drawing.draw_landmarks(image, results.pose_landmarks, mp_holistic.POSE_CONNECTIONS,\n",
    "                             mp_drawing.DrawingSpec(color=(80,22,10), thickness=2, circle_radius=4), \n",
    "                             mp_drawing.DrawingSpec(color=(80,44,121), thickness=2, circle_radius=2)\n",
    "                             ) \n",
    "    # Draw left hand connections\n",
    "    mp_drawing.draw_landmarks(image, results.left_hand_landmarks, mp_holistic.HAND_CONNECTIONS, \n",
    "                             mp_drawing.DrawingSpec(color=(121,22,76), thickness=2, circle_radius=4), \n",
    "                             mp_drawing.DrawingSpec(color=(121,44,250), thickness=2, circle_radius=2)\n",
    "                             ) \n",
    "    # Draw right hand connections  \n",
    "    mp_drawing.draw_landmarks(image, results.right_hand_landmarks, mp_holistic.HAND_CONNECTIONS, \n",
    "                             mp_drawing.DrawingSpec(color=(245,117,66), thickness=2, circle_radius=4), \n",
    "                             mp_drawing.DrawingSpec(color=(245,66,230), thickness=2, circle_radius=2)\n",
    "                             ) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a0fbba02",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n"
     ]
    }
   ],
   "source": [
    "cap = cv2.VideoCapture(0)\n",
    "# Set mediapipe model \n",
    "with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:\n",
    "    while cap.isOpened():\n",
    "\n",
    "        # Read feed\n",
    "        ret, frame = cap.read()\n",
    "\n",
    "        # Make detections\n",
    "        image, results = mediapipe_detection(frame, holistic)\n",
    "        print(results)\n",
    "        \n",
    "        # Draw landmarks\n",
    "        draw_styled_landmarks(image, results)\n",
    "\n",
    "        # Show to screen\n",
    "        cv2.imshow('OpenCV Feed', image)\n",
    "\n",
    "        # Break gracefully\n",
    "        if cv2.waitKey(10) & 0xFF == ord('q'):\n",
    "            break\n",
    "    cap.release()\n",
    "    cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "0fbdc637",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(results.left_hand_landmarks.landmark)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c8bfa6f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mediapipe.python.solution_base.SolutionOutputs"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "961e43e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "draw_landmarks(frame, results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "ebda91cf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1e4e1818908>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "726af990",
   "metadata": {},
   "source": [
    "# Extract Keypoint Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "5dec5a97",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(results.left_hand_landmarks.landmark)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "12cf9546",
   "metadata": {},
   "outputs": [],
   "source": [
    "pose = []\n",
    "for res in results.pose_landmarks.landmark:\n",
    "    test = np.array([res.x, res.y, res.z, res.visibility])\n",
    "    pose.append(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "39b0aacc",
   "metadata": {},
   "outputs": [],
   "source": [
    "pose = np.array([[res.x, res.y, res.z, res.visibility] for res in results.pose_landmarks.landmark]).flatten() if results.pose_landmarks else np.zeros(132)\n",
    "face = np.array([[res.x, res.y, res.z] for res in results.face_landmarks.landmark]).flatten() if results.face_landmarks else np.zeros(1404)\n",
    "lh = np.array([[res.x, res.y, res.z] for res in results.left_hand_landmarks.landmark]).flatten() if results.left_hand_landmarks else np.zeros(21*3)\n",
    "rh = np.array([[res.x, res.y, res.z] for res in results.right_hand_landmarks.landmark]).flatten() if results.right_hand_landmarks else np.zeros(21*3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "c070c1f3",
   "metadata": {},
   "outputs": [
    {
     "ename": "IndentationError",
     "evalue": "unexpected indent (<ipython-input-18-3fe2adb7da0b>, line 2)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-18-3fe2adb7da0b>\"\u001b[1;36m, line \u001b[1;32m2\u001b[0m\n\u001b[1;33m    if results.face_landmarks\u001b[0m\n\u001b[1;37m    ^\u001b[0m\n\u001b[1;31mIndentationError\u001b[0m\u001b[1;31m:\u001b[0m unexpected indent\n"
     ]
    }
   ],
   "source": [
    "face = np.array([[res.x, res.y, res.z] for res in results.face_landmarks.landmark]).flatten() \n",
    "    if results.face_landmarks \n",
    "    else np.zeros(1404)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "17a25c3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_keypoints(results):\n",
    "    pose = np.array([[res.x, res.y, res.z, res.visibility] for res in results.pose_landmarks.landmark]).flatten() if results.pose_landmarks else np.zeros(33*4)\n",
    "    face = np.array([[res.x, res.y, res.z] for res in results.face_landmarks.landmark]).flatten() if results.face_landmarks else np.zeros(468*3)\n",
    "    lh = np.array([[res.x, res.y, res.z] for res in results.left_hand_landmarks.landmark]).flatten() if results.left_hand_landmarks else np.zeros(21*3)\n",
    "    rh = np.array([[res.x, res.y, res.z] for res in results.right_hand_landmarks.landmark]).flatten() if results.right_hand_landmarks else np.zeros(21*3)\n",
    "    return np.concatenate([pose, face, lh, rh])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "c446a77c",
   "metadata": {},
   "outputs": [],
   "source": [
    "result_test = extract_keypoints(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ebb955be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.50489241,  0.48859787, -0.41671914, ...,  0.        ,\n",
       "        0.        ,  0.        ])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "964eefea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1662"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "468*3+33*4+21*3+21*3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4e2454a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('0', result_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "eadbcfdb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.50489241,  0.48859787, -0.41671914, ...,  0.        ,\n",
       "        0.        ,  0.        ])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.load('0.npy')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "091a6a12",
   "metadata": {},
   "source": [
    "# Setup Folders for Collection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "051e512a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Path for exported data, numpy arrays\n",
    "DATA_PATH = os.path.join('MP_Data') \n",
    "\n",
    "# Actions that we try to detect\n",
    "actions = np.array(['hello', 'thanks', 'iloveyou'])\n",
    "\n",
    "# Thirty videos worth of data\n",
    "no_sequences = 30\n",
    "\n",
    "# Videos are going to be 30 frames in length\n",
    "sequence_length = 30"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "339217f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# hello\n",
    "## 0\n",
    "## 1\n",
    "## 2\n",
    "## ...\n",
    "## 29\n",
    "# thanks\n",
    "\n",
    "# I love you"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "e5251edf",
   "metadata": {},
   "outputs": [],
   "source": [
    "for action in actions: \n",
    "    for sequence in range(no_sequences):\n",
    "        try: \n",
    "            os.makedirs(os.path.join(DATA_PATH, action, str(sequence)))\n",
    "        except:\n",
    "            pass"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49b09585",
   "metadata": {},
   "source": [
    "# Collect Keypoint Values for Training and Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "8825e9d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "cap = cv2.VideoCapture(0)\n",
    "# Set mediapipe model \n",
    "with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:\n",
    "    \n",
    "    # NEW LOOP\n",
    "    # Loop through actions\n",
    "    for action in actions:\n",
    "        # Loop through sequences aka videos\n",
    "        for sequence in range(no_sequences):\n",
    "            # Loop through video length aka sequence length\n",
    "            for frame_num in range(sequence_length):\n",
    "\n",
    "                # Read feed\n",
    "                ret, frame = cap.read()\n",
    "\n",
    "                # Make detections\n",
    "                image, results = mediapipe_detection(frame, holistic)\n",
    "#                 print(results)\n",
    "\n",
    "                # Draw landmarks\n",
    "                draw_styled_landmarks(image, results)\n",
    "                \n",
    "                # NEW Apply wait logic\n",
    "                if frame_num == 0: \n",
    "                    cv2.putText(image, 'STARTING COLLECTION', (120,200), \n",
    "                               cv2.FONT_HERSHEY_SIMPLEX, 1, (0,255, 0), 4, cv2.LINE_AA)\n",
    "                    cv2.putText(image, 'Collecting frames for {} Video Number {}'.format(action, sequence), (15,12), \n",
    "                               cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1, cv2.LINE_AA)\n",
    "                    # Show to screen\n",
    "                    cv2.imshow('OpenCV Feed', image)\n",
    "                    cv2.waitKey(2000)\n",
    "                else: \n",
    "                    cv2.putText(image, 'Collecting frames for {} Video Number {}'.format(action, sequence), (15,12), \n",
    "                               cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1, cv2.LINE_AA)\n",
    "                    # Show to screen\n",
    "                    cv2.imshow('OpenCV Feed', image)\n",
    "                \n",
    "                # NEW Export keypoints\n",
    "                keypoints = extract_keypoints(results)\n",
    "                npy_path = os.path.join(DATA_PATH, action, str(sequence), str(frame_num))\n",
    "                np.save(npy_path, keypoints)\n",
    "\n",
    "                # Break gracefully\n",
    "                if cv2.waitKey(10) & 0xFF == ord('q'):\n",
    "                    break\n",
    "                    \n",
    "    cap.release()\n",
    "    cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "5c198678",
   "metadata": {},
   "outputs": [],
   "source": [
    "cap.release()\n",
    "cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d338470f",
   "metadata": {},
   "source": [
    "# Preprocess Data and Create Labels and Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "748b0795",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "63932656",
   "metadata": {},
   "outputs": [],
   "source": [
    "label_map = {label:num for num, label in enumerate(actions)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "0d1871a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'hello': 0, 'thanks': 1, 'iloveyou': 2}"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_map"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5caf568d",
   "metadata": {},
   "source": [
    "# sequences, labels = [], []\n",
    "for action in actions:\n",
    "    for sequence in range(no_sequences):\n",
    "        window = []\n",
    "        for frame_num in range(sequence_length):\n",
    "            res = np.load(os.path.join(DATA_PATH, action, str(sequence), \"0.npy\".format(frame_num)))\n",
    "            window.append(res)\n",
    "        sequences.append(window)\n",
    "        labels.append(label_map[action])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "6d690287",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(90, 30, 1662)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(sequences).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "16862903",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(90,)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "np.array(labels).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "bd5f32a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.array(sequences)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "7a13fc89",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(90, 30, 1662)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "c0facbd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "y = to_categorical(labels).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "7b1f792d",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "5e9fefbb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5, 3)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37b088a7",
   "metadata": {},
   "source": [
    "# Build and Train LSTM Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "dceb3e5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import LSTM, Dense\n",
    "from tensorflow.keras.callbacks import TensorBoard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "62eed6e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "log_dir = os.path.join('Logs')\n",
    "tb_callback = TensorBoard(log_dir=log_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "13df65a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM(64, return_sequences=True, activation='relu', input_shape=(30,1662)))\n",
    "model.add(LSTM(128, return_sequences=True, activation='relu'))\n",
    "model.add(LSTM(64, return_sequences=False, activation='relu'))\n",
    "model.add(Dense(64, activation='relu'))\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(actions.shape[0], activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "4b8d5a6d",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = [.7, 0.2, 0.1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "0c1cb0fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'hello'"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actions[np.argmax(res)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "376f35b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='Adam', loss='categorical_crossentropy', metrics=['categorical_accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "5b1c65a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/2000\n",
      "3/3 [==============================] - 9s 638ms/step - loss: 1.2209 - categorical_accuracy: 0.3698\n",
      "Epoch 2/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 1.1520 - categorical_accuracy: 0.3249\n",
      "Epoch 3/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.2215 - categorical_accuracy: 0.3034\n",
      "Epoch 4/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.1885 - categorical_accuracy: 0.3679\n",
      "Epoch 5/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.0872 - categorical_accuracy: 0.3854\n",
      "Epoch 6/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.0569 - categorical_accuracy: 0.3425\n",
      "Epoch 7/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.0084 - categorical_accuracy: 0.3405\n",
      "Epoch 8/2000\n",
      "3/3 [==============================] - 0s 155ms/step - loss: 0.9901 - categorical_accuracy: 0.3581\n",
      "Epoch 9/2000\n",
      "3/3 [==============================] - 0s 147ms/step - loss: 1.1316 - categorical_accuracy: 0.2056\n",
      "Epoch 10/2000\n",
      "3/3 [==============================] - 0s 150ms/step - loss: 1.1207 - categorical_accuracy: 0.3620\n",
      "Epoch 11/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.1191 - categorical_accuracy: 0.3327\n",
      "Epoch 12/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.1054 - categorical_accuracy: 0.3523\n",
      "Epoch 13/2000\n",
      "3/3 [==============================] - 0s 150ms/step - loss: 1.0993 - categorical_accuracy: 0.3405\n",
      "Epoch 14/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.1030 - categorical_accuracy: 0.3327\n",
      "Epoch 15/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0844 - categorical_accuracy: 0.4167\n",
      "Epoch 16/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0791 - categorical_accuracy: 0.4207\n",
      "Epoch 17/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0714 - categorical_accuracy: 0.3131\n",
      "Epoch 18/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0636 - categorical_accuracy: 0.3268\n",
      "Epoch 19/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0279 - categorical_accuracy: 0.3522\n",
      "Epoch 20/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9818 - categorical_accuracy: 0.3893\n",
      "Epoch 21/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0191 - categorical_accuracy: 0.4873\n",
      "Epoch 22/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0051 - categorical_accuracy: 0.4736\n",
      "Epoch 23/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.1150 - categorical_accuracy: 0.3503\n",
      "Epoch 24/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0358 - categorical_accuracy: 0.5400\n",
      "Epoch 25/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 1.0261 - categorical_accuracy: 0.4227\n",
      "Epoch 26/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0327 - categorical_accuracy: 0.3034\n",
      "Epoch 27/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.0499 - categorical_accuracy: 0.4109\n",
      "Epoch 28/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9934 - categorical_accuracy: 0.3307\n",
      "Epoch 29/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9933 - categorical_accuracy: 0.3738\n",
      "Epoch 30/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9545 - categorical_accuracy: 0.5930\n",
      "Epoch 31/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 0.9614 - categorical_accuracy: 0.6653\n",
      "Epoch 32/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9469 - categorical_accuracy: 0.5010\n",
      "Epoch 33/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9402 - categorical_accuracy: 0.4755\n",
      "Epoch 34/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.8948 - categorical_accuracy: 0.5420\n",
      "Epoch 35/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9038 - categorical_accuracy: 0.5675\n",
      "Epoch 36/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9052 - categorical_accuracy: 0.5597\n",
      "Epoch 37/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9088 - categorical_accuracy: 0.5264\n",
      "Epoch 38/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.8957 - categorical_accuracy: 0.4658\n",
      "Epoch 39/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9248 - categorical_accuracy: 0.5166\n",
      "Epoch 40/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9649 - categorical_accuracy: 0.4403\n",
      "Epoch 41/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9452 - categorical_accuracy: 0.5068\n",
      "Epoch 42/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9203 - categorical_accuracy: 0.4893\n",
      "Epoch 43/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.9023 - categorical_accuracy: 0.5362\n",
      "Epoch 44/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8953 - categorical_accuracy: 0.5518\n",
      "Epoch 45/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8976 - categorical_accuracy: 0.4638\n",
      "Epoch 46/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.8584 - categorical_accuracy: 0.6027\n",
      "Epoch 47/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9192 - categorical_accuracy: 0.5323\n",
      "Epoch 48/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.8674 - categorical_accuracy: 0.5499\n",
      "Epoch 49/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.8683 - categorical_accuracy: 0.5499\n",
      "Epoch 50/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8508 - categorical_accuracy: 0.4834\n",
      "Epoch 51/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7969 - categorical_accuracy: 0.5226\n",
      "Epoch 52/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.7676 - categorical_accuracy: 0.6066\n",
      "Epoch 53/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7873 - categorical_accuracy: 0.5832\n",
      "Epoch 54/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8081 - categorical_accuracy: 0.5421\n",
      "Epoch 55/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7798 - categorical_accuracy: 0.5852\n",
      "Epoch 56/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 3.1008 - categorical_accuracy: 0.3405\n",
      "Epoch 57/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0512 - categorical_accuracy: 0.5283\n",
      "Epoch 58/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0640 - categorical_accuracy: 0.3210\n",
      "Epoch 59/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0701 - categorical_accuracy: 0.2994\n",
      "Epoch 60/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0555 - categorical_accuracy: 0.3561\n",
      "Epoch 61/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.0874 - categorical_accuracy: 0.3307\n",
      "Epoch 62/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0976 - categorical_accuracy: 0.3347\n",
      "Epoch 63/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0979 - categorical_accuracy: 0.3014\n",
      "Epoch 64/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0964 - categorical_accuracy: 0.3640\n",
      "Epoch 65/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0954 - categorical_accuracy: 0.5341\n",
      "Epoch 66/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0939 - categorical_accuracy: 0.3405\n",
      "Epoch 67/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0915 - categorical_accuracy: 0.3874\n",
      "Epoch 68/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0860 - categorical_accuracy: 0.4108\n",
      "Epoch 69/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0782 - categorical_accuracy: 0.3327\n",
      "Epoch 70/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.0316 - categorical_accuracy: 0.4109\n",
      "Epoch 71/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0891 - categorical_accuracy: 0.2721\n",
      "Epoch 72/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.0922 - categorical_accuracy: 0.3112\n",
      "Epoch 73/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0859 - categorical_accuracy: 0.3425\n",
      "Epoch 74/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0830 - categorical_accuracy: 0.3620\n",
      "Epoch 75/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0830 - categorical_accuracy: 0.3268\n",
      "Epoch 76/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0621 - categorical_accuracy: 0.3776\n",
      "Epoch 77/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0757 - categorical_accuracy: 0.3151\n",
      "Epoch 78/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0433 - categorical_accuracy: 0.3483\n",
      "Epoch 79/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0350 - categorical_accuracy: 0.3483\n",
      "Epoch 80/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0803 - categorical_accuracy: 0.3798\n",
      "Epoch 81/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0418 - categorical_accuracy: 0.4696\n",
      "Epoch 82/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0464 - categorical_accuracy: 0.3914\n",
      "Epoch 83/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0694 - categorical_accuracy: 0.5244\n",
      "Epoch 84/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0875 - categorical_accuracy: 0.4207\n",
      "Epoch 85/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0860 - categorical_accuracy: 0.4071\n",
      "Epoch 86/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0679 - categorical_accuracy: 0.6320\n",
      "Epoch 87/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0388 - categorical_accuracy: 0.4676\n",
      "Epoch 88/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0255 - categorical_accuracy: 0.3718\n",
      "Epoch 89/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 1.0182 - categorical_accuracy: 0.3718\n",
      "Epoch 90/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0112 - categorical_accuracy: 0.4932\n",
      "Epoch 91/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0009 - categorical_accuracy: 0.5793\n",
      "Epoch 92/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9921 - categorical_accuracy: 0.5108\n",
      "Epoch 93/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9235 - categorical_accuracy: 0.5773\n",
      "Epoch 94/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0392 - categorical_accuracy: 0.4345\n",
      "Epoch 95/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0626 - categorical_accuracy: 0.4110\n",
      "Epoch 96/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9406 - categorical_accuracy: 0.6223\n",
      "Epoch 97/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8873 - categorical_accuracy: 0.5597\n",
      "Epoch 98/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9494 - categorical_accuracy: 0.5499\n",
      "Epoch 99/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9118 - categorical_accuracy: 0.6320\n",
      "Epoch 100/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9147 - categorical_accuracy: 0.6067\n",
      "Epoch 101/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8844 - categorical_accuracy: 0.5988\n",
      "Epoch 102/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8846 - categorical_accuracy: 0.5930\n",
      "Epoch 103/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8436 - categorical_accuracy: 0.6380\n",
      "Epoch 104/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8450 - categorical_accuracy: 0.5676\n",
      "Epoch 105/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.8723 - categorical_accuracy: 0.5891\n",
      "Epoch 106/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8660 - categorical_accuracy: 0.5538\n",
      "Epoch 107/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8310 - categorical_accuracy: 0.6595\n",
      "Epoch 108/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7958 - categorical_accuracy: 0.6693\n",
      "Epoch 109/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8091 - categorical_accuracy: 0.5989\n",
      "Epoch 110/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7919 - categorical_accuracy: 0.6654\n",
      "Epoch 111/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7099 - categorical_accuracy: 0.7338\n",
      "Epoch 112/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7667 - categorical_accuracy: 0.6183\n",
      "Epoch 113/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8189 - categorical_accuracy: 0.5812\n",
      "Epoch 114/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7144 - categorical_accuracy: 0.6674\n",
      "Epoch 115/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7045 - categorical_accuracy: 0.6340\n",
      "Epoch 116/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6742 - categorical_accuracy: 0.6771\n",
      "Epoch 117/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7273 - categorical_accuracy: 0.5714\n",
      "Epoch 118/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7078 - categorical_accuracy: 0.6556\n",
      "Epoch 119/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6883 - categorical_accuracy: 0.6380\n",
      "Epoch 120/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.6506 - categorical_accuracy: 0.6516\n",
      "Epoch 121/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6178 - categorical_accuracy: 0.6066\n",
      "Epoch 122/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6047 - categorical_accuracy: 0.6399\n",
      "Epoch 123/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5609 - categorical_accuracy: 0.7084\n",
      "Epoch 124/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6453 - categorical_accuracy: 0.6614\n",
      "Epoch 125/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6577 - categorical_accuracy: 0.6438\n",
      "Epoch 126/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6458 - categorical_accuracy: 0.6262\n",
      "Epoch 127/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6640 - categorical_accuracy: 0.5460\n",
      "Epoch 128/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5905 - categorical_accuracy: 0.7299\n",
      "Epoch 129/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.6462 - categorical_accuracy: 0.6869\n",
      "Epoch 130/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5843 - categorical_accuracy: 0.7006\n",
      "Epoch 131/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5855 - categorical_accuracy: 0.6869\n",
      "Epoch 132/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6109 - categorical_accuracy: 0.6772\n",
      "Epoch 133/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5735 - categorical_accuracy: 0.8278\n",
      "Epoch 134/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5924 - categorical_accuracy: 0.7104\n",
      "Epoch 135/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.6001 - categorical_accuracy: 0.7828\n",
      "Epoch 136/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.5706 - categorical_accuracy: 0.7710\n",
      "Epoch 137/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.6681 - categorical_accuracy: 0.6927\n",
      "Epoch 138/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.6290 - categorical_accuracy: 0.6751\n",
      "Epoch 139/2000\n",
      "3/3 [==============================] - 1s 164ms/step - loss: 0.7622 - categorical_accuracy: 0.5773\n",
      "Epoch 140/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7738 - categorical_accuracy: 0.5402\n",
      "Epoch 141/2000\n",
      "3/3 [==============================] - 0s 159ms/step - loss: 0.6277 - categorical_accuracy: 0.6282\n",
      "Epoch 142/2000\n",
      "3/3 [==============================] - 0s 156ms/step - loss: 0.6358 - categorical_accuracy: 0.7457\n",
      "Epoch 143/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.6406 - categorical_accuracy: 0.6380\n",
      "Epoch 144/2000\n",
      "3/3 [==============================] - 0s 156ms/step - loss: 0.5545 - categorical_accuracy: 0.7397\n",
      "Epoch 145/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 148ms/step - loss: 0.5734 - categorical_accuracy: 0.7416\n",
      "Epoch 146/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.5048 - categorical_accuracy: 0.7553\n",
      "Epoch 147/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5954 - categorical_accuracy: 0.7319\n",
      "Epoch 148/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.6424 - categorical_accuracy: 0.6438\n",
      "Epoch 149/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6753 - categorical_accuracy: 0.6537\n",
      "Epoch 150/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6344 - categorical_accuracy: 0.6438\n",
      "Epoch 151/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5338 - categorical_accuracy: 0.7515\n",
      "Epoch 152/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5286 - categorical_accuracy: 0.7045\n",
      "Epoch 153/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7627 - categorical_accuracy: 0.6966\n",
      "Epoch 154/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8099 - categorical_accuracy: 0.6320\n",
      "Epoch 155/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.6725 - categorical_accuracy: 0.6771\n",
      "Epoch 156/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6297 - categorical_accuracy: 0.6203\n",
      "Epoch 157/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7135 - categorical_accuracy: 0.6145\n",
      "Epoch 158/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6645 - categorical_accuracy: 0.6340\n",
      "Epoch 159/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.6711 - categorical_accuracy: 0.5852\n",
      "Epoch 160/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6430 - categorical_accuracy: 0.5949\n",
      "Epoch 161/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5818 - categorical_accuracy: 0.7378\n",
      "Epoch 162/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.5660 - categorical_accuracy: 0.7847\n",
      "Epoch 163/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5288 - categorical_accuracy: 0.7182\n",
      "Epoch 164/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5549 - categorical_accuracy: 0.6869\n",
      "Epoch 165/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5584 - categorical_accuracy: 0.7398\n",
      "Epoch 166/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.5623 - categorical_accuracy: 0.6751\n",
      "Epoch 167/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.4639 - categorical_accuracy: 0.8434\n",
      "Epoch 168/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.4607 - categorical_accuracy: 0.8336\n",
      "Epoch 169/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5383 - categorical_accuracy: 0.7593\n",
      "Epoch 170/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.4666 - categorical_accuracy: 0.8493\n",
      "Epoch 171/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5352 - categorical_accuracy: 0.7163\n",
      "Epoch 172/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4677 - categorical_accuracy: 0.8121\n",
      "Epoch 173/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4792 - categorical_accuracy: 0.7710\n",
      "Epoch 174/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4176 - categorical_accuracy: 0.8024\n",
      "Epoch 175/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4192 - categorical_accuracy: 0.8102\n",
      "Epoch 176/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4676 - categorical_accuracy: 0.8121\n",
      "Epoch 177/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4476 - categorical_accuracy: 0.7808\n",
      "Epoch 178/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9514 - categorical_accuracy: 0.6557\n",
      "Epoch 179/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6267 - categorical_accuracy: 0.6987\n",
      "Epoch 180/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.6666 - categorical_accuracy: 0.6849\n",
      "Epoch 181/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5133 - categorical_accuracy: 0.7338\n",
      "Epoch 182/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5356 - categorical_accuracy: 0.6889\n",
      "Epoch 183/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5604 - categorical_accuracy: 0.7476\n",
      "Epoch 184/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.5431 - categorical_accuracy: 0.7378\n",
      "Epoch 185/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4499 - categorical_accuracy: 0.8199\n",
      "Epoch 186/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4852 - categorical_accuracy: 0.7848\n",
      "Epoch 187/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5219 - categorical_accuracy: 0.7750\n",
      "Epoch 188/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5469 - categorical_accuracy: 0.7417\n",
      "Epoch 189/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4446 - categorical_accuracy: 0.8787\n",
      "Epoch 190/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3441 - categorical_accuracy: 0.9060\n",
      "Epoch 191/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4011 - categorical_accuracy: 0.8533\n",
      "Epoch 192/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.3619 - categorical_accuracy: 0.8982\n",
      "Epoch 193/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3892 - categorical_accuracy: 0.8493\n",
      "Epoch 194/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3969 - categorical_accuracy: 0.8572\n",
      "Epoch 195/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3762 - categorical_accuracy: 0.8748\n",
      "Epoch 196/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2988 - categorical_accuracy: 0.8689\n",
      "Epoch 197/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4521 - categorical_accuracy: 0.8239\n",
      "Epoch 198/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3457 - categorical_accuracy: 0.8767\n",
      "Epoch 199/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4271 - categorical_accuracy: 0.8102\n",
      "Epoch 200/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4467 - categorical_accuracy: 0.8004\n",
      "Epoch 201/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.3903 - categorical_accuracy: 0.8552\n",
      "Epoch 202/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3938 - categorical_accuracy: 0.8239\n",
      "Epoch 203/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.5351 - categorical_accuracy: 0.7710\n",
      "Epoch 204/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4736 - categorical_accuracy: 0.8083\n",
      "Epoch 205/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5048 - categorical_accuracy: 0.7750\n",
      "Epoch 206/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3253 - categorical_accuracy: 0.8669\n",
      "Epoch 207/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4055 - categorical_accuracy: 0.8318\n",
      "Epoch 208/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3156 - categorical_accuracy: 0.9002\n",
      "Epoch 209/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2987 - categorical_accuracy: 0.9100\n",
      "Epoch 210/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3101 - categorical_accuracy: 0.9041\n",
      "Epoch 211/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3000 - categorical_accuracy: 0.9022\n",
      "Epoch 212/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2815 - categorical_accuracy: 0.8807\n",
      "Epoch 213/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2328 - categorical_accuracy: 0.9119\n",
      "Epoch 214/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2354 - categorical_accuracy: 0.9198\n",
      "Epoch 215/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2358 - categorical_accuracy: 0.9256\n",
      "Epoch 216/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2612 - categorical_accuracy: 0.9040\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 217/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.3979 - categorical_accuracy: 0.8180\n",
      "Epoch 218/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.5742 - categorical_accuracy: 0.7612\n",
      "Epoch 219/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5949 - categorical_accuracy: 0.8004\n",
      "Epoch 220/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.5871 - categorical_accuracy: 0.7261\n",
      "Epoch 221/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6265 - categorical_accuracy: 0.7574\n",
      "Epoch 222/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8385 - categorical_accuracy: 0.6359\n",
      "Epoch 223/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5284 - categorical_accuracy: 0.7377\n",
      "Epoch 224/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5162 - categorical_accuracy: 0.7750\n",
      "Epoch 225/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5938 - categorical_accuracy: 0.6966\n",
      "Epoch 226/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5185 - categorical_accuracy: 0.7593\n",
      "Epoch 227/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4787 - categorical_accuracy: 0.7807\n",
      "Epoch 228/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.3948 - categorical_accuracy: 0.8239\n",
      "Epoch 229/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4177 - categorical_accuracy: 0.8375\n",
      "Epoch 230/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4862 - categorical_accuracy: 0.7887\n",
      "Epoch 231/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4810 - categorical_accuracy: 0.8102\n",
      "Epoch 232/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.3348 - categorical_accuracy: 0.9041\n",
      "Epoch 233/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.4288 - categorical_accuracy: 0.8318\n",
      "Epoch 234/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.3671 - categorical_accuracy: 0.9237\n",
      "Epoch 235/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3204 - categorical_accuracy: 0.9472\n",
      "Epoch 236/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.3220 - categorical_accuracy: 0.9119\n",
      "Epoch 237/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3120 - categorical_accuracy: 0.9472\n",
      "Epoch 238/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.2564 - categorical_accuracy: 0.9335\n",
      "Epoch 239/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2671 - categorical_accuracy: 0.9217\n",
      "Epoch 240/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2856 - categorical_accuracy: 0.9374\n",
      "Epoch 241/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2815 - categorical_accuracy: 0.9354\n",
      "Epoch 242/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2872 - categorical_accuracy: 0.9119\n",
      "Epoch 243/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3035 - categorical_accuracy: 0.9080\n",
      "Epoch 244/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2558 - categorical_accuracy: 0.9080\n",
      "Epoch 245/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.2237 - categorical_accuracy: 0.9335\n",
      "Epoch 246/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2427 - categorical_accuracy: 0.9198\n",
      "Epoch 247/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1878 - categorical_accuracy: 0.9491\n",
      "Epoch 248/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.1962 - categorical_accuracy: 0.9511\n",
      "Epoch 249/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2289 - categorical_accuracy: 0.9217\n",
      "Epoch 250/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.2353 - categorical_accuracy: 0.9002\n",
      "Epoch 251/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3517 - categorical_accuracy: 0.9080\n",
      "Epoch 252/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1183 - categorical_accuracy: 0.6752\n",
      "Epoch 253/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.5382 - categorical_accuracy: 0.8160\n",
      "Epoch 254/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6545 - categorical_accuracy: 0.6594\n",
      "Epoch 255/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5385 - categorical_accuracy: 0.8610\n",
      "Epoch 256/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.5151 - categorical_accuracy: 0.7985\n",
      "Epoch 257/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4551 - categorical_accuracy: 0.8709\n",
      "Epoch 258/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4152 - categorical_accuracy: 0.9061\n",
      "Epoch 259/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.4417 - categorical_accuracy: 0.8160\n",
      "Epoch 260/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4534 - categorical_accuracy: 0.8122\n",
      "Epoch 261/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.3907 - categorical_accuracy: 0.8474\n",
      "Epoch 262/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3205 - categorical_accuracy: 0.8533\n",
      "Epoch 263/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.3034 - categorical_accuracy: 0.9100\n",
      "Epoch 264/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3085 - categorical_accuracy: 0.9374\n",
      "Epoch 265/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.2291 - categorical_accuracy: 0.9569\n",
      "Epoch 266/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2196 - categorical_accuracy: 0.9374\n",
      "Epoch 267/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1827 - categorical_accuracy: 0.9452\n",
      "Epoch 268/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3432 - categorical_accuracy: 0.8474\n",
      "Epoch 269/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.2653 - categorical_accuracy: 0.9198\n",
      "Epoch 270/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2676 - categorical_accuracy: 0.9002\n",
      "Epoch 271/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2290 - categorical_accuracy: 0.9276\n",
      "Epoch 272/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2696 - categorical_accuracy: 0.8983\n",
      "Epoch 273/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.2376 - categorical_accuracy: 0.8885\n",
      "Epoch 274/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1869 - categorical_accuracy: 0.9472\n",
      "Epoch 275/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1735 - categorical_accuracy: 0.9432\n",
      "Epoch 276/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1846 - categorical_accuracy: 0.9530\n",
      "Epoch 277/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1576 - categorical_accuracy: 0.9530\n",
      "Epoch 278/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2336 - categorical_accuracy: 0.9080\n",
      "Epoch 279/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1712 - categorical_accuracy: 0.9550\n",
      "Epoch 280/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1800 - categorical_accuracy: 0.9530\n",
      "Epoch 281/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2110 - categorical_accuracy: 0.9237\n",
      "Epoch 282/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2454 - categorical_accuracy: 0.9061\n",
      "Epoch 283/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1371 - categorical_accuracy: 0.9667\n",
      "Epoch 284/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1542 - categorical_accuracy: 0.9393\n",
      "Epoch 285/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1165 - categorical_accuracy: 0.9648\n",
      "Epoch 286/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2020 - categorical_accuracy: 0.9080\n",
      "Epoch 287/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2381 - categorical_accuracy: 0.9119\n",
      "Epoch 288/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2180 - categorical_accuracy: 0.9295\n",
      "Epoch 289/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1319 - categorical_accuracy: 0.9569\n",
      "Epoch 290/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4505 - categorical_accuracy: 0.8650\n",
      "Epoch 291/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2215 - categorical_accuracy: 0.9119\n",
      "Epoch 292/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2831 - categorical_accuracy: 0.9040\n",
      "Epoch 293/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3464 - categorical_accuracy: 0.8688\n",
      "Epoch 294/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2390 - categorical_accuracy: 0.9256\n",
      "Epoch 295/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2285 - categorical_accuracy: 0.9315\n",
      "Epoch 296/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2289 - categorical_accuracy: 0.9041\n",
      "Epoch 297/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2933 - categorical_accuracy: 0.9100\n",
      "Epoch 298/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.1860 - categorical_accuracy: 0.9706\n",
      "Epoch 299/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.2459 - categorical_accuracy: 0.9217\n",
      "Epoch 300/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2259 - categorical_accuracy: 0.9374\n",
      "Epoch 301/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.2024 - categorical_accuracy: 0.9374\n",
      "Epoch 302/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2162 - categorical_accuracy: 0.9374\n",
      "Epoch 303/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1583 - categorical_accuracy: 0.9530\n",
      "Epoch 304/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1869 - categorical_accuracy: 0.9256\n",
      "Epoch 305/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1505 - categorical_accuracy: 0.9432\n",
      "Epoch 306/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.1145 - categorical_accuracy: 0.9687\n",
      "Epoch 307/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2205 - categorical_accuracy: 0.9276\n",
      "Epoch 308/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1430 - categorical_accuracy: 0.9491\n",
      "Epoch 309/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.6108 - categorical_accuracy: 0.7886\n",
      "Epoch 310/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.5743 - categorical_accuracy: 0.8160\n",
      "Epoch 311/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3295 - categorical_accuracy: 0.8884\n",
      "Epoch 312/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2550 - categorical_accuracy: 0.8904\n",
      "Epoch 313/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2107 - categorical_accuracy: 0.9511\n",
      "Epoch 314/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1494 - categorical_accuracy: 0.9648\n",
      "Epoch 315/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2462 - categorical_accuracy: 0.9315\n",
      "Epoch 316/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.2432 - categorical_accuracy: 0.9296\n",
      "Epoch 317/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1657 - categorical_accuracy: 0.9628\n",
      "Epoch 318/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1602 - categorical_accuracy: 0.9413\n",
      "Epoch 319/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1920 - categorical_accuracy: 0.9726\n",
      "Epoch 320/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1162 - categorical_accuracy: 0.9824\n",
      "Epoch 321/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1634 - categorical_accuracy: 0.9608\n",
      "Epoch 322/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1292 - categorical_accuracy: 0.9745\n",
      "Epoch 323/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1257 - categorical_accuracy: 0.9804\n",
      "Epoch 324/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.0983 - categorical_accuracy: 0.9589\n",
      "Epoch 325/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1123 - categorical_accuracy: 0.9628\n",
      "Epoch 326/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1358 - categorical_accuracy: 0.9765\n",
      "Epoch 327/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2451 - categorical_accuracy: 0.9159\n",
      "Epoch 328/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2315 - categorical_accuracy: 0.9159\n",
      "Epoch 329/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4046 - categorical_accuracy: 0.8650\n",
      "Epoch 330/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4353 - categorical_accuracy: 0.8513\n",
      "Epoch 331/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3789 - categorical_accuracy: 0.8199\n",
      "Epoch 332/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7230 - categorical_accuracy: 0.7319\n",
      "Epoch 333/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3347 - categorical_accuracy: 0.8747\n",
      "Epoch 334/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4459 - categorical_accuracy: 0.8375\n",
      "Epoch 335/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.4835 - categorical_accuracy: 0.8394\n",
      "Epoch 336/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3349 - categorical_accuracy: 0.8415\n",
      "Epoch 337/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4768 - categorical_accuracy: 0.7868\n",
      "Epoch 338/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3657 - categorical_accuracy: 0.8317\n",
      "Epoch 339/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.2327 - categorical_accuracy: 0.9432\n",
      "Epoch 340/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.2576 - categorical_accuracy: 0.9159\n",
      "Epoch 341/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2966 - categorical_accuracy: 0.9276\n",
      "Epoch 342/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2265 - categorical_accuracy: 0.9452\n",
      "Epoch 343/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.2442 - categorical_accuracy: 0.9550\n",
      "Epoch 344/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1880 - categorical_accuracy: 0.9393\n",
      "Epoch 345/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2180 - categorical_accuracy: 0.9550\n",
      "Epoch 346/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2049 - categorical_accuracy: 0.9726\n",
      "Epoch 347/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2092 - categorical_accuracy: 0.9335\n",
      "Epoch 348/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1742 - categorical_accuracy: 0.9608\n",
      "Epoch 349/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1876 - categorical_accuracy: 0.9628\n",
      "Epoch 350/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2444 - categorical_accuracy: 0.9217\n",
      "Epoch 351/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1633 - categorical_accuracy: 0.9452\n",
      "Epoch 352/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1237 - categorical_accuracy: 0.9608\n",
      "Epoch 353/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1379 - categorical_accuracy: 0.9393\n",
      "Epoch 354/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.1461 - categorical_accuracy: 0.9726\n",
      "Epoch 355/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1279 - categorical_accuracy: 0.9413\n",
      "Epoch 356/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.1549 - categorical_accuracy: 0.9608\n",
      "Epoch 357/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1115 - categorical_accuracy: 0.9667\n",
      "Epoch 358/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1433 - categorical_accuracy: 0.9491\n",
      "Epoch 359/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1807 - categorical_accuracy: 0.9256\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 360/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.2287 - categorical_accuracy: 0.9217\n",
      "Epoch 361/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2372 - categorical_accuracy: 0.9276\n",
      "Epoch 362/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.1878 - categorical_accuracy: 0.9491\n",
      "Epoch 363/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1405 - categorical_accuracy: 0.9628\n",
      "Epoch 364/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1632 - categorical_accuracy: 0.9452\n",
      "Epoch 365/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1371 - categorical_accuracy: 0.9628\n",
      "Epoch 366/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1447 - categorical_accuracy: 0.9530\n",
      "Epoch 367/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1213 - categorical_accuracy: 0.9589\n",
      "Epoch 368/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1533 - categorical_accuracy: 0.9530\n",
      "Epoch 369/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1203 - categorical_accuracy: 0.9491\n",
      "Epoch 370/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1452 - categorical_accuracy: 0.9452\n",
      "Epoch 371/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1196 - categorical_accuracy: 0.9550\n",
      "Epoch 372/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1079 - categorical_accuracy: 0.9843\n",
      "Epoch 373/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6727 - categorical_accuracy: 0.7926\n",
      "Epoch 374/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7108 - categorical_accuracy: 0.7476\n",
      "Epoch 375/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8107 - categorical_accuracy: 0.7202\n",
      "Epoch 376/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.6363 - categorical_accuracy: 0.7867\n",
      "Epoch 377/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.1561 - categorical_accuracy: 0.6183\n",
      "Epoch 378/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6551 - categorical_accuracy: 0.7318\n",
      "Epoch 379/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6469 - categorical_accuracy: 0.7143\n",
      "Epoch 380/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.6605 - categorical_accuracy: 0.7573\n",
      "Epoch 381/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4385 - categorical_accuracy: 0.8825\n",
      "Epoch 382/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.5132 - categorical_accuracy: 0.8317\n",
      "Epoch 383/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4759 - categorical_accuracy: 0.8689\n",
      "Epoch 384/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3805 - categorical_accuracy: 0.8787\n",
      "Epoch 385/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.3624 - categorical_accuracy: 0.8904\n",
      "Epoch 386/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.3431 - categorical_accuracy: 0.8982\n",
      "Epoch 387/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3564 - categorical_accuracy: 0.8767\n",
      "Epoch 388/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.2898 - categorical_accuracy: 0.8963\n",
      "Epoch 389/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2959 - categorical_accuracy: 0.8806\n",
      "Epoch 390/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3678 - categorical_accuracy: 0.8807\n",
      "Epoch 391/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.3267 - categorical_accuracy: 0.8963\n",
      "Epoch 392/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3049 - categorical_accuracy: 0.9041\n",
      "Epoch 393/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2869 - categorical_accuracy: 0.9217\n",
      "Epoch 394/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.2278 - categorical_accuracy: 0.9256\n",
      "Epoch 395/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2917 - categorical_accuracy: 0.8748\n",
      "Epoch 396/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2486 - categorical_accuracy: 0.9080\n",
      "Epoch 397/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2678 - categorical_accuracy: 0.9198\n",
      "Epoch 398/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.2293 - categorical_accuracy: 0.9080\n",
      "Epoch 399/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2658 - categorical_accuracy: 0.9159\n",
      "Epoch 400/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2205 - categorical_accuracy: 0.9198\n",
      "Epoch 401/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1913 - categorical_accuracy: 0.9413\n",
      "Epoch 402/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1857 - categorical_accuracy: 0.9452\n",
      "Epoch 403/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2344 - categorical_accuracy: 0.9276\n",
      "Epoch 404/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2383 - categorical_accuracy: 0.9100\n",
      "Epoch 405/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.1910 - categorical_accuracy: 0.9511\n",
      "Epoch 406/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1746 - categorical_accuracy: 0.9491\n",
      "Epoch 407/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1971 - categorical_accuracy: 0.9374\n",
      "Epoch 408/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2175 - categorical_accuracy: 0.9295\n",
      "Epoch 409/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2172 - categorical_accuracy: 0.9256\n",
      "Epoch 410/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2062 - categorical_accuracy: 0.9374\n",
      "Epoch 411/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1713 - categorical_accuracy: 0.9608\n",
      "Epoch 412/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1510 - categorical_accuracy: 0.9648\n",
      "Epoch 413/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1825 - categorical_accuracy: 0.9472\n",
      "Epoch 414/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2119 - categorical_accuracy: 0.9237\n",
      "Epoch 415/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.2096 - categorical_accuracy: 0.9491\n",
      "Epoch 416/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1693 - categorical_accuracy: 0.9608\n",
      "Epoch 417/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.2237 - categorical_accuracy: 0.9276\n",
      "Epoch 418/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2135 - categorical_accuracy: 0.9393\n",
      "Epoch 419/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1623 - categorical_accuracy: 0.9608\n",
      "Epoch 420/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7146 - categorical_accuracy: 0.8081\n",
      "Epoch 421/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.5182 - categorical_accuracy: 0.7944\n",
      "Epoch 422/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.4529 - categorical_accuracy: 0.8394\n",
      "Epoch 423/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.4450 - categorical_accuracy: 0.8435\n",
      "Epoch 424/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.4106 - categorical_accuracy: 0.8259\n",
      "Epoch 425/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.3018 - categorical_accuracy: 0.9217\n",
      "Epoch 426/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.2724 - categorical_accuracy: 0.9354\n",
      "Epoch 427/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2741 - categorical_accuracy: 0.9217\n",
      "Epoch 428/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2507 - categorical_accuracy: 0.9198\n",
      "Epoch 429/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2217 - categorical_accuracy: 0.9491\n",
      "Epoch 430/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2493 - categorical_accuracy: 0.9452\n",
      "Epoch 431/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2237 - categorical_accuracy: 0.9393\n",
      "Epoch 432/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.2444 - categorical_accuracy: 0.9315\n",
      "Epoch 433/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1761 - categorical_accuracy: 0.9511\n",
      "Epoch 434/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2036 - categorical_accuracy: 0.9432\n",
      "Epoch 435/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2633 - categorical_accuracy: 0.9374\n",
      "Epoch 436/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2254 - categorical_accuracy: 0.9276\n",
      "Epoch 437/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1900 - categorical_accuracy: 0.9491\n",
      "Epoch 438/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1596 - categorical_accuracy: 0.9706\n",
      "Epoch 439/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2369 - categorical_accuracy: 0.9452\n",
      "Epoch 440/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1790 - categorical_accuracy: 0.9335\n",
      "Epoch 441/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1875 - categorical_accuracy: 0.9667\n",
      "Epoch 442/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2329 - categorical_accuracy: 0.9119\n",
      "Epoch 443/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2103 - categorical_accuracy: 0.9589\n",
      "Epoch 444/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1954 - categorical_accuracy: 0.9452\n",
      "Epoch 445/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2110 - categorical_accuracy: 0.9315\n",
      "Epoch 446/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.2195 - categorical_accuracy: 0.9393\n",
      "Epoch 447/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2374 - categorical_accuracy: 0.9041\n",
      "Epoch 448/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1914 - categorical_accuracy: 0.9530\n",
      "Epoch 449/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1309 - categorical_accuracy: 0.9628\n",
      "Epoch 450/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1885 - categorical_accuracy: 0.9413\n",
      "Epoch 451/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1782 - categorical_accuracy: 0.9432\n",
      "Epoch 452/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.2746 - categorical_accuracy: 0.9178\n",
      "Epoch 453/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.3025 - categorical_accuracy: 0.8943\n",
      "Epoch 454/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2316 - categorical_accuracy: 0.9276\n",
      "Epoch 455/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.2302 - categorical_accuracy: 0.9198\n",
      "Epoch 456/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1331 - categorical_accuracy: 0.9511\n",
      "Epoch 457/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2053 - categorical_accuracy: 0.9472\n",
      "Epoch 458/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1888 - categorical_accuracy: 0.9393\n",
      "Epoch 459/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2136 - categorical_accuracy: 0.9159\n",
      "Epoch 460/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2051 - categorical_accuracy: 0.9315\n",
      "Epoch 461/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2189 - categorical_accuracy: 0.9335\n",
      "Epoch 462/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2700 - categorical_accuracy: 0.9021\n",
      "Epoch 463/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1687 - categorical_accuracy: 0.9432\n",
      "Epoch 464/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.4241 - categorical_accuracy: 0.8337\n",
      "Epoch 465/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2930 - categorical_accuracy: 0.9100\n",
      "Epoch 466/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1995 - categorical_accuracy: 0.9296\n",
      "Epoch 467/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2537 - categorical_accuracy: 0.9237\n",
      "Epoch 468/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1855 - categorical_accuracy: 0.9295\n",
      "Epoch 469/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2221 - categorical_accuracy: 0.9099\n",
      "Epoch 470/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.1913 - categorical_accuracy: 0.9374\n",
      "Epoch 471/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2318 - categorical_accuracy: 0.9217\n",
      "Epoch 472/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.2542 - categorical_accuracy: 0.9061\n",
      "Epoch 473/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1885 - categorical_accuracy: 0.9569\n",
      "Epoch 474/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1902 - categorical_accuracy: 0.9315\n",
      "Epoch 475/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1912 - categorical_accuracy: 0.9217\n",
      "Epoch 476/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.2188 - categorical_accuracy: 0.9452\n",
      "Epoch 477/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1682 - categorical_accuracy: 0.9354\n",
      "Epoch 478/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1294 - categorical_accuracy: 0.9648\n",
      "Epoch 479/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1478 - categorical_accuracy: 0.9472\n",
      "Epoch 480/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1157 - categorical_accuracy: 0.9550\n",
      "Epoch 481/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1311 - categorical_accuracy: 0.9491\n",
      "Epoch 482/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1394 - categorical_accuracy: 0.9667\n",
      "Epoch 483/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1535 - categorical_accuracy: 0.9530\n",
      "Epoch 484/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1473 - categorical_accuracy: 0.9315\n",
      "Epoch 485/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1819 - categorical_accuracy: 0.9472\n",
      "Epoch 486/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1567 - categorical_accuracy: 0.9589\n",
      "Epoch 487/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.1353 - categorical_accuracy: 0.9393\n",
      "Epoch 488/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1913 - categorical_accuracy: 0.9315\n",
      "Epoch 489/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1581 - categorical_accuracy: 0.9315\n",
      "Epoch 490/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1962 - categorical_accuracy: 0.9315\n",
      "Epoch 491/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1611 - categorical_accuracy: 0.9530\n",
      "Epoch 492/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2282 - categorical_accuracy: 0.9120\n",
      "Epoch 493/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1486 - categorical_accuracy: 0.9589\n",
      "Epoch 494/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1444 - categorical_accuracy: 0.9550\n",
      "Epoch 495/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.1362 - categorical_accuracy: 0.9452\n",
      "Epoch 496/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1407 - categorical_accuracy: 0.9589\n",
      "Epoch 497/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1179 - categorical_accuracy: 0.9667\n",
      "Epoch 498/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.1543 - categorical_accuracy: 0.9374\n",
      "Epoch 499/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1093 - categorical_accuracy: 0.9608\n",
      "Epoch 500/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.0837 - categorical_accuracy: 0.9784\n",
      "Epoch 501/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1290 - categorical_accuracy: 0.9530\n",
      "Epoch 502/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.1355 - categorical_accuracy: 0.9452\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 503/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1195 - categorical_accuracy: 0.9608\n",
      "Epoch 504/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.0864 - categorical_accuracy: 0.9648\n",
      "Epoch 505/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1018 - categorical_accuracy: 0.9706\n",
      "Epoch 506/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1148 - categorical_accuracy: 0.9530\n",
      "Epoch 507/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.3630 - categorical_accuracy: 0.8533\n",
      "Epoch 508/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.3712 - categorical_accuracy: 0.8631\n",
      "Epoch 509/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1769 - categorical_accuracy: 0.9276\n",
      "Epoch 510/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1645 - categorical_accuracy: 0.9432\n",
      "Epoch 511/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.1191 - categorical_accuracy: 0.9393\n",
      "Epoch 512/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1417 - categorical_accuracy: 0.9452\n",
      "Epoch 513/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1537 - categorical_accuracy: 0.9413\n",
      "Epoch 514/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1304 - categorical_accuracy: 0.9511\n",
      "Epoch 515/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1384 - categorical_accuracy: 0.9491\n",
      "Epoch 516/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1264 - categorical_accuracy: 0.9491\n",
      "Epoch 517/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1110 - categorical_accuracy: 0.9530\n",
      "Epoch 518/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1627 - categorical_accuracy: 0.9452\n",
      "Epoch 519/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1140 - categorical_accuracy: 0.9569\n",
      "Epoch 520/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1330 - categorical_accuracy: 0.9491\n",
      "Epoch 521/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.0894 - categorical_accuracy: 0.9687\n",
      "Epoch 522/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1107 - categorical_accuracy: 0.9452\n",
      "Epoch 523/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.1133 - categorical_accuracy: 0.9765\n",
      "Epoch 524/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.1016 - categorical_accuracy: 0.9530\n",
      "Epoch 525/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1150 - categorical_accuracy: 0.9726\n",
      "Epoch 526/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.1134 - categorical_accuracy: 0.9452\n",
      "Epoch 527/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.1508 - categorical_accuracy: 0.9296\n",
      "Epoch 528/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.1110 - categorical_accuracy: 0.9726\n",
      "Epoch 529/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.0964 - categorical_accuracy: 0.9667\n",
      "Epoch 530/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.0865 - categorical_accuracy: 0.9804\n",
      "Epoch 531/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.1219 - categorical_accuracy: 0.9589\n",
      "Epoch 532/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.0973 - categorical_accuracy: 0.9726\n",
      "Epoch 533/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.0950 - categorical_accuracy: 0.9765\n",
      "Epoch 534/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.0901 - categorical_accuracy: 0.9765\n",
      "Epoch 535/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.0664 - categorical_accuracy: 0.9843\n",
      "Epoch 536/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.2980 - categorical_accuracy: 0.9080\n",
      "Epoch 537/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.5137 - categorical_accuracy: 0.8591\n",
      "Epoch 538/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.7020 - categorical_accuracy: 0.6262\n",
      "Epoch 539/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 2.0719 - categorical_accuracy: 0.5459\n",
      "Epoch 540/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 8.6369 - categorical_accuracy: 0.3620\n",
      "Epoch 541/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 18.7519 - categorical_accuracy: 0.4070\n",
      "Epoch 542/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 14.9310 - categorical_accuracy: 0.3856\n",
      "Epoch 543/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 57.0404 - categorical_accuracy: 0.4130\n",
      "Epoch 544/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 23.7210 - categorical_accuracy: 0.3386\n",
      "Epoch 545/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 20.0459 - categorical_accuracy: 0.3484\n",
      "Epoch 546/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 148.1293 - categorical_accuracy: 0.4892\n",
      "Epoch 547/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 85.0117 - categorical_accuracy: 0.3209\n",
      "Epoch 548/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 61.3025 - categorical_accuracy: 0.4091\n",
      "Epoch 549/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 30.9410 - categorical_accuracy: 0.3620\n",
      "Epoch 550/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 10.0802 - categorical_accuracy: 0.2622\n",
      "Epoch 551/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 327.9776 - categorical_accuracy: 0.2525\n",
      "Epoch 552/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 8.3556 - categorical_accuracy: 0.3561\n",
      "Epoch 553/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 5.1462 - categorical_accuracy: 0.3562\n",
      "Epoch 554/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 5.0568 - categorical_accuracy: 0.2837\n",
      "Epoch 555/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 9.0240 - categorical_accuracy: 0.2916\n",
      "Epoch 556/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 3.2452 - categorical_accuracy: 0.3503\n",
      "Epoch 557/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 2.3063 - categorical_accuracy: 0.3288\n",
      "Epoch 558/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 2.8983 - categorical_accuracy: 0.3052\n",
      "Epoch 559/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 20.1812 - categorical_accuracy: 0.2682\n",
      "Epoch 560/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 2.3346 - categorical_accuracy: 0.3424\n",
      "Epoch 561/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 2.0362 - categorical_accuracy: 0.2994\n",
      "Epoch 562/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 2.9014 - categorical_accuracy: 0.3542\n",
      "Epoch 563/2000\n",
      "3/3 [==============================] - 0s 156ms/step - loss: 2.5227 - categorical_accuracy: 0.2936\n",
      "Epoch 564/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.9947 - categorical_accuracy: 0.3523\n",
      "Epoch 565/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.4579 - categorical_accuracy: 0.3092\n",
      "Epoch 566/2000\n",
      "3/3 [==============================] - 0s 149ms/step - loss: 2.9673 - categorical_accuracy: 0.3777\n",
      "Epoch 567/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 3.2973 - categorical_accuracy: 0.3757\n",
      "Epoch 568/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 2.9147 - categorical_accuracy: 0.2271\n",
      "Epoch 569/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 2.0454 - categorical_accuracy: 0.3699\n",
      "Epoch 570/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.4526 - categorical_accuracy: 0.3053\n",
      "Epoch 571/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.5509 - categorical_accuracy: 0.3170\n",
      "Epoch 572/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1769 - categorical_accuracy: 0.4266\n",
      "Epoch 573/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.3347 - categorical_accuracy: 0.4754\n",
      "Epoch 574/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 140ms/step - loss: 1.3196 - categorical_accuracy: 0.3092\n",
      "Epoch 575/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.4346 - categorical_accuracy: 0.2544\n",
      "Epoch 576/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.4138 - categorical_accuracy: 0.3072\n",
      "Epoch 577/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.3402 - categorical_accuracy: 0.3855\n",
      "Epoch 578/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.6235 - categorical_accuracy: 0.3112\n",
      "Epoch 579/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.3521 - categorical_accuracy: 0.3679\n",
      "Epoch 580/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.4401 - categorical_accuracy: 0.2760\n",
      "Epoch 581/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.2965 - categorical_accuracy: 0.3562\n",
      "Epoch 582/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.5169 - categorical_accuracy: 0.2779\n",
      "Epoch 583/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.2711 - categorical_accuracy: 0.3933\n",
      "Epoch 584/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.4373 - categorical_accuracy: 0.3287\n",
      "Epoch 585/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.5185 - categorical_accuracy: 0.4168\n",
      "Epoch 586/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.2494 - categorical_accuracy: 0.3796\n",
      "Epoch 587/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.6553 - categorical_accuracy: 0.3249\n",
      "Epoch 588/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.3191 - categorical_accuracy: 0.3717\n",
      "Epoch 589/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1208 - categorical_accuracy: 0.3347\n",
      "Epoch 590/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1888 - categorical_accuracy: 0.3619\n",
      "Epoch 591/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.2320 - categorical_accuracy: 0.2486\n",
      "Epoch 592/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0316 - categorical_accuracy: 0.4266\n",
      "Epoch 593/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.2443 - categorical_accuracy: 0.2720\n",
      "Epoch 594/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.2911 - categorical_accuracy: 0.2583\n",
      "Epoch 595/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.3018 - categorical_accuracy: 0.3698\n",
      "Epoch 596/2000\n",
      "3/3 [==============================] - 0s 139ms/step - loss: 1.2040 - categorical_accuracy: 0.3796\n",
      "Epoch 597/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1445 - categorical_accuracy: 0.3425\n",
      "Epoch 598/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1840 - categorical_accuracy: 0.3973\n",
      "Epoch 599/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.1402 - categorical_accuracy: 0.1840\n",
      "Epoch 600/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1297 - categorical_accuracy: 0.3092\n",
      "Epoch 601/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1554 - categorical_accuracy: 0.3092\n",
      "Epoch 602/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1022 - categorical_accuracy: 0.3640\n",
      "Epoch 603/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0669 - categorical_accuracy: 0.4852\n",
      "Epoch 604/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.1103 - categorical_accuracy: 0.2877\n",
      "Epoch 605/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.2317 - categorical_accuracy: 0.3288\n",
      "Epoch 606/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0693 - categorical_accuracy: 0.3679\n",
      "Epoch 607/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.1865 - categorical_accuracy: 0.2505\n",
      "Epoch 608/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.2973 - categorical_accuracy: 0.3229\n",
      "Epoch 609/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1683 - categorical_accuracy: 0.3229\n",
      "Epoch 610/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1910 - categorical_accuracy: 0.2701\n",
      "Epoch 611/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.2445 - categorical_accuracy: 0.3151\n",
      "Epoch 612/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.1795 - categorical_accuracy: 0.3033\n",
      "Epoch 613/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1919 - categorical_accuracy: 0.3092\n",
      "Epoch 614/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.1668 - categorical_accuracy: 0.3151\n",
      "Epoch 615/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1406 - categorical_accuracy: 0.2916\n",
      "Epoch 616/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1733 - categorical_accuracy: 0.2623\n",
      "Epoch 617/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1182 - categorical_accuracy: 0.3737\n",
      "Epoch 618/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1576 - categorical_accuracy: 0.3425\n",
      "Epoch 619/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0983 - categorical_accuracy: 0.3405\n",
      "Epoch 620/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.1213 - categorical_accuracy: 0.3347\n",
      "Epoch 621/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.1236 - categorical_accuracy: 0.3444\n",
      "Epoch 622/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.1554 - categorical_accuracy: 0.2818\n",
      "Epoch 623/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1069 - categorical_accuracy: 0.3836\n",
      "Epoch 624/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.1766 - categorical_accuracy: 0.3699\n",
      "Epoch 625/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0988 - categorical_accuracy: 0.4129\n",
      "Epoch 626/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1206 - categorical_accuracy: 0.3621\n",
      "Epoch 627/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.1007 - categorical_accuracy: 0.3875\n",
      "Epoch 628/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0821 - categorical_accuracy: 0.4501\n",
      "Epoch 629/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0770 - categorical_accuracy: 0.3913\n",
      "Epoch 630/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1147 - categorical_accuracy: 0.3562\n",
      "Epoch 631/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.1849 - categorical_accuracy: 0.3425\n",
      "Epoch 632/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.1466 - categorical_accuracy: 0.3170\n",
      "Epoch 633/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.1210 - categorical_accuracy: 0.4031\n",
      "Epoch 634/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0624 - categorical_accuracy: 0.2857\n",
      "Epoch 635/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1218 - categorical_accuracy: 0.2955\n",
      "Epoch 636/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 1.1451 - categorical_accuracy: 0.3718\n",
      "Epoch 637/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1415 - categorical_accuracy: 0.3601\n",
      "Epoch 638/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1508 - categorical_accuracy: 0.3229\n",
      "Epoch 639/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1252 - categorical_accuracy: 0.4931\n",
      "Epoch 640/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1228 - categorical_accuracy: 0.3503\n",
      "Epoch 641/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1242 - categorical_accuracy: 0.3953\n",
      "Epoch 642/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1395 - categorical_accuracy: 0.3698\n",
      "Epoch 643/2000\n",
      "3/3 [==============================] - 0s 139ms/step - loss: 1.1794 - categorical_accuracy: 0.4070\n",
      "Epoch 644/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1095 - categorical_accuracy: 0.2916\n",
      "Epoch 645/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.2138 - categorical_accuracy: 0.2917\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 646/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1250 - categorical_accuracy: 0.3033\n",
      "Epoch 647/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.1376 - categorical_accuracy: 0.3171\n",
      "Epoch 648/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1151 - categorical_accuracy: 0.2896\n",
      "Epoch 649/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0343 - categorical_accuracy: 0.3895\n",
      "Epoch 650/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1001 - categorical_accuracy: 0.4206\n",
      "Epoch 651/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0673 - categorical_accuracy: 0.3249\n",
      "Epoch 652/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.1023 - categorical_accuracy: 0.4756\n",
      "Epoch 653/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0960 - categorical_accuracy: 0.3053\n",
      "Epoch 654/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0507 - categorical_accuracy: 0.3523\n",
      "Epoch 655/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0446 - categorical_accuracy: 0.4911\n",
      "Epoch 656/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1259 - categorical_accuracy: 0.2779\n",
      "Epoch 657/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0374 - categorical_accuracy: 0.3719\n",
      "Epoch 658/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0946 - categorical_accuracy: 0.4677\n",
      "Epoch 659/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0662 - categorical_accuracy: 0.2701\n",
      "Epoch 660/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0534 - categorical_accuracy: 0.4501\n",
      "Epoch 661/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.1169 - categorical_accuracy: 0.3072\n",
      "Epoch 662/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0919 - categorical_accuracy: 0.4091\n",
      "Epoch 663/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0778 - categorical_accuracy: 0.4872\n",
      "Epoch 664/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0217 - categorical_accuracy: 0.3404\n",
      "Epoch 665/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.0499 - categorical_accuracy: 0.5303\n",
      "Epoch 666/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0324 - categorical_accuracy: 0.3816\n",
      "Epoch 667/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0902 - categorical_accuracy: 0.3347\n",
      "Epoch 668/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0773 - categorical_accuracy: 0.4089\n",
      "Epoch 669/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0271 - categorical_accuracy: 0.4971\n",
      "Epoch 670/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0383 - categorical_accuracy: 0.4932\n",
      "Epoch 671/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0253 - categorical_accuracy: 0.3209\n",
      "Epoch 672/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0421 - categorical_accuracy: 0.3798\n",
      "Epoch 673/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0697 - categorical_accuracy: 0.5147\n",
      "Epoch 674/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0933 - categorical_accuracy: 0.2955\n",
      "Epoch 675/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0508 - categorical_accuracy: 0.3386\n",
      "Epoch 676/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1021 - categorical_accuracy: 0.4696\n",
      "Epoch 677/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.1008 - categorical_accuracy: 0.3033\n",
      "Epoch 678/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0277 - categorical_accuracy: 0.4247\n",
      "Epoch 679/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.0118 - categorical_accuracy: 0.5361\n",
      "Epoch 680/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0849 - categorical_accuracy: 0.3229\n",
      "Epoch 681/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0589 - categorical_accuracy: 0.4540\n",
      "Epoch 682/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0380 - categorical_accuracy: 0.4089\n",
      "Epoch 683/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0616 - categorical_accuracy: 0.3269\n",
      "Epoch 684/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0479 - categorical_accuracy: 0.4951\n",
      "Epoch 685/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0421 - categorical_accuracy: 0.3796\n",
      "Epoch 686/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0330 - categorical_accuracy: 0.3739\n",
      "Epoch 687/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0546 - categorical_accuracy: 0.4285\n",
      "Epoch 688/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0455 - categorical_accuracy: 0.3934\n",
      "Epoch 689/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0743 - categorical_accuracy: 0.3640\n",
      "Epoch 690/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0520 - categorical_accuracy: 0.5323\n",
      "Epoch 691/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0113 - categorical_accuracy: 0.4618\n",
      "Epoch 692/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0458 - categorical_accuracy: 0.2701\n",
      "Epoch 693/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0744 - categorical_accuracy: 0.4247\n",
      "Epoch 694/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0153 - categorical_accuracy: 0.4324\n",
      "Epoch 695/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0384 - categorical_accuracy: 0.3406\n",
      "Epoch 696/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0294 - categorical_accuracy: 0.5087\n",
      "Epoch 697/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0275 - categorical_accuracy: 0.4227\n",
      "Epoch 698/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0301 - categorical_accuracy: 0.5147\n",
      "Epoch 699/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0017 - categorical_accuracy: 0.4696\n",
      "Epoch 700/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0401 - categorical_accuracy: 0.3131\n",
      "Epoch 701/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0568 - categorical_accuracy: 0.4599\n",
      "Epoch 702/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0080 - categorical_accuracy: 0.5087\n",
      "Epoch 703/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 1.0502 - categorical_accuracy: 0.2995\n",
      "Epoch 704/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0438 - categorical_accuracy: 0.4658\n",
      "Epoch 705/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0173 - categorical_accuracy: 0.5010\n",
      "Epoch 706/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0214 - categorical_accuracy: 0.4814\n",
      "Epoch 707/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0436 - categorical_accuracy: 0.3718\n",
      "Epoch 708/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 1.0258 - categorical_accuracy: 0.4658\n",
      "Epoch 709/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0241 - categorical_accuracy: 0.4520\n",
      "Epoch 710/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0193 - categorical_accuracy: 0.3934\n",
      "Epoch 711/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0497 - categorical_accuracy: 0.4717\n",
      "Epoch 712/2000\n",
      "3/3 [==============================] - 0s 156ms/step - loss: 1.0031 - categorical_accuracy: 0.4618\n",
      "Epoch 713/2000\n",
      "3/3 [==============================] - 0s 160ms/step - loss: 1.0310 - categorical_accuracy: 0.4501\n",
      "Epoch 714/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 1.0127 - categorical_accuracy: 0.4775\n",
      "Epoch 715/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 0.9901 - categorical_accuracy: 0.4676\n",
      "Epoch 716/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.0299 - categorical_accuracy: 0.3993\n",
      "Epoch 717/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0053 - categorical_accuracy: 0.4403\n",
      "Epoch 718/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0271 - categorical_accuracy: 0.4599\n",
      "Epoch 719/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0240 - categorical_accuracy: 0.3835\n",
      "Epoch 720/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0858 - categorical_accuracy: 0.3386\n",
      "Epoch 721/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 1.0019 - categorical_accuracy: 0.4364\n",
      "Epoch 722/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0729 - categorical_accuracy: 0.2662\n",
      "Epoch 723/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 1.0287 - categorical_accuracy: 0.4873\n",
      "Epoch 724/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.0233 - categorical_accuracy: 0.4109\n",
      "Epoch 725/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0202 - categorical_accuracy: 0.3229\n",
      "Epoch 726/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0261 - categorical_accuracy: 0.4873\n",
      "Epoch 727/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.0712 - categorical_accuracy: 0.3287\n",
      "Epoch 728/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0173 - categorical_accuracy: 0.4306\n",
      "Epoch 729/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0852 - categorical_accuracy: 0.4677\n",
      "Epoch 730/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0294 - categorical_accuracy: 0.3425\n",
      "Epoch 731/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0068 - categorical_accuracy: 0.5244\n",
      "Epoch 732/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9969 - categorical_accuracy: 0.4129\n",
      "Epoch 733/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0023 - categorical_accuracy: 0.4716\n",
      "Epoch 734/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0162 - categorical_accuracy: 0.4207\n",
      "Epoch 735/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0113 - categorical_accuracy: 0.4736\n",
      "Epoch 736/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9885 - categorical_accuracy: 0.5537\n",
      "Epoch 737/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0277 - categorical_accuracy: 0.3738\n",
      "Epoch 738/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0033 - categorical_accuracy: 0.4521\n",
      "Epoch 739/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0178 - categorical_accuracy: 0.4580\n",
      "Epoch 740/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9909 - categorical_accuracy: 0.4128\n",
      "Epoch 741/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0047 - categorical_accuracy: 0.4678\n",
      "Epoch 742/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9650 - categorical_accuracy: 0.5264\n",
      "Epoch 743/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0606 - categorical_accuracy: 0.3307\n",
      "Epoch 744/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0305 - categorical_accuracy: 0.5303\n",
      "Epoch 745/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0707 - categorical_accuracy: 0.2896\n",
      "Epoch 746/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9757 - categorical_accuracy: 0.4736\n",
      "Epoch 747/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0953 - categorical_accuracy: 0.4208\n",
      "Epoch 748/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9625 - categorical_accuracy: 0.5146\n",
      "Epoch 749/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.0292 - categorical_accuracy: 0.4834\n",
      "Epoch 750/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 1.0148 - categorical_accuracy: 0.4130\n",
      "Epoch 751/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 0.9858 - categorical_accuracy: 0.4872\n",
      "Epoch 752/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0021 - categorical_accuracy: 0.4306\n",
      "Epoch 753/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9818 - categorical_accuracy: 0.4912\n",
      "Epoch 754/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 1.0161 - categorical_accuracy: 0.4364\n",
      "Epoch 755/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0058 - categorical_accuracy: 0.5577\n",
      "Epoch 756/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9744 - categorical_accuracy: 0.5205\n",
      "Epoch 757/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 1.0259 - categorical_accuracy: 0.4462\n",
      "Epoch 758/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0015 - categorical_accuracy: 0.4109\n",
      "Epoch 759/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9766 - categorical_accuracy: 0.4286\n",
      "Epoch 760/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0188 - categorical_accuracy: 0.4892\n",
      "Epoch 761/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9805 - categorical_accuracy: 0.4462\n",
      "Epoch 762/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9856 - categorical_accuracy: 0.5596\n",
      "Epoch 763/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9727 - categorical_accuracy: 0.4657\n",
      "Epoch 764/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9661 - categorical_accuracy: 0.4618\n",
      "Epoch 765/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9697 - categorical_accuracy: 0.4814\n",
      "Epoch 766/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9811 - categorical_accuracy: 0.4892\n",
      "Epoch 767/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9791 - categorical_accuracy: 0.4677\n",
      "Epoch 768/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9726 - categorical_accuracy: 0.4462\n",
      "Epoch 769/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9813 - categorical_accuracy: 0.4814\n",
      "Epoch 770/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.0040 - categorical_accuracy: 0.3719\n",
      "Epoch 771/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 1.0212 - categorical_accuracy: 0.4443\n",
      "Epoch 772/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9792 - categorical_accuracy: 0.4618\n",
      "Epoch 773/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9904 - categorical_accuracy: 0.3797\n",
      "Epoch 774/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 1.0027 - categorical_accuracy: 0.5147\n",
      "Epoch 775/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9763 - categorical_accuracy: 0.4619\n",
      "Epoch 776/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9847 - categorical_accuracy: 0.4932\n",
      "Epoch 777/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9774 - categorical_accuracy: 0.4403\n",
      "Epoch 778/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.9508 - categorical_accuracy: 0.4735\n",
      "Epoch 779/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9846 - categorical_accuracy: 0.5421\n",
      "Epoch 780/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0201 - categorical_accuracy: 0.3894\n",
      "Epoch 781/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9795 - categorical_accuracy: 0.4716\n",
      "Epoch 782/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9816 - categorical_accuracy: 0.5557\n",
      "Epoch 783/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 1.0742 - categorical_accuracy: 0.3522\n",
      "Epoch 784/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 1.0337 - categorical_accuracy: 0.3875\n",
      "Epoch 785/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9723 - categorical_accuracy: 0.5067\n",
      "Epoch 786/2000\n",
      "3/3 [==============================] - 0s 154ms/step - loss: 0.9670 - categorical_accuracy: 0.3406\n",
      "Epoch 787/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 1.0521 - categorical_accuracy: 0.4618\n",
      "Epoch 788/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9860 - categorical_accuracy: 0.3913\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 789/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0061 - categorical_accuracy: 0.4130\n",
      "Epoch 790/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.9865 - categorical_accuracy: 0.5166\n",
      "Epoch 791/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9921 - categorical_accuracy: 0.4462\n",
      "Epoch 792/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9583 - categorical_accuracy: 0.6046\n",
      "Epoch 793/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9644 - categorical_accuracy: 0.4697\n",
      "Epoch 794/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9490 - categorical_accuracy: 0.4872\n",
      "Epoch 795/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 1.0285 - categorical_accuracy: 0.5087\n",
      "Epoch 796/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0179 - categorical_accuracy: 0.4090\n",
      "Epoch 797/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9909 - categorical_accuracy: 0.3954\n",
      "Epoch 798/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9779 - categorical_accuracy: 0.5733\n",
      "Epoch 799/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9650 - categorical_accuracy: 0.4950\n",
      "Epoch 800/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9539 - categorical_accuracy: 0.4990\n",
      "Epoch 801/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0108 - categorical_accuracy: 0.5792\n",
      "Epoch 802/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0045 - categorical_accuracy: 0.3719\n",
      "Epoch 803/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9554 - categorical_accuracy: 0.4990\n",
      "Epoch 804/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0135 - categorical_accuracy: 0.5284\n",
      "Epoch 805/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9454 - categorical_accuracy: 0.4657\n",
      "Epoch 806/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9457 - categorical_accuracy: 0.4756\n",
      "Epoch 807/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9551 - categorical_accuracy: 0.4834\n",
      "Epoch 808/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9222 - categorical_accuracy: 0.4852\n",
      "Epoch 809/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9637 - categorical_accuracy: 0.5421\n",
      "Epoch 810/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9267 - categorical_accuracy: 0.4931\n",
      "Epoch 811/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9350 - categorical_accuracy: 0.5010\n",
      "Epoch 812/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9204 - categorical_accuracy: 0.4873\n",
      "Epoch 813/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9303 - categorical_accuracy: 0.4599\n",
      "Epoch 814/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9320 - categorical_accuracy: 0.4423\n",
      "Epoch 815/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8895 - categorical_accuracy: 0.5244\n",
      "Epoch 816/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9203 - categorical_accuracy: 0.5323\n",
      "Epoch 817/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9779 - categorical_accuracy: 0.4227\n",
      "Epoch 818/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9171 - categorical_accuracy: 0.5538\n",
      "Epoch 819/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9851 - categorical_accuracy: 0.4443\n",
      "Epoch 820/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9735 - categorical_accuracy: 0.5205\n",
      "Epoch 821/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9811 - categorical_accuracy: 0.4149\n",
      "Epoch 822/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9816 - categorical_accuracy: 0.5479\n",
      "Epoch 823/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9853 - categorical_accuracy: 0.4167\n",
      "Epoch 824/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9744 - categorical_accuracy: 0.5636\n",
      "Epoch 825/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0939 - categorical_accuracy: 0.4520\n",
      "Epoch 826/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0911 - categorical_accuracy: 0.3366\n",
      "Epoch 827/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9859 - categorical_accuracy: 0.5361\n",
      "Epoch 828/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9740 - categorical_accuracy: 0.3347\n",
      "Epoch 829/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 1.0018 - categorical_accuracy: 0.5166\n",
      "Epoch 830/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9333 - categorical_accuracy: 0.5088\n",
      "Epoch 831/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9391 - categorical_accuracy: 0.4716\n",
      "Epoch 832/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9311 - categorical_accuracy: 0.4423\n",
      "Epoch 833/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9400 - categorical_accuracy: 0.5030\n",
      "Epoch 834/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9164 - categorical_accuracy: 0.4931\n",
      "Epoch 835/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8895 - categorical_accuracy: 0.5518\n",
      "Epoch 836/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8950 - categorical_accuracy: 0.4853\n",
      "Epoch 837/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9351 - categorical_accuracy: 0.4599\n",
      "Epoch 838/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9069 - categorical_accuracy: 0.5381\n",
      "Epoch 839/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8833 - categorical_accuracy: 0.5362\n",
      "Epoch 840/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 0.9212 - categorical_accuracy: 0.5460\n",
      "Epoch 841/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9075 - categorical_accuracy: 0.5147\n",
      "Epoch 842/2000\n",
      "3/3 [==============================] - 0s 153ms/step - loss: 0.9073 - categorical_accuracy: 0.5225\n",
      "Epoch 843/2000\n",
      "3/3 [==============================] - 0s 150ms/step - loss: 0.9122 - categorical_accuracy: 0.4912\n",
      "Epoch 844/2000\n",
      "3/3 [==============================] - 0s 158ms/step - loss: 0.9058 - categorical_accuracy: 0.5245\n",
      "Epoch 845/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9127 - categorical_accuracy: 0.4893\n",
      "Epoch 846/2000\n",
      "3/3 [==============================] - 0s 146ms/step - loss: 0.9033 - categorical_accuracy: 0.5264\n",
      "Epoch 847/2000\n",
      "3/3 [==============================] - 0s 151ms/step - loss: 0.9642 - categorical_accuracy: 0.5068\n",
      "Epoch 848/2000\n",
      "3/3 [==============================] - 1s 193ms/step - loss: 0.9610 - categorical_accuracy: 0.5187\n",
      "Epoch 849/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8886 - categorical_accuracy: 0.5146\n",
      "Epoch 850/2000\n",
      "3/3 [==============================] - 1s 174ms/step - loss: 0.9533 - categorical_accuracy: 0.5478\n",
      "Epoch 851/2000\n",
      "3/3 [==============================] - 0s 162ms/step - loss: 0.9679 - categorical_accuracy: 0.5439\n",
      "Epoch 852/2000\n",
      "3/3 [==============================] - 0s 163ms/step - loss: 0.9392 - categorical_accuracy: 0.5264\n",
      "Epoch 853/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 18.9603 - categorical_accuracy: 0.3796\n",
      "Epoch 854/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 2.0652 - categorical_accuracy: 0.3778\n",
      "Epoch 855/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 3.2093 - categorical_accuracy: 0.3288\n",
      "Epoch 856/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 2.1686 - categorical_accuracy: 0.3249\n",
      "Epoch 857/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.9598 - categorical_accuracy: 0.3385\n",
      "Epoch 858/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 2.0022 - categorical_accuracy: 0.3347\n",
      "Epoch 859/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.6765 - categorical_accuracy: 0.3738\n",
      "Epoch 860/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 130ms/step - loss: 1.4409 - categorical_accuracy: 0.3679\n",
      "Epoch 861/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.2626 - categorical_accuracy: 0.3465\n",
      "Epoch 862/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 1.1370 - categorical_accuracy: 0.3562\n",
      "Epoch 863/2000\n",
      "3/3 [==============================] - 1s 224ms/step - loss: 1.1265 - categorical_accuracy: 0.3229\n",
      "Epoch 864/2000\n",
      "3/3 [==============================] - 1s 174ms/step - loss: 1.0350 - categorical_accuracy: 0.4247\n",
      "Epoch 865/2000\n",
      "3/3 [==============================] - 1s 178ms/step - loss: 1.0694 - categorical_accuracy: 0.3307\n",
      "Epoch 866/2000\n",
      "3/3 [==============================] - 1s 169ms/step - loss: 1.0564 - categorical_accuracy: 0.5146\n",
      "Epoch 867/2000\n",
      "3/3 [==============================] - 1s 168ms/step - loss: 1.0607 - categorical_accuracy: 0.4364\n",
      "Epoch 868/2000\n",
      "3/3 [==============================] - 1s 204ms/step - loss: 1.0618 - categorical_accuracy: 0.4286\n",
      "Epoch 869/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0477 - categorical_accuracy: 0.4384\n",
      "Epoch 870/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0350 - categorical_accuracy: 0.4716\n",
      "Epoch 871/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0284 - categorical_accuracy: 0.4755\n",
      "Epoch 872/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0470 - categorical_accuracy: 0.3992\n",
      "Epoch 873/2000\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 1.0490 - categorical_accuracy: 0.2721\n",
      "Epoch 874/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0456 - categorical_accuracy: 0.4286\n",
      "Epoch 875/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0532 - categorical_accuracy: 0.5049\n",
      "Epoch 876/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0350 - categorical_accuracy: 0.3855\n",
      "Epoch 877/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0370 - categorical_accuracy: 0.4384\n",
      "Epoch 878/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0405 - categorical_accuracy: 0.4462\n",
      "Epoch 879/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0333 - categorical_accuracy: 0.4657\n",
      "Epoch 880/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0316 - categorical_accuracy: 0.4462\n",
      "Epoch 881/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0327 - categorical_accuracy: 0.4540\n",
      "Epoch 882/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0197 - categorical_accuracy: 0.4696\n",
      "Epoch 883/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0159 - categorical_accuracy: 0.4579\n",
      "Epoch 884/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0294 - categorical_accuracy: 0.4560\n",
      "Epoch 885/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0156 - categorical_accuracy: 0.4950\n",
      "Epoch 886/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0268 - categorical_accuracy: 0.4364\n",
      "Epoch 887/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0299 - categorical_accuracy: 0.4345\n",
      "Epoch 888/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0215 - categorical_accuracy: 0.5264\n",
      "Epoch 889/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0256 - categorical_accuracy: 0.4912\n",
      "Epoch 890/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0323 - categorical_accuracy: 0.4384\n",
      "Epoch 891/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0170 - categorical_accuracy: 0.4775\n",
      "Epoch 892/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0376 - categorical_accuracy: 0.4716\n",
      "Epoch 893/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0135 - categorical_accuracy: 0.4559\n",
      "Epoch 894/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0351 - categorical_accuracy: 0.4521\n",
      "Epoch 895/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0300 - categorical_accuracy: 0.4990\n",
      "Epoch 896/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0266 - categorical_accuracy: 0.4970\n",
      "Epoch 897/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0022 - categorical_accuracy: 0.5342\n",
      "Epoch 898/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0158 - categorical_accuracy: 0.4462\n",
      "Epoch 899/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0261 - categorical_accuracy: 0.4306\n",
      "Epoch 900/2000\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 1.0093 - categorical_accuracy: 0.4462\n",
      "Epoch 901/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0354 - categorical_accuracy: 0.4109\n",
      "Epoch 902/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 1.0096 - categorical_accuracy: 0.5186\n",
      "Epoch 903/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0434 - categorical_accuracy: 0.4657\n",
      "Epoch 904/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0039 - categorical_accuracy: 0.4735\n",
      "Epoch 905/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0318 - categorical_accuracy: 0.4404\n",
      "Epoch 906/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0290 - categorical_accuracy: 0.4991\n",
      "Epoch 907/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0195 - categorical_accuracy: 0.4834\n",
      "Epoch 908/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 1.0104 - categorical_accuracy: 0.5147\n",
      "Epoch 909/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0166 - categorical_accuracy: 0.4795\n",
      "Epoch 910/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0102 - categorical_accuracy: 0.4384\n",
      "Epoch 911/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.9878 - categorical_accuracy: 0.5028\n",
      "Epoch 912/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0097 - categorical_accuracy: 0.4853\n",
      "Epoch 913/2000\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 1.0148 - categorical_accuracy: 0.4932\n",
      "Epoch 914/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0192 - categorical_accuracy: 0.4795\n",
      "Epoch 915/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0097 - categorical_accuracy: 0.5049\n",
      "Epoch 916/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0395 - categorical_accuracy: 0.4814\n",
      "Epoch 917/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0262 - categorical_accuracy: 0.4364\n",
      "Epoch 918/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0172 - categorical_accuracy: 0.4540\n",
      "Epoch 919/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0250 - categorical_accuracy: 0.4638\n",
      "Epoch 920/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9859 - categorical_accuracy: 0.4735\n",
      "Epoch 921/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9894 - categorical_accuracy: 0.4834\n",
      "Epoch 922/2000\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 0.9967 - categorical_accuracy: 0.4540\n",
      "Epoch 923/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9905 - categorical_accuracy: 0.5166\n",
      "Epoch 924/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9983 - categorical_accuracy: 0.4931\n",
      "Epoch 925/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0013 - categorical_accuracy: 0.4091\n",
      "Epoch 926/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9922 - categorical_accuracy: 0.4619\n",
      "Epoch 927/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9792 - categorical_accuracy: 0.4716\n",
      "Epoch 928/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9920 - categorical_accuracy: 0.4540\n",
      "Epoch 929/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9862 - categorical_accuracy: 0.4384\n",
      "Epoch 930/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9893 - categorical_accuracy: 0.4677\n",
      "Epoch 931/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0015 - categorical_accuracy: 0.4952\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 932/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0051 - categorical_accuracy: 0.4756\n",
      "Epoch 933/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9974 - categorical_accuracy: 0.4304\n",
      "Epoch 934/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9917 - categorical_accuracy: 0.4618\n",
      "Epoch 935/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9896 - categorical_accuracy: 0.5107\n",
      "Epoch 936/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9937 - categorical_accuracy: 0.4031\n",
      "Epoch 937/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9925 - categorical_accuracy: 0.4893\n",
      "Epoch 938/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9887 - categorical_accuracy: 0.5499\n",
      "Epoch 939/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0114 - categorical_accuracy: 0.4384\n",
      "Epoch 940/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9836 - categorical_accuracy: 0.4306\n",
      "Epoch 941/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9620 - categorical_accuracy: 0.5440\n",
      "Epoch 942/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0164 - categorical_accuracy: 0.3972\n",
      "Epoch 943/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0261 - categorical_accuracy: 0.4873\n",
      "Epoch 944/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0108 - categorical_accuracy: 0.4911\n",
      "Epoch 945/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0054 - categorical_accuracy: 0.3641\n",
      "Epoch 946/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9662 - categorical_accuracy: 0.4560\n",
      "Epoch 947/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9706 - categorical_accuracy: 0.4207\n",
      "Epoch 948/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9815 - categorical_accuracy: 0.4990\n",
      "Epoch 949/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9686 - categorical_accuracy: 0.4462\n",
      "Epoch 950/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9804 - categorical_accuracy: 0.4462\n",
      "Epoch 951/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9763 - categorical_accuracy: 0.4267\n",
      "Epoch 952/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9800 - categorical_accuracy: 0.4501\n",
      "Epoch 953/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9473 - categorical_accuracy: 0.4657\n",
      "Epoch 954/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9491 - categorical_accuracy: 0.5342\n",
      "Epoch 955/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9652 - categorical_accuracy: 0.4697\n",
      "Epoch 956/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9469 - categorical_accuracy: 0.5225\n",
      "Epoch 957/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9895 - categorical_accuracy: 0.4971\n",
      "Epoch 958/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9947 - categorical_accuracy: 0.4639\n",
      "Epoch 959/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9543 - categorical_accuracy: 0.4814\n",
      "Epoch 960/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9510 - categorical_accuracy: 0.4815\n",
      "Epoch 961/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9688 - categorical_accuracy: 0.4599\n",
      "Epoch 962/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9583 - categorical_accuracy: 0.5029\n",
      "Epoch 963/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9390 - categorical_accuracy: 0.5146\n",
      "Epoch 964/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9360 - categorical_accuracy: 0.4775\n",
      "Epoch 965/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9836 - categorical_accuracy: 0.4853\n",
      "Epoch 966/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9451 - categorical_accuracy: 0.5733\n",
      "Epoch 967/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9519 - categorical_accuracy: 0.4990\n",
      "Epoch 968/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9410 - categorical_accuracy: 0.4384\n",
      "Epoch 969/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9624 - categorical_accuracy: 0.4853\n",
      "Epoch 970/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9274 - categorical_accuracy: 0.4129\n",
      "Epoch 971/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0381 - categorical_accuracy: 0.2916\n",
      "Epoch 972/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0035 - categorical_accuracy: 0.5245\n",
      "Epoch 973/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9806 - categorical_accuracy: 0.4697\n",
      "Epoch 974/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9947 - categorical_accuracy: 0.4521\n",
      "Epoch 975/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9834 - categorical_accuracy: 0.4834\n",
      "Epoch 976/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9562 - categorical_accuracy: 0.4580\n",
      "Epoch 977/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9247 - categorical_accuracy: 0.5343\n",
      "Epoch 978/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9668 - categorical_accuracy: 0.5147\n",
      "Epoch 979/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9025 - categorical_accuracy: 0.5263\n",
      "Epoch 980/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9496 - categorical_accuracy: 0.5185\n",
      "Epoch 981/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9751 - categorical_accuracy: 0.5029\n",
      "Epoch 982/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9657 - categorical_accuracy: 0.4110\n",
      "Epoch 983/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9483 - categorical_accuracy: 0.4931\n",
      "Epoch 984/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9084 - categorical_accuracy: 0.5107\n",
      "Epoch 985/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9511 - categorical_accuracy: 0.5088\n",
      "Epoch 986/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9181 - categorical_accuracy: 0.5010\n",
      "Epoch 987/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9608 - categorical_accuracy: 0.5323\n",
      "Epoch 988/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9371 - categorical_accuracy: 0.4638\n",
      "Epoch 989/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9168 - categorical_accuracy: 0.4697\n",
      "Epoch 990/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9260 - categorical_accuracy: 0.4580\n",
      "Epoch 991/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9225 - categorical_accuracy: 0.4951\n",
      "Epoch 992/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9343 - categorical_accuracy: 0.4364\n",
      "Epoch 993/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8843 - categorical_accuracy: 0.4736\n",
      "Epoch 994/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0578 - categorical_accuracy: 0.2524\n",
      "Epoch 995/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8989 - categorical_accuracy: 0.5577\n",
      "Epoch 996/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9520 - categorical_accuracy: 0.4403\n",
      "Epoch 997/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9327 - categorical_accuracy: 0.5499\n",
      "Epoch 998/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9332 - categorical_accuracy: 0.4306\n",
      "Epoch 999/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9389 - categorical_accuracy: 0.4716\n",
      "Epoch 1000/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0331 - categorical_accuracy: 0.6006\n",
      "Epoch 1001/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0965 - categorical_accuracy: 0.3933\n",
      "Epoch 1002/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0085 - categorical_accuracy: 0.4207\n",
      "Epoch 1003/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0251 - categorical_accuracy: 0.4540\n",
      "Epoch 1004/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9949 - categorical_accuracy: 0.4265\n",
      "Epoch 1005/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9713 - categorical_accuracy: 0.3994\n",
      "Epoch 1006/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0550 - categorical_accuracy: 0.5499\n",
      "Epoch 1007/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9537 - categorical_accuracy: 0.5186\n",
      "Epoch 1008/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9816 - categorical_accuracy: 0.4657\n",
      "Epoch 1009/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9404 - categorical_accuracy: 0.4951\n",
      "Epoch 1010/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9242 - categorical_accuracy: 0.6027\n",
      "Epoch 1011/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9584 - categorical_accuracy: 0.5323\n",
      "Epoch 1012/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9455 - categorical_accuracy: 0.5088\n",
      "Epoch 1013/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9533 - categorical_accuracy: 0.5500\n",
      "Epoch 1014/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9606 - categorical_accuracy: 0.5205\n",
      "Epoch 1015/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9357 - categorical_accuracy: 0.4619\n",
      "Epoch 1016/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9489 - categorical_accuracy: 0.3915\n",
      "Epoch 1017/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9389 - categorical_accuracy: 0.4345\n",
      "Epoch 1018/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9068 - categorical_accuracy: 0.4423\n",
      "Epoch 1019/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8978 - categorical_accuracy: 0.4696\n",
      "Epoch 1020/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9215 - categorical_accuracy: 0.4423\n",
      "Epoch 1021/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9166 - categorical_accuracy: 0.4462\n",
      "Epoch 1022/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9121 - categorical_accuracy: 0.4775\n",
      "Epoch 1023/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9442 - categorical_accuracy: 0.4716\n",
      "Epoch 1024/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9036 - categorical_accuracy: 0.4560\n",
      "Epoch 1025/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8893 - categorical_accuracy: 0.5636\n",
      "Epoch 1026/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8726 - categorical_accuracy: 0.5714\n",
      "Epoch 1027/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.9065 - categorical_accuracy: 0.5694\n",
      "Epoch 1028/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8840 - categorical_accuracy: 0.4677\n",
      "Epoch 1029/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9119 - categorical_accuracy: 0.4697\n",
      "Epoch 1030/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9046 - categorical_accuracy: 0.4501\n",
      "Epoch 1031/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.8422 - categorical_accuracy: 0.5675\n",
      "Epoch 1032/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9945 - categorical_accuracy: 0.4012\n",
      "Epoch 1033/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9469 - categorical_accuracy: 0.5440\n",
      "Epoch 1034/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0654 - categorical_accuracy: 0.4130\n",
      "Epoch 1035/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9508 - categorical_accuracy: 0.4324\n",
      "Epoch 1036/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9024 - categorical_accuracy: 0.6496\n",
      "Epoch 1037/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9509 - categorical_accuracy: 0.5050\n",
      "Epoch 1038/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9274 - categorical_accuracy: 0.5460\n",
      "Epoch 1039/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9134 - categorical_accuracy: 0.4756\n",
      "Epoch 1040/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9322 - categorical_accuracy: 0.4638\n",
      "Epoch 1041/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.9539 - categorical_accuracy: 0.4814\n",
      "Epoch 1042/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9092 - categorical_accuracy: 0.5206\n",
      "Epoch 1043/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9351 - categorical_accuracy: 0.5186\n",
      "Epoch 1044/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.9189 - categorical_accuracy: 0.5323\n",
      "Epoch 1045/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8725 - categorical_accuracy: 0.5088\n",
      "Epoch 1046/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9420 - categorical_accuracy: 0.4423\n",
      "Epoch 1047/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8767 - categorical_accuracy: 0.5205\n",
      "Epoch 1048/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8862 - categorical_accuracy: 0.6321\n",
      "Epoch 1049/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.8930 - categorical_accuracy: 0.4149\n",
      "Epoch 1050/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9448 - categorical_accuracy: 0.5264\n",
      "Epoch 1051/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8687 - categorical_accuracy: 0.5342\n",
      "Epoch 1052/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9130 - categorical_accuracy: 0.5676\n",
      "Epoch 1053/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8356 - categorical_accuracy: 0.5576\n",
      "Epoch 1054/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8817 - categorical_accuracy: 0.5146\n",
      "Epoch 1055/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9096 - categorical_accuracy: 0.5636\n",
      "Epoch 1056/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9070 - categorical_accuracy: 0.5792\n",
      "Epoch 1057/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9677 - categorical_accuracy: 0.3797\n",
      "Epoch 1058/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9352 - categorical_accuracy: 0.5244\n",
      "Epoch 1059/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0657 - categorical_accuracy: 0.3484\n",
      "Epoch 1060/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9205 - categorical_accuracy: 0.5128\n",
      "Epoch 1061/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0115 - categorical_accuracy: 0.5187\n",
      "Epoch 1062/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9128 - categorical_accuracy: 0.5186\n",
      "Epoch 1063/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8910 - categorical_accuracy: 0.5088\n",
      "Epoch 1064/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8813 - categorical_accuracy: 0.5929\n",
      "Epoch 1065/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.8807 - categorical_accuracy: 0.5029\n",
      "Epoch 1066/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8751 - categorical_accuracy: 0.5049\n",
      "Epoch 1067/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8746 - categorical_accuracy: 0.5851\n",
      "Epoch 1068/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8709 - categorical_accuracy: 0.4951\n",
      "Epoch 1069/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8677 - categorical_accuracy: 0.5421\n",
      "Epoch 1070/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8639 - categorical_accuracy: 0.4873\n",
      "Epoch 1071/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8416 - categorical_accuracy: 0.6477\n",
      "Epoch 1072/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8507 - categorical_accuracy: 0.6476\n",
      "Epoch 1073/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8283 - categorical_accuracy: 0.5811\n",
      "Epoch 1074/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8662 - categorical_accuracy: 0.5205\n",
      "Epoch 1075/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8412 - categorical_accuracy: 0.4951\n",
      "Epoch 1076/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.8147 - categorical_accuracy: 0.6066\n",
      "Epoch 1077/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9799 - categorical_accuracy: 0.4932\n",
      "Epoch 1078/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9906 - categorical_accuracy: 0.4756\n",
      "Epoch 1079/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8812 - categorical_accuracy: 0.5851\n",
      "Epoch 1080/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.9177 - categorical_accuracy: 0.4736\n",
      "Epoch 1081/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9627 - categorical_accuracy: 0.4991\n",
      "Epoch 1082/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8342 - categorical_accuracy: 0.5420\n",
      "Epoch 1083/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.8055 - categorical_accuracy: 0.5460\n",
      "Epoch 1084/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8323 - categorical_accuracy: 0.5499\n",
      "Epoch 1085/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8292 - categorical_accuracy: 0.5284\n",
      "Epoch 1086/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8805 - categorical_accuracy: 0.5753\n",
      "Epoch 1087/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8219 - categorical_accuracy: 0.5244\n",
      "Epoch 1088/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8019 - categorical_accuracy: 0.5734\n",
      "Epoch 1089/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.8169 - categorical_accuracy: 0.6829\n",
      "Epoch 1090/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9703 - categorical_accuracy: 0.5186\n",
      "Epoch 1091/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.9558 - categorical_accuracy: 0.5578\n",
      "Epoch 1092/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9671 - categorical_accuracy: 0.5225\n",
      "Epoch 1093/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.7943 - categorical_accuracy: 0.6085\n",
      "Epoch 1094/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9118 - categorical_accuracy: 0.4990\n",
      "Epoch 1095/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8134 - categorical_accuracy: 0.5870\n",
      "Epoch 1096/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.7858 - categorical_accuracy: 0.5890\n",
      "Epoch 1097/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8310 - categorical_accuracy: 0.5871\n",
      "Epoch 1098/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.9212 - categorical_accuracy: 0.5813\n",
      "Epoch 1099/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8484 - categorical_accuracy: 0.5597\n",
      "Epoch 1100/2000\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 0.8428 - categorical_accuracy: 0.5578\n",
      "Epoch 1101/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9143 - categorical_accuracy: 0.5304\n",
      "Epoch 1102/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.7844 - categorical_accuracy: 0.6105\n",
      "Epoch 1103/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.8586 - categorical_accuracy: 0.5597\n",
      "Epoch 1104/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.7809 - categorical_accuracy: 0.6320\n",
      "Epoch 1105/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7916 - categorical_accuracy: 0.5949\n",
      "Epoch 1106/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8272 - categorical_accuracy: 0.5871\n",
      "Epoch 1107/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8162 - categorical_accuracy: 0.5636\n",
      "Epoch 1108/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.7644 - categorical_accuracy: 0.6223\n",
      "Epoch 1109/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8082 - categorical_accuracy: 0.5577\n",
      "Epoch 1110/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.7730 - categorical_accuracy: 0.6086\n",
      "Epoch 1111/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.7960 - categorical_accuracy: 0.6028\n",
      "Epoch 1112/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8564 - categorical_accuracy: 0.5949\n",
      "Epoch 1113/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7847 - categorical_accuracy: 0.6517\n",
      "Epoch 1114/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7630 - categorical_accuracy: 0.6027\n",
      "Epoch 1115/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8669 - categorical_accuracy: 0.4560\n",
      "Epoch 1116/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8617 - categorical_accuracy: 0.4892\n",
      "Epoch 1117/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.7421 - categorical_accuracy: 0.6204\n",
      "Epoch 1118/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.7370 - categorical_accuracy: 0.5597\n",
      "Epoch 1119/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8677 - categorical_accuracy: 0.6067\n",
      "Epoch 1120/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7807 - categorical_accuracy: 0.6144\n",
      "Epoch 1121/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9258 - categorical_accuracy: 0.4423\n",
      "Epoch 1122/2000\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 1.0373 - categorical_accuracy: 0.4344\n",
      "Epoch 1123/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.4931 - categorical_accuracy: 0.3660\n",
      "Epoch 1124/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.1919 - categorical_accuracy: 0.3543\n",
      "Epoch 1125/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.8974 - categorical_accuracy: 0.5714\n",
      "Epoch 1126/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8851 - categorical_accuracy: 0.5264\n",
      "Epoch 1127/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9100 - categorical_accuracy: 0.4795\n",
      "Epoch 1128/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8495 - categorical_accuracy: 0.5576\n",
      "Epoch 1129/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9183 - categorical_accuracy: 0.4813\n",
      "Epoch 1130/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 0.8691 - categorical_accuracy: 0.5871\n",
      "Epoch 1131/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8730 - categorical_accuracy: 0.5558\n",
      "Epoch 1132/2000\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 0.8861 - categorical_accuracy: 0.5851\n",
      "Epoch 1133/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8565 - categorical_accuracy: 0.5402\n",
      "Epoch 1134/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8792 - categorical_accuracy: 0.5460\n",
      "Epoch 1135/2000\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 0.8233 - categorical_accuracy: 0.5754\n",
      "Epoch 1136/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8452 - categorical_accuracy: 0.5421\n",
      "Epoch 1137/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8288 - categorical_accuracy: 0.5852\n",
      "Epoch 1138/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.8356 - categorical_accuracy: 0.5851\n",
      "Epoch 1139/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8405 - categorical_accuracy: 0.5734\n",
      "Epoch 1140/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8413 - categorical_accuracy: 0.5852\n",
      "Epoch 1141/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8523 - categorical_accuracy: 0.5245\n",
      "Epoch 1142/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8038 - categorical_accuracy: 0.5558\n",
      "Epoch 1143/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.7889 - categorical_accuracy: 0.6281\n",
      "Epoch 1144/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8655 - categorical_accuracy: 0.5636\n",
      "Epoch 1145/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8193 - categorical_accuracy: 0.5929\n",
      "Epoch 1146/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8197 - categorical_accuracy: 0.6027\n",
      "Epoch 1147/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.7956 - categorical_accuracy: 0.6634\n",
      "Epoch 1148/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8154 - categorical_accuracy: 0.5655\n",
      "Epoch 1149/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.7874 - categorical_accuracy: 0.6497\n",
      "Epoch 1150/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7786 - categorical_accuracy: 0.5813\n",
      "Epoch 1151/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.7996 - categorical_accuracy: 0.6047\n",
      "Epoch 1152/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.7555 - categorical_accuracy: 0.6536\n",
      "Epoch 1153/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7804 - categorical_accuracy: 0.6027\n",
      "Epoch 1154/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.7597 - categorical_accuracy: 0.6243\n",
      "Epoch 1155/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.7619 - categorical_accuracy: 0.5988\n",
      "Epoch 1156/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.2785 - categorical_accuracy: 0.4540\n",
      "Epoch 1157/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8437 - categorical_accuracy: 0.5988\n",
      "Epoch 1158/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.9336 - categorical_accuracy: 0.4482\n",
      "Epoch 1159/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9324 - categorical_accuracy: 0.5714\n",
      "Epoch 1160/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9710 - categorical_accuracy: 0.5363\n",
      "Epoch 1161/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8675 - categorical_accuracy: 0.6086\n",
      "Epoch 1162/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.8536 - categorical_accuracy: 0.5381\n",
      "Epoch 1163/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8369 - categorical_accuracy: 0.5401\n",
      "Epoch 1164/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8655 - categorical_accuracy: 0.5264\n",
      "Epoch 1165/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8654 - categorical_accuracy: 0.4540\n",
      "Epoch 1166/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8649 - categorical_accuracy: 0.5010\n",
      "Epoch 1167/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8593 - categorical_accuracy: 0.5499\n",
      "Epoch 1168/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8379 - categorical_accuracy: 0.5479\n",
      "Epoch 1169/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 0.8278 - categorical_accuracy: 0.5869\n",
      "Epoch 1170/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8414 - categorical_accuracy: 0.5402\n",
      "Epoch 1171/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8801 - categorical_accuracy: 0.6106\n",
      "Epoch 1172/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8108 - categorical_accuracy: 0.5909\n",
      "Epoch 1173/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.8516 - categorical_accuracy: 0.5440\n",
      "Epoch 1174/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8112 - categorical_accuracy: 0.5694\n",
      "Epoch 1175/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7839 - categorical_accuracy: 0.6672\n",
      "Epoch 1176/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.7716 - categorical_accuracy: 0.6360\n",
      "Epoch 1177/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8673 - categorical_accuracy: 0.5832\n",
      "Epoch 1178/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7797 - categorical_accuracy: 0.5851\n",
      "Epoch 1179/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.8017 - categorical_accuracy: 0.6106\n",
      "Epoch 1180/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8192 - categorical_accuracy: 0.5655\n",
      "Epoch 1181/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8242 - categorical_accuracy: 0.6477\n",
      "Epoch 1182/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8182 - categorical_accuracy: 0.5637\n",
      "Epoch 1183/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.7719 - categorical_accuracy: 0.5949\n",
      "Epoch 1184/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.8913 - categorical_accuracy: 0.6243\n",
      "Epoch 1185/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0670 - categorical_accuracy: 0.5108\n",
      "Epoch 1186/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8595 - categorical_accuracy: 0.5674\n",
      "Epoch 1187/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7672 - categorical_accuracy: 0.6380\n",
      "Epoch 1188/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.8286 - categorical_accuracy: 0.5441\n",
      "Epoch 1189/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.7936 - categorical_accuracy: 0.6418\n",
      "Epoch 1190/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.9059 - categorical_accuracy: 0.4599\n",
      "Epoch 1191/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.9008 - categorical_accuracy: 0.5362\n",
      "Epoch 1192/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7709 - categorical_accuracy: 0.6184\n",
      "Epoch 1193/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8434 - categorical_accuracy: 0.4912\n",
      "Epoch 1194/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8281 - categorical_accuracy: 0.5519\n",
      "Epoch 1195/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.7973 - categorical_accuracy: 0.6360\n",
      "Epoch 1196/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7912 - categorical_accuracy: 0.5480\n",
      "Epoch 1197/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.8011 - categorical_accuracy: 0.6731\n",
      "Epoch 1198/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.8890 - categorical_accuracy: 0.5616\n",
      "Epoch 1199/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.8380 - categorical_accuracy: 0.5753\n",
      "Epoch 1200/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.8170 - categorical_accuracy: 0.5851\n",
      "Epoch 1201/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.7797 - categorical_accuracy: 0.6478\n",
      "Epoch 1202/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 0.7852 - categorical_accuracy: 0.6145\n",
      "Epoch 1203/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.7549 - categorical_accuracy: 0.6869\n",
      "Epoch 1204/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.7828 - categorical_accuracy: 0.5519\n",
      "Epoch 1205/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.7928 - categorical_accuracy: 0.6144\n",
      "Epoch 1206/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 0.7803 - categorical_accuracy: 0.6066\n",
      "Epoch 1207/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 0.8327 - categorical_accuracy: 0.6183\n",
      "Epoch 1208/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 2.8605 - categorical_accuracy: 0.4324\n",
      "Epoch 1209/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 1.9866 - categorical_accuracy: 0.3425\n",
      "Epoch 1210/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.6647 - categorical_accuracy: 0.3796\n",
      "Epoch 1211/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.5485 - categorical_accuracy: 0.3327\n",
      "Epoch 1212/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.2980 - categorical_accuracy: 0.3483\n",
      "Epoch 1213/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.1557 - categorical_accuracy: 0.3483\n",
      "Epoch 1214/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0752 - categorical_accuracy: 0.3796\n",
      "Epoch 1215/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0743 - categorical_accuracy: 0.3856\n",
      "Epoch 1216/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0772 - categorical_accuracy: 0.4892\n",
      "Epoch 1217/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0816 - categorical_accuracy: 0.4756\n",
      "Epoch 1218/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.1927 - categorical_accuracy: 0.3248\n",
      "Epoch 1219/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0961 - categorical_accuracy: 0.3601\n",
      "Epoch 1220/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.1691 - categorical_accuracy: 0.3132\n",
      "Epoch 1221/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.1328 - categorical_accuracy: 0.3523\n",
      "Epoch 1222/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 1.1220 - categorical_accuracy: 0.3640\n",
      "Epoch 1223/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.1033 - categorical_accuracy: 0.3366\n",
      "Epoch 1224/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0775 - categorical_accuracy: 0.3562\n",
      "Epoch 1225/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0688 - categorical_accuracy: 0.3288\n",
      "Epoch 1226/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0540 - categorical_accuracy: 0.3679\n",
      "Epoch 1227/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 4.6065 - categorical_accuracy: 0.3640\n",
      "Epoch 1228/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0523 - categorical_accuracy: 0.3288\n",
      "Epoch 1229/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0658 - categorical_accuracy: 0.5558\n",
      "Epoch 1230/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0868 - categorical_accuracy: 0.4286\n",
      "Epoch 1231/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 1.0641 - categorical_accuracy: 0.4911\n",
      "Epoch 1232/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0749 - categorical_accuracy: 0.4403\n",
      "Epoch 1233/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0526 - categorical_accuracy: 0.5088\n",
      "Epoch 1234/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0457 - categorical_accuracy: 0.3444\n",
      "Epoch 1235/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0510 - categorical_accuracy: 0.3444\n",
      "Epoch 1236/2000\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 1.0407 - categorical_accuracy: 0.3835\n",
      "Epoch 1237/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0284 - categorical_accuracy: 0.3835\n",
      "Epoch 1238/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0514 - categorical_accuracy: 0.3562\n",
      "Epoch 1239/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 1.0482 - categorical_accuracy: 0.3327\n",
      "Epoch 1240/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 1.0477 - categorical_accuracy: 0.3523\n",
      "Epoch 1241/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0360 - categorical_accuracy: 0.3972\n",
      "Epoch 1242/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0235 - categorical_accuracy: 0.4364\n",
      "Epoch 1243/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0366 - categorical_accuracy: 0.5577\n",
      "Epoch 1244/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0380 - categorical_accuracy: 0.5226\n",
      "Epoch 1245/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0304 - categorical_accuracy: 0.6085\n",
      "Epoch 1246/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0330 - categorical_accuracy: 0.5988\n",
      "Epoch 1247/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0265 - categorical_accuracy: 0.5656\n",
      "Epoch 1248/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0343 - categorical_accuracy: 0.5265\n",
      "Epoch 1249/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0289 - categorical_accuracy: 0.5538\n",
      "Epoch 1250/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0197 - categorical_accuracy: 0.5773\n",
      "Epoch 1251/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0232 - categorical_accuracy: 0.5694\n",
      "Epoch 1252/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0150 - categorical_accuracy: 0.5851\n",
      "Epoch 1253/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0224 - categorical_accuracy: 0.5636\n",
      "Epoch 1254/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0260 - categorical_accuracy: 0.6144\n",
      "Epoch 1255/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0344 - categorical_accuracy: 0.5597\n",
      "Epoch 1256/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0262 - categorical_accuracy: 0.5929\n",
      "Epoch 1257/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0302 - categorical_accuracy: 0.5617\n",
      "Epoch 1258/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0174 - categorical_accuracy: 0.5890\n",
      "Epoch 1259/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0265 - categorical_accuracy: 0.5245\n",
      "Epoch 1260/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0225 - categorical_accuracy: 0.5597\n",
      "Epoch 1261/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0270 - categorical_accuracy: 0.5343\n",
      "Epoch 1262/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0309 - categorical_accuracy: 0.5265\n",
      "Epoch 1263/2000\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 1.0247 - categorical_accuracy: 0.5362\n",
      "Epoch 1264/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0208 - categorical_accuracy: 0.5499\n",
      "Epoch 1265/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0191 - categorical_accuracy: 0.5792\n",
      "Epoch 1266/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0138 - categorical_accuracy: 0.5656\n",
      "Epoch 1267/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0115 - categorical_accuracy: 0.6242\n",
      "Epoch 1268/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0113 - categorical_accuracy: 0.5831\n",
      "Epoch 1269/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 1.0175 - categorical_accuracy: 0.5323\n",
      "Epoch 1270/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0261 - categorical_accuracy: 0.4522\n",
      "Epoch 1271/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0154 - categorical_accuracy: 0.5362\n",
      "Epoch 1272/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 39.5431 - categorical_accuracy: 0.4286\n",
      "Epoch 1273/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 15.6159 - categorical_accuracy: 0.3582\n",
      "Epoch 1274/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.1460 - categorical_accuracy: 0.3288\n",
      "Epoch 1275/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 1.2446 - categorical_accuracy: 0.3366\n",
      "Epoch 1276/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.2223 - categorical_accuracy: 0.3757\n",
      "Epoch 1277/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.2419 - categorical_accuracy: 0.3562\n",
      "Epoch 1278/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.1276 - categorical_accuracy: 0.3483\n",
      "Epoch 1279/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0549 - categorical_accuracy: 0.3523\n",
      "Epoch 1280/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0171 - categorical_accuracy: 0.4285\n",
      "Epoch 1281/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0439 - categorical_accuracy: 0.5675\n",
      "Epoch 1282/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0445 - categorical_accuracy: 0.4951\n",
      "Epoch 1283/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0492 - categorical_accuracy: 0.5127\n",
      "Epoch 1284/2000\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 1.0415 - categorical_accuracy: 0.5596\n",
      "Epoch 1285/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0454 - categorical_accuracy: 0.4951\n",
      "Epoch 1286/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0338 - categorical_accuracy: 0.5949\n",
      "Epoch 1287/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 119ms/step - loss: 1.0307 - categorical_accuracy: 0.5401\n",
      "Epoch 1288/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 1.0391 - categorical_accuracy: 0.3797\n",
      "Epoch 1289/2000\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 1.0313 - categorical_accuracy: 0.3464\n",
      "Epoch 1290/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0342 - categorical_accuracy: 0.3327\n",
      "Epoch 1291/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 1.0243 - categorical_accuracy: 0.3366\n",
      "Epoch 1292/2000\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 1.0098 - categorical_accuracy: 0.3718\n",
      "Epoch 1293/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 1.0149 - categorical_accuracy: 0.3718\n",
      "Epoch 1294/2000\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 1.0190 - categorical_accuracy: 0.3738\n",
      "Epoch 1295/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9953 - categorical_accuracy: 0.5166\n",
      "Epoch 1296/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0096 - categorical_accuracy: 0.5832\n",
      "Epoch 1297/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9950 - categorical_accuracy: 0.5988\n",
      "Epoch 1298/2000\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 0.9971 - categorical_accuracy: 0.5480\n",
      "Epoch 1299/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 1.0007 - categorical_accuracy: 0.5909\n",
      "Epoch 1300/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9886 - categorical_accuracy: 0.6183\n",
      "Epoch 1301/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 1.0031 - categorical_accuracy: 0.5402\n",
      "Epoch 1302/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9850 - categorical_accuracy: 0.5968\n",
      "Epoch 1303/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9913 - categorical_accuracy: 0.5851\n",
      "Epoch 1304/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9799 - categorical_accuracy: 0.5851\n",
      "Epoch 1305/2000\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 0.9980 - categorical_accuracy: 0.5636\n",
      "Epoch 1306/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.9849 - categorical_accuracy: 0.5616\n",
      "Epoch 1307/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 1.0035 - categorical_accuracy: 0.5166\n",
      "Epoch 1308/2000\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 0.9986 - categorical_accuracy: 0.5068\n",
      "Epoch 1309/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 0.9978 - categorical_accuracy: 0.5069\n",
      "Epoch 1310/2000\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 1.0014 - categorical_accuracy: 0.5069\n",
      "Epoch 1311/2000\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 0.9845 - categorical_accuracy: 0.5616\n",
      "Epoch 1312/2000\n",
      "3/3 [==============================] - 1s 187ms/step - loss: 0.9743 - categorical_accuracy: 0.6046\n",
      "Epoch 1313/2000\n",
      "3/3 [==============================] - 0s 154ms/step - loss: 0.9824 - categorical_accuracy: 0.5616\n",
      "Epoch 1314/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9661 - categorical_accuracy: 0.5714\n",
      "Epoch 1315/2000\n",
      "3/3 [==============================] - 1s 189ms/step - loss: 0.9737 - categorical_accuracy: 0.5381\n",
      "Epoch 1316/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9768 - categorical_accuracy: 0.5421\n",
      "Epoch 1317/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9731 - categorical_accuracy: 0.5304\n",
      "Epoch 1318/2000\n",
      "3/3 [==============================] - 0s 154ms/step - loss: 0.9880 - categorical_accuracy: 0.5186\n",
      "Epoch 1319/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9926 - categorical_accuracy: 0.4756\n",
      "Epoch 1320/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9856 - categorical_accuracy: 0.5108\n",
      "Epoch 1321/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9582 - categorical_accuracy: 0.5733\n",
      "Epoch 1322/2000\n",
      "3/3 [==============================] - 0s 160ms/step - loss: 0.9766 - categorical_accuracy: 0.5421\n",
      "Epoch 1323/2000\n",
      "3/3 [==============================] - 0s 151ms/step - loss: 0.9843 - categorical_accuracy: 0.5499\n",
      "Epoch 1324/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9800 - categorical_accuracy: 0.5148\n",
      "Epoch 1325/2000\n",
      "3/3 [==============================] - 0s 156ms/step - loss: 0.9806 - categorical_accuracy: 0.5225\n",
      "Epoch 1326/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 0.9710 - categorical_accuracy: 0.4951\n",
      "Epoch 1327/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9636 - categorical_accuracy: 0.4951\n",
      "Epoch 1328/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9720 - categorical_accuracy: 0.5127\n",
      "Epoch 1329/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9793 - categorical_accuracy: 0.5382\n",
      "Epoch 1330/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9571 - categorical_accuracy: 0.5909\n",
      "Epoch 1331/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9633 - categorical_accuracy: 0.5675\n",
      "Epoch 1332/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9707 - categorical_accuracy: 0.5870\n",
      "Epoch 1333/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9795 - categorical_accuracy: 0.4834\n",
      "Epoch 1334/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9571 - categorical_accuracy: 0.5714\n",
      "Epoch 1335/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9646 - categorical_accuracy: 0.5636\n",
      "Epoch 1336/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9770 - categorical_accuracy: 0.5089\n",
      "Epoch 1337/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9677 - categorical_accuracy: 0.5029\n",
      "Epoch 1338/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9589 - categorical_accuracy: 0.4990\n",
      "Epoch 1339/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9725 - categorical_accuracy: 0.5127\n",
      "Epoch 1340/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9763 - categorical_accuracy: 0.5479\n",
      "Epoch 1341/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9654 - categorical_accuracy: 0.5342\n",
      "Epoch 1342/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 0.9547 - categorical_accuracy: 0.5401\n",
      "Epoch 1343/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9415 - categorical_accuracy: 0.5557\n",
      "Epoch 1344/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9592 - categorical_accuracy: 0.5362\n",
      "Epoch 1345/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9516 - categorical_accuracy: 0.5362\n",
      "Epoch 1346/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9695 - categorical_accuracy: 0.5186\n",
      "Epoch 1347/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9368 - categorical_accuracy: 0.5655\n",
      "Epoch 1348/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9499 - categorical_accuracy: 0.5440\n",
      "Epoch 1349/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9408 - categorical_accuracy: 0.5675\n",
      "Epoch 1350/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9676 - categorical_accuracy: 0.4737\n",
      "Epoch 1351/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9427 - categorical_accuracy: 0.6144\n",
      "Epoch 1352/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9501 - categorical_accuracy: 0.5890\n",
      "Epoch 1353/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9560 - categorical_accuracy: 0.5363\n",
      "Epoch 1354/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9412 - categorical_accuracy: 0.5636\n",
      "Epoch 1355/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9502 - categorical_accuracy: 0.5069\n",
      "Epoch 1356/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9735 - categorical_accuracy: 0.5206\n",
      "Epoch 1357/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9411 - categorical_accuracy: 0.5930\n",
      "Epoch 1358/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9357 - categorical_accuracy: 0.6046\n",
      "Epoch 1359/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9504 - categorical_accuracy: 0.5890\n",
      "Epoch 1360/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.9248 - categorical_accuracy: 0.5988\n",
      "Epoch 1361/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9528 - categorical_accuracy: 0.5831\n",
      "Epoch 1362/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9562 - categorical_accuracy: 0.5578\n",
      "Epoch 1363/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9244 - categorical_accuracy: 0.5676\n",
      "Epoch 1364/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9507 - categorical_accuracy: 0.5676\n",
      "Epoch 1365/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9450 - categorical_accuracy: 0.5773\n",
      "Epoch 1366/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9418 - categorical_accuracy: 0.5870\n",
      "Epoch 1367/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9221 - categorical_accuracy: 0.5792\n",
      "Epoch 1368/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9612 - categorical_accuracy: 0.5480\n",
      "Epoch 1369/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9349 - categorical_accuracy: 0.5049\n",
      "Epoch 1370/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9122 - categorical_accuracy: 0.5812\n",
      "Epoch 1371/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9390 - categorical_accuracy: 0.6086\n",
      "Epoch 1372/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9085 - categorical_accuracy: 0.5773\n",
      "Epoch 1373/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9498 - categorical_accuracy: 0.5480\n",
      "Epoch 1374/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9106 - categorical_accuracy: 0.6164\n",
      "Epoch 1375/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9467 - categorical_accuracy: 0.5891\n",
      "Epoch 1376/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9103 - categorical_accuracy: 0.6164\n",
      "Epoch 1377/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9551 - categorical_accuracy: 0.6066\n",
      "Epoch 1378/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9293 - categorical_accuracy: 0.6028\n",
      "Epoch 1379/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.9381 - categorical_accuracy: 0.5460\n",
      "Epoch 1380/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9125 - categorical_accuracy: 0.5597\n",
      "Epoch 1381/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9201 - categorical_accuracy: 0.5832\n",
      "Epoch 1382/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9334 - categorical_accuracy: 0.5186\n",
      "Epoch 1383/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9107 - categorical_accuracy: 0.5754\n",
      "Epoch 1384/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9431 - categorical_accuracy: 0.5128\n",
      "Epoch 1385/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9207 - categorical_accuracy: 0.5147\n",
      "Epoch 1386/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.9324 - categorical_accuracy: 0.5656\n",
      "Epoch 1387/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9449 - categorical_accuracy: 0.5460\n",
      "Epoch 1388/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9351 - categorical_accuracy: 0.5676\n",
      "Epoch 1389/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9062 - categorical_accuracy: 0.5949\n",
      "Epoch 1390/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9314 - categorical_accuracy: 0.5636\n",
      "Epoch 1391/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9166 - categorical_accuracy: 0.5832\n",
      "Epoch 1392/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9363 - categorical_accuracy: 0.5186\n",
      "Epoch 1393/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9262 - categorical_accuracy: 0.5519\n",
      "Epoch 1394/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9091 - categorical_accuracy: 0.6105\n",
      "Epoch 1395/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9257 - categorical_accuracy: 0.5264\n",
      "Epoch 1396/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.8976 - categorical_accuracy: 0.6379\n",
      "Epoch 1397/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9318 - categorical_accuracy: 0.5577\n",
      "Epoch 1398/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9050 - categorical_accuracy: 0.5636\n",
      "Epoch 1399/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9124 - categorical_accuracy: 0.5538\n",
      "Epoch 1400/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9099 - categorical_accuracy: 0.5714\n",
      "Epoch 1401/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9337 - categorical_accuracy: 0.5793\n",
      "Epoch 1402/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9088 - categorical_accuracy: 0.5382\n",
      "Epoch 1403/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9107 - categorical_accuracy: 0.6379\n",
      "Epoch 1404/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9251 - categorical_accuracy: 0.5324\n",
      "Epoch 1405/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9201 - categorical_accuracy: 0.5420\n",
      "Epoch 1406/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9181 - categorical_accuracy: 0.5597\n",
      "Epoch 1407/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9255 - categorical_accuracy: 0.5851\n",
      "Epoch 1408/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9221 - categorical_accuracy: 0.5382\n",
      "Epoch 1409/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9363 - categorical_accuracy: 0.6341\n",
      "Epoch 1410/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9151 - categorical_accuracy: 0.5147\n",
      "Epoch 1411/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9038 - categorical_accuracy: 0.5988\n",
      "Epoch 1412/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9060 - categorical_accuracy: 0.5499\n",
      "Epoch 1413/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9078 - categorical_accuracy: 0.5459\n",
      "Epoch 1414/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8686 - categorical_accuracy: 0.6418\n",
      "Epoch 1415/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8951 - categorical_accuracy: 0.6046\n",
      "Epoch 1416/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9358 - categorical_accuracy: 0.5186\n",
      "Epoch 1417/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8986 - categorical_accuracy: 0.6046\n",
      "Epoch 1418/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9414 - categorical_accuracy: 0.5148\n",
      "Epoch 1419/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9394 - categorical_accuracy: 0.5127\n",
      "Epoch 1420/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9617 - categorical_accuracy: 0.4873\n",
      "Epoch 1421/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9049 - categorical_accuracy: 0.5753\n",
      "Epoch 1422/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9649 - categorical_accuracy: 0.4892\n",
      "Epoch 1423/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9290 - categorical_accuracy: 0.5343\n",
      "Epoch 1424/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.9119 - categorical_accuracy: 0.5225\n",
      "Epoch 1425/2000\n",
      "3/3 [==============================] - 0s 139ms/step - loss: 0.9289 - categorical_accuracy: 0.5441\n",
      "Epoch 1426/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9592 - categorical_accuracy: 0.5597\n",
      "Epoch 1427/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.9244 - categorical_accuracy: 0.5578\n",
      "Epoch 1428/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.9007 - categorical_accuracy: 0.5890\n",
      "Epoch 1429/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9026 - categorical_accuracy: 0.6320\n",
      "Epoch 1430/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8866 - categorical_accuracy: 0.5342\n",
      "Epoch 1431/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8825 - categorical_accuracy: 0.5949\n",
      "Epoch 1432/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8953 - categorical_accuracy: 0.5871\n",
      "Epoch 1433/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9137 - categorical_accuracy: 0.5382\n",
      "Epoch 1434/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9085 - categorical_accuracy: 0.5324\n",
      "Epoch 1435/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8954 - categorical_accuracy: 0.6007\n",
      "Epoch 1436/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.8902 - categorical_accuracy: 0.5733\n",
      "Epoch 1437/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8994 - categorical_accuracy: 0.5851\n",
      "Epoch 1438/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8938 - categorical_accuracy: 0.6261\n",
      "Epoch 1439/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9082 - categorical_accuracy: 0.5499\n",
      "Epoch 1440/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9362 - categorical_accuracy: 0.5872\n",
      "Epoch 1441/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.8806 - categorical_accuracy: 0.6516\n",
      "Epoch 1442/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9130 - categorical_accuracy: 0.5030\n",
      "Epoch 1443/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8868 - categorical_accuracy: 0.6066\n",
      "Epoch 1444/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8758 - categorical_accuracy: 0.6829\n",
      "Epoch 1445/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9040 - categorical_accuracy: 0.5108\n",
      "Epoch 1446/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9193 - categorical_accuracy: 0.5304\n",
      "Epoch 1447/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8953 - categorical_accuracy: 0.5342\n",
      "Epoch 1448/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 0.8704 - categorical_accuracy: 0.5910\n",
      "Epoch 1449/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8816 - categorical_accuracy: 0.5851\n",
      "Epoch 1450/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8779 - categorical_accuracy: 0.5812\n",
      "Epoch 1451/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8809 - categorical_accuracy: 0.6066\n",
      "Epoch 1452/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.8912 - categorical_accuracy: 0.5420\n",
      "Epoch 1453/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8837 - categorical_accuracy: 0.5519\n",
      "Epoch 1454/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8865 - categorical_accuracy: 0.5734\n",
      "Epoch 1455/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8824 - categorical_accuracy: 0.5284\n",
      "Epoch 1456/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8676 - categorical_accuracy: 0.6556\n",
      "Epoch 1457/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8589 - categorical_accuracy: 0.5538\n",
      "Epoch 1458/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8822 - categorical_accuracy: 0.6223\n",
      "Epoch 1459/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8841 - categorical_accuracy: 0.6282\n",
      "Epoch 1460/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.9371 - categorical_accuracy: 0.5147\n",
      "Epoch 1461/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9721 - categorical_accuracy: 0.5029\n",
      "Epoch 1462/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9032 - categorical_accuracy: 0.5616\n",
      "Epoch 1463/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9282 - categorical_accuracy: 0.5206\n",
      "Epoch 1464/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9085 - categorical_accuracy: 0.5420\n",
      "Epoch 1465/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9003 - categorical_accuracy: 0.5303\n",
      "Epoch 1466/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9167 - categorical_accuracy: 0.5949\n",
      "Epoch 1467/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.8970 - categorical_accuracy: 0.6106\n",
      "Epoch 1468/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8760 - categorical_accuracy: 0.5186\n",
      "Epoch 1469/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8649 - categorical_accuracy: 0.6066\n",
      "Epoch 1470/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8826 - categorical_accuracy: 0.6144\n",
      "Epoch 1471/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8701 - categorical_accuracy: 0.5245\n",
      "Epoch 1472/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8840 - categorical_accuracy: 0.5558\n",
      "Epoch 1473/2000\n",
      "3/3 [==============================] - 0s 146ms/step - loss: 0.8880 - categorical_accuracy: 0.6184\n",
      "Epoch 1474/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8667 - categorical_accuracy: 0.5460\n",
      "Epoch 1475/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.8860 - categorical_accuracy: 0.5715\n",
      "Epoch 1476/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8691 - categorical_accuracy: 0.5929\n",
      "Epoch 1477/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8654 - categorical_accuracy: 0.5655\n",
      "Epoch 1478/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.8652 - categorical_accuracy: 0.6067\n",
      "Epoch 1479/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8724 - categorical_accuracy: 0.5303\n",
      "Epoch 1480/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8672 - categorical_accuracy: 0.5714\n",
      "Epoch 1481/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9118 - categorical_accuracy: 0.5558\n",
      "Epoch 1482/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8786 - categorical_accuracy: 0.5889\n",
      "Epoch 1483/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.9843 - categorical_accuracy: 0.4208\n",
      "Epoch 1484/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9668 - categorical_accuracy: 0.5675\n",
      "Epoch 1485/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.9451 - categorical_accuracy: 0.5244\n",
      "Epoch 1486/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8911 - categorical_accuracy: 0.5773\n",
      "Epoch 1487/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8846 - categorical_accuracy: 0.5617\n",
      "Epoch 1488/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8909 - categorical_accuracy: 0.5459\n",
      "Epoch 1489/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8938 - categorical_accuracy: 0.5577\n",
      "Epoch 1490/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9075 - categorical_accuracy: 0.5597\n",
      "Epoch 1491/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9291 - categorical_accuracy: 0.5402\n",
      "Epoch 1492/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9310 - categorical_accuracy: 0.4951\n",
      "Epoch 1493/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8700 - categorical_accuracy: 0.5753\n",
      "Epoch 1494/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9220 - categorical_accuracy: 0.5420\n",
      "Epoch 1495/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8867 - categorical_accuracy: 0.5597\n",
      "Epoch 1496/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.9109 - categorical_accuracy: 0.5264\n",
      "Epoch 1497/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8979 - categorical_accuracy: 0.5363\n",
      "Epoch 1498/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9033 - categorical_accuracy: 0.5578\n",
      "Epoch 1499/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8812 - categorical_accuracy: 0.5284\n",
      "Epoch 1500/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9219 - categorical_accuracy: 0.4717\n",
      "Epoch 1501/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8742 - categorical_accuracy: 0.6007\n",
      "Epoch 1502/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8672 - categorical_accuracy: 0.6105\n",
      "Epoch 1503/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8850 - categorical_accuracy: 0.5069\n",
      "Epoch 1504/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8885 - categorical_accuracy: 0.5851\n",
      "Epoch 1505/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.9007 - categorical_accuracy: 0.6125\n",
      "Epoch 1506/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8869 - categorical_accuracy: 0.5030\n",
      "Epoch 1507/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8745 - categorical_accuracy: 0.6028\n",
      "Epoch 1508/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8484 - categorical_accuracy: 0.5421\n",
      "Epoch 1509/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8682 - categorical_accuracy: 0.5147\n",
      "Epoch 1510/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.8592 - categorical_accuracy: 0.5773\n",
      "Epoch 1511/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8366 - categorical_accuracy: 0.5655\n",
      "Epoch 1512/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8580 - categorical_accuracy: 0.6145\n",
      "Epoch 1513/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8528 - categorical_accuracy: 0.6516\n",
      "Epoch 1514/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8477 - categorical_accuracy: 0.5676\n",
      "Epoch 1515/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8452 - categorical_accuracy: 0.6517\n",
      "Epoch 1516/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8532 - categorical_accuracy: 0.5342\n",
      "Epoch 1517/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8364 - categorical_accuracy: 0.6106\n",
      "Epoch 1518/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8451 - categorical_accuracy: 0.5773\n",
      "Epoch 1519/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8624 - categorical_accuracy: 0.5323\n",
      "Epoch 1520/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8408 - categorical_accuracy: 0.6204\n",
      "Epoch 1521/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8558 - categorical_accuracy: 0.5225\n",
      "Epoch 1522/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8621 - categorical_accuracy: 0.5539\n",
      "Epoch 1523/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8392 - categorical_accuracy: 0.5831\n",
      "Epoch 1524/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8485 - categorical_accuracy: 0.5050\n",
      "Epoch 1525/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8430 - categorical_accuracy: 0.6770\n",
      "Epoch 1526/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8364 - categorical_accuracy: 0.5460\n",
      "Epoch 1527/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8317 - categorical_accuracy: 0.5480\n",
      "Epoch 1528/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8521 - categorical_accuracy: 0.6419\n",
      "Epoch 1529/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 31.7702 - categorical_accuracy: 0.6106\n",
      "Epoch 1530/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.9085 - categorical_accuracy: 0.5108\n",
      "Epoch 1531/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8838 - categorical_accuracy: 0.5675\n",
      "Epoch 1532/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9077 - categorical_accuracy: 0.5813\n",
      "Epoch 1533/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8956 - categorical_accuracy: 0.5616\n",
      "Epoch 1534/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9045 - categorical_accuracy: 0.5617\n",
      "Epoch 1535/2000\n",
      "3/3 [==============================] - 0s 146ms/step - loss: 0.9176 - categorical_accuracy: 0.5148\n",
      "Epoch 1536/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8560 - categorical_accuracy: 0.5989\n",
      "Epoch 1537/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8621 - categorical_accuracy: 0.6243\n",
      "Epoch 1538/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8233 - categorical_accuracy: 0.6047\n",
      "Epoch 1539/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.8908 - categorical_accuracy: 0.5127\n",
      "Epoch 1540/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8568 - categorical_accuracy: 0.5754\n",
      "Epoch 1541/2000\n",
      "3/3 [==============================] - 0s 135ms/step - loss: 0.8286 - categorical_accuracy: 0.6261\n",
      "Epoch 1542/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8481 - categorical_accuracy: 0.6047\n",
      "Epoch 1543/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8218 - categorical_accuracy: 0.6496\n",
      "Epoch 1544/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8528 - categorical_accuracy: 0.5792\n",
      "Epoch 1545/2000\n",
      "3/3 [==============================] - 0s 142ms/step - loss: 0.9149 - categorical_accuracy: 0.5011\n",
      "Epoch 1546/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8391 - categorical_accuracy: 0.7084\n",
      "Epoch 1547/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8573 - categorical_accuracy: 0.5441\n",
      "Epoch 1548/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8184 - categorical_accuracy: 0.6243\n",
      "Epoch 1549/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8496 - categorical_accuracy: 0.6653\n",
      "Epoch 1550/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8434 - categorical_accuracy: 0.5813\n",
      "Epoch 1551/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8537 - categorical_accuracy: 0.6204\n",
      "Epoch 1552/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8555 - categorical_accuracy: 0.6067\n",
      "Epoch 1553/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8166 - categorical_accuracy: 0.6164\n",
      "Epoch 1554/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8286 - categorical_accuracy: 0.6048\n",
      "Epoch 1555/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8241 - categorical_accuracy: 0.6497\n",
      "Epoch 1556/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8252 - categorical_accuracy: 0.5734\n",
      "Epoch 1557/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7993 - categorical_accuracy: 0.6750\n",
      "Epoch 1558/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8266 - categorical_accuracy: 0.6497\n",
      "Epoch 1559/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8597 - categorical_accuracy: 0.5636\n",
      "Epoch 1560/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8584 - categorical_accuracy: 0.5597\n",
      "Epoch 1561/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8597 - categorical_accuracy: 0.5754\n",
      "Epoch 1562/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8143 - categorical_accuracy: 0.6125\n",
      "Epoch 1563/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8615 - categorical_accuracy: 0.6146\n",
      "Epoch 1564/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8504 - categorical_accuracy: 0.5323\n",
      "Epoch 1565/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8400 - categorical_accuracy: 0.5910\n",
      "Epoch 1566/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8102 - categorical_accuracy: 0.5949\n",
      "Epoch 1567/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8605 - categorical_accuracy: 0.5226\n",
      "Epoch 1568/2000\n",
      "3/3 [==============================] - 0s 149ms/step - loss: 0.8039 - categorical_accuracy: 0.6477\n",
      "Epoch 1569/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8041 - categorical_accuracy: 0.5675\n",
      "Epoch 1570/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.8132 - categorical_accuracy: 0.6477\n",
      "Epoch 1571/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8188 - categorical_accuracy: 0.5871\n",
      "Epoch 1572/2000\n",
      "3/3 [==============================] - 0s 145ms/step - loss: 0.7975 - categorical_accuracy: 0.6555\n",
      "Epoch 1573/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8160 - categorical_accuracy: 0.6458\n",
      "Epoch 1574/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8145 - categorical_accuracy: 0.5695\n",
      "Epoch 1575/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8243 - categorical_accuracy: 0.6380\n",
      "Epoch 1576/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8259 - categorical_accuracy: 0.5519\n",
      "Epoch 1577/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8135 - categorical_accuracy: 0.6634\n",
      "Epoch 1578/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8105 - categorical_accuracy: 0.6810\n",
      "Epoch 1579/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.8143 - categorical_accuracy: 0.6007\n",
      "Epoch 1580/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7943 - categorical_accuracy: 0.6614\n",
      "Epoch 1581/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8297 - categorical_accuracy: 0.5480\n",
      "Epoch 1582/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7997 - categorical_accuracy: 0.6477\n",
      "Epoch 1583/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7878 - categorical_accuracy: 0.6399\n",
      "Epoch 1584/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7976 - categorical_accuracy: 0.6105\n",
      "Epoch 1585/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8115 - categorical_accuracy: 0.6478\n",
      "Epoch 1586/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7964 - categorical_accuracy: 0.6477\n",
      "Epoch 1587/2000\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 0.8022 - categorical_accuracy: 0.6321\n",
      "Epoch 1588/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7923 - categorical_accuracy: 0.5597\n",
      "Epoch 1589/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8111 - categorical_accuracy: 0.6927\n",
      "Epoch 1590/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.8343 - categorical_accuracy: 0.5851\n",
      "Epoch 1591/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8228 - categorical_accuracy: 0.6360\n",
      "Epoch 1592/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7860 - categorical_accuracy: 0.6575\n",
      "Epoch 1593/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.8077 - categorical_accuracy: 0.6693\n",
      "Epoch 1594/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7751 - categorical_accuracy: 0.6439\n",
      "Epoch 1595/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7976 - categorical_accuracy: 0.6008\n",
      "Epoch 1596/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7985 - categorical_accuracy: 0.6419\n",
      "Epoch 1597/2000\n",
      "3/3 [==============================] - 0s 143ms/step - loss: 0.7898 - categorical_accuracy: 0.5969\n",
      "Epoch 1598/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7894 - categorical_accuracy: 0.6888\n",
      "Epoch 1599/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7731 - categorical_accuracy: 0.6985\n",
      "Epoch 1600/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7861 - categorical_accuracy: 0.6791\n",
      "Epoch 1601/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7917 - categorical_accuracy: 0.6536\n",
      "Epoch 1602/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7764 - categorical_accuracy: 0.7103\n",
      "Epoch 1603/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7989 - categorical_accuracy: 0.5753\n",
      "Epoch 1604/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8136 - categorical_accuracy: 0.5813\n",
      "Epoch 1605/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7960 - categorical_accuracy: 0.5831\n",
      "Epoch 1606/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8277 - categorical_accuracy: 0.6398\n",
      "Epoch 1607/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.8178 - categorical_accuracy: 0.5479\n",
      "Epoch 1608/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.8446 - categorical_accuracy: 0.5636\n",
      "Epoch 1609/2000\n",
      "3/3 [==============================] - 0s 154ms/step - loss: 0.8580 - categorical_accuracy: 0.5362\n",
      "Epoch 1610/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9696 - categorical_accuracy: 0.5245\n",
      "Epoch 1611/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.9006 - categorical_accuracy: 0.5128\n",
      "Epoch 1612/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8250 - categorical_accuracy: 0.5538\n",
      "Epoch 1613/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.9588 - categorical_accuracy: 0.5538\n",
      "Epoch 1614/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7861 - categorical_accuracy: 0.6301\n",
      "Epoch 1615/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7936 - categorical_accuracy: 0.5969\n",
      "Epoch 1616/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8950 - categorical_accuracy: 0.5088\n",
      "Epoch 1617/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7876 - categorical_accuracy: 0.5753\n",
      "Epoch 1618/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7868 - categorical_accuracy: 0.6281\n",
      "Epoch 1619/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.8398 - categorical_accuracy: 0.5461\n",
      "Epoch 1620/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7559 - categorical_accuracy: 0.6634\n",
      "Epoch 1621/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7505 - categorical_accuracy: 0.7260\n",
      "Epoch 1622/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.7740 - categorical_accuracy: 0.6145\n",
      "Epoch 1623/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7502 - categorical_accuracy: 0.6223\n",
      "Epoch 1624/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7532 - categorical_accuracy: 0.6634\n",
      "Epoch 1625/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7635 - categorical_accuracy: 0.6204\n",
      "Epoch 1626/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7759 - categorical_accuracy: 0.6380\n",
      "Epoch 1627/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7443 - categorical_accuracy: 0.7142\n",
      "Epoch 1628/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7724 - categorical_accuracy: 0.6360\n",
      "Epoch 1629/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7425 - categorical_accuracy: 0.6712\n",
      "Epoch 1630/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7627 - categorical_accuracy: 0.6596\n",
      "Epoch 1631/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7461 - categorical_accuracy: 0.6732\n",
      "Epoch 1632/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7558 - categorical_accuracy: 0.6360\n",
      "Epoch 1633/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7595 - categorical_accuracy: 0.6908\n",
      "Epoch 1634/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.7487 - categorical_accuracy: 0.5694\n",
      "Epoch 1635/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7796 - categorical_accuracy: 0.5949\n",
      "Epoch 1636/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8054 - categorical_accuracy: 0.5734\n",
      "Epoch 1637/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7832 - categorical_accuracy: 0.5480\n",
      "Epoch 1638/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7425 - categorical_accuracy: 0.6419\n",
      "Epoch 1639/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7454 - categorical_accuracy: 0.6634\n",
      "Epoch 1640/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7243 - categorical_accuracy: 0.6183\n",
      "Epoch 1641/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7311 - categorical_accuracy: 0.7456\n",
      "Epoch 1642/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 149ms/step - loss: 0.7887 - categorical_accuracy: 0.6536\n",
      "Epoch 1643/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7209 - categorical_accuracy: 0.6399\n",
      "Epoch 1644/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.7232 - categorical_accuracy: 0.6692\n",
      "Epoch 1645/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7357 - categorical_accuracy: 0.6869\n",
      "Epoch 1646/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7315 - categorical_accuracy: 0.6458\n",
      "Epoch 1647/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7423 - categorical_accuracy: 0.6204\n",
      "Epoch 1648/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7556 - categorical_accuracy: 0.6262\n",
      "Epoch 1649/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6971 - categorical_accuracy: 0.6595\n",
      "Epoch 1650/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.7673 - categorical_accuracy: 0.6477\n",
      "Epoch 1651/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7082 - categorical_accuracy: 0.6360\n",
      "Epoch 1652/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7477 - categorical_accuracy: 0.5852\n",
      "Epoch 1653/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7444 - categorical_accuracy: 0.6164\n",
      "Epoch 1654/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7558 - categorical_accuracy: 0.6085\n",
      "Epoch 1655/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8161 - categorical_accuracy: 0.6008\n",
      "Epoch 1656/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7803 - categorical_accuracy: 0.5616\n",
      "Epoch 1657/2000\n",
      "3/3 [==============================] - 0s 132ms/step - loss: 0.7342 - categorical_accuracy: 0.7260\n",
      "Epoch 1658/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.8225 - categorical_accuracy: 0.5694\n",
      "Epoch 1659/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8825 - categorical_accuracy: 0.5010\n",
      "Epoch 1660/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8933 - categorical_accuracy: 0.6144\n",
      "Epoch 1661/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7586 - categorical_accuracy: 0.6380\n",
      "Epoch 1662/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7552 - categorical_accuracy: 0.5852\n",
      "Epoch 1663/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7842 - categorical_accuracy: 0.5949\n",
      "Epoch 1664/2000\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 0.7268 - categorical_accuracy: 0.6379\n",
      "Epoch 1665/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.7430 - categorical_accuracy: 0.6028\n",
      "Epoch 1666/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7500 - categorical_accuracy: 0.5753\n",
      "Epoch 1667/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.7317 - categorical_accuracy: 0.5930\n",
      "Epoch 1668/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.7217 - categorical_accuracy: 0.6693\n",
      "Epoch 1669/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7704 - categorical_accuracy: 0.5480\n",
      "Epoch 1670/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6932 - categorical_accuracy: 0.6575\n",
      "Epoch 1671/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6942 - categorical_accuracy: 0.6947\n",
      "Epoch 1672/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7213 - categorical_accuracy: 0.6302\n",
      "Epoch 1673/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.6724 - categorical_accuracy: 0.7240\n",
      "Epoch 1674/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7096 - categorical_accuracy: 0.6380\n",
      "Epoch 1675/2000\n",
      "3/3 [==============================] - 0s 144ms/step - loss: 0.6726 - categorical_accuracy: 0.6165\n",
      "Epoch 1676/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6993 - categorical_accuracy: 0.6125\n",
      "Epoch 1677/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.6950 - categorical_accuracy: 0.6438\n",
      "Epoch 1678/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6489 - categorical_accuracy: 0.6497\n",
      "Epoch 1679/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6360 - categorical_accuracy: 0.6594\n",
      "Epoch 1680/2000\n",
      "3/3 [==============================] - 0s 148ms/step - loss: 0.6576 - categorical_accuracy: 0.7201\n",
      "Epoch 1681/2000\n",
      "3/3 [==============================] - 1s 184ms/step - loss: 0.6600 - categorical_accuracy: 0.7613\n",
      "Epoch 1682/2000\n",
      "3/3 [==============================] - 0s 150ms/step - loss: 0.6554 - categorical_accuracy: 0.6556\n",
      "Epoch 1683/2000\n",
      "3/3 [==============================] - 0s 152ms/step - loss: 0.6656 - categorical_accuracy: 0.6732\n",
      "Epoch 1684/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.6356 - categorical_accuracy: 0.7026\n",
      "Epoch 1685/2000\n",
      "3/3 [==============================] - 0s 138ms/step - loss: 0.6120 - categorical_accuracy: 0.7259\n",
      "Epoch 1686/2000\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 0.6452 - categorical_accuracy: 0.7221\n",
      "Epoch 1687/2000\n",
      "3/3 [==============================] - 0s 147ms/step - loss: 0.6597 - categorical_accuracy: 0.6947\n",
      "Epoch 1688/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.6302 - categorical_accuracy: 0.6830\n",
      "Epoch 1689/2000\n",
      "3/3 [==============================] - 0s 180ms/step - loss: 0.6530 - categorical_accuracy: 0.6594\n",
      "Epoch 1690/2000\n",
      "3/3 [==============================] - 1s 165ms/step - loss: 0.7503 - categorical_accuracy: 0.6301\n",
      "Epoch 1691/2000\n",
      "3/3 [==============================] - 0s 140ms/step - loss: 0.8206 - categorical_accuracy: 0.6301\n",
      "Epoch 1692/2000\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 0.7500 - categorical_accuracy: 0.6869\n",
      "Epoch 1693/2000\n",
      "3/3 [==============================] - 0s 133ms/step - loss: 0.6947 - categorical_accuracy: 0.6556\n",
      "Epoch 1694/2000\n",
      "3/3 [==============================] - 0s 131ms/step - loss: 0.7808 - categorical_accuracy: 0.4815\n",
      "Epoch 1695/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.7304 - categorical_accuracy: 0.5850\n",
      "Epoch 1696/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.6802 - categorical_accuracy: 0.6908\n",
      "Epoch 1697/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6747 - categorical_accuracy: 0.7239\n",
      "Epoch 1698/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6841 - categorical_accuracy: 0.6771\n",
      "Epoch 1699/2000\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 0.6617 - categorical_accuracy: 0.6575\n",
      "Epoch 1700/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.7065 - categorical_accuracy: 0.5949\n",
      "Epoch 1701/2000\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 0.7137 - categorical_accuracy: 0.5755\n",
      "Epoch 1702/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6468 - categorical_accuracy: 0.6497\n",
      "Epoch 1703/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6596 - categorical_accuracy: 0.7240\n",
      "Epoch 1704/2000\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 0.6661 - categorical_accuracy: 0.6379\n",
      "Epoch 1705/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.6631 - categorical_accuracy: 0.6614\n",
      "Epoch 1706/2000\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 0.6960 - categorical_accuracy: 0.5793\n",
      "Epoch 1707/2000\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 0.6482 - categorical_accuracy: 0.6557\n",
      "Epoch 1708/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.6659 - categorical_accuracy: 0.6537\n",
      "Epoch 1709/2000\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 0.6489 - categorical_accuracy: 0.7182\n",
      "Epoch 1710/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.6179 - categorical_accuracy: 0.6986\n",
      "Epoch 1711/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.6119 - categorical_accuracy: 0.6360\n",
      "Epoch 1712/2000\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 0.6611 - categorical_accuracy: 0.5793\n",
      "Epoch 1713/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 117ms/step - loss: 0.6652 - categorical_accuracy: 0.6067\n",
      "Epoch 1714/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6351 - categorical_accuracy: 0.6438\n",
      "Epoch 1715/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5984 - categorical_accuracy: 0.7201\n",
      "Epoch 1716/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6258 - categorical_accuracy: 0.6830\n",
      "Epoch 1717/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.6455 - categorical_accuracy: 0.6771\n",
      "Epoch 1718/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6336 - categorical_accuracy: 0.6928\n",
      "Epoch 1719/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.5900 - categorical_accuracy: 0.7261\n",
      "Epoch 1720/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6077 - categorical_accuracy: 0.5969\n",
      "Epoch 1721/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5699 - categorical_accuracy: 0.7593\n",
      "Epoch 1722/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5894 - categorical_accuracy: 0.7573\n",
      "Epoch 1723/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.7277 - categorical_accuracy: 0.5813\n",
      "Epoch 1724/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6258 - categorical_accuracy: 0.7162\n",
      "Epoch 1725/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5820 - categorical_accuracy: 0.7163\n",
      "Epoch 1726/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.8086 - categorical_accuracy: 0.5421\n",
      "Epoch 1727/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 1.0127 - categorical_accuracy: 0.5070\n",
      "Epoch 1728/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.7489 - categorical_accuracy: 0.5460\n",
      "Epoch 1729/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.7858 - categorical_accuracy: 0.6126\n",
      "Epoch 1730/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.7761 - categorical_accuracy: 0.5989\n",
      "Epoch 1731/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6885 - categorical_accuracy: 0.6146\n",
      "Epoch 1732/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6488 - categorical_accuracy: 0.6750\n",
      "Epoch 1733/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6824 - categorical_accuracy: 0.5715\n",
      "Epoch 1734/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5919 - categorical_accuracy: 0.7241\n",
      "Epoch 1735/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6866 - categorical_accuracy: 0.6145\n",
      "Epoch 1736/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.6041 - categorical_accuracy: 0.7652\n",
      "Epoch 1737/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6406 - categorical_accuracy: 0.7182\n",
      "Epoch 1738/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.6220 - categorical_accuracy: 0.6399\n",
      "Epoch 1739/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6992 - categorical_accuracy: 0.6183\n",
      "Epoch 1740/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.8536 - categorical_accuracy: 0.6066\n",
      "Epoch 1741/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.7559 - categorical_accuracy: 0.5754\n",
      "Epoch 1742/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6263 - categorical_accuracy: 0.7201\n",
      "Epoch 1743/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.7215 - categorical_accuracy: 0.6439\n",
      "Epoch 1744/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.6315 - categorical_accuracy: 0.6575\n",
      "Epoch 1745/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.7244 - categorical_accuracy: 0.5871\n",
      "Epoch 1746/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5800 - categorical_accuracy: 0.7671\n",
      "Epoch 1747/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6227 - categorical_accuracy: 0.7123\n",
      "Epoch 1748/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5909 - categorical_accuracy: 0.6771\n",
      "Epoch 1749/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.5947 - categorical_accuracy: 0.7182\n",
      "Epoch 1750/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5640 - categorical_accuracy: 0.6458\n",
      "Epoch 1751/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5579 - categorical_accuracy: 0.7163\n",
      "Epoch 1752/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5645 - categorical_accuracy: 0.7398\n",
      "Epoch 1753/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5540 - categorical_accuracy: 0.7925\n",
      "Epoch 1754/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5818 - categorical_accuracy: 0.7026\n",
      "Epoch 1755/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6211 - categorical_accuracy: 0.7045\n",
      "Epoch 1756/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5807 - categorical_accuracy: 0.7045\n",
      "Epoch 1757/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6531 - categorical_accuracy: 0.6635\n",
      "Epoch 1758/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6497 - categorical_accuracy: 0.6713\n",
      "Epoch 1759/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5604 - categorical_accuracy: 0.6693\n",
      "Epoch 1760/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6072 - categorical_accuracy: 0.7495\n",
      "Epoch 1761/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5572 - categorical_accuracy: 0.7377\n",
      "Epoch 1762/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.7188 - categorical_accuracy: 0.6144\n",
      "Epoch 1763/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5807 - categorical_accuracy: 0.7142\n",
      "Epoch 1764/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6433 - categorical_accuracy: 0.6693\n",
      "Epoch 1765/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6210 - categorical_accuracy: 0.6908\n",
      "Epoch 1766/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6362 - categorical_accuracy: 0.6477\n",
      "Epoch 1767/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6258 - categorical_accuracy: 0.6653\n",
      "Epoch 1768/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.7099 - categorical_accuracy: 0.6438\n",
      "Epoch 1769/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.6512 - categorical_accuracy: 0.6653\n",
      "Epoch 1770/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6788 - categorical_accuracy: 0.6751\n",
      "Epoch 1771/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5638 - categorical_accuracy: 0.7397\n",
      "Epoch 1772/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5908 - categorical_accuracy: 0.6693\n",
      "Epoch 1773/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6166 - categorical_accuracy: 0.7084\n",
      "Epoch 1774/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5869 - categorical_accuracy: 0.6927\n",
      "Epoch 1775/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.6254 - categorical_accuracy: 0.7201\n",
      "Epoch 1776/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5925 - categorical_accuracy: 0.7163\n",
      "Epoch 1777/2000\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 0.6168 - categorical_accuracy: 0.6615\n",
      "Epoch 1778/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5816 - categorical_accuracy: 0.7163\n",
      "Epoch 1779/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5644 - categorical_accuracy: 0.6635\n",
      "Epoch 1780/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5142 - categorical_accuracy: 0.7984\n",
      "Epoch 1781/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5248 - categorical_accuracy: 0.7593\n",
      "Epoch 1782/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5394 - categorical_accuracy: 0.7398\n",
      "Epoch 1783/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5531 - categorical_accuracy: 0.7848\n",
      "Epoch 1784/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5737 - categorical_accuracy: 0.7691\n",
      "Epoch 1785/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5863 - categorical_accuracy: 0.7378\n",
      "Epoch 1786/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5020 - categorical_accuracy: 0.7847\n",
      "Epoch 1787/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5586 - categorical_accuracy: 0.7534\n",
      "Epoch 1788/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5340 - categorical_accuracy: 0.7397\n",
      "Epoch 1789/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6455 - categorical_accuracy: 0.7182\n",
      "Epoch 1790/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 1.1071 - categorical_accuracy: 0.4971\n",
      "Epoch 1791/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.9106 - categorical_accuracy: 0.6653\n",
      "Epoch 1792/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 1.2587 - categorical_accuracy: 0.5069\n",
      "Epoch 1793/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 1.1683 - categorical_accuracy: 0.5303\n",
      "Epoch 1794/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.9156 - categorical_accuracy: 0.5949\n",
      "Epoch 1795/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.8787 - categorical_accuracy: 0.5244\n",
      "Epoch 1796/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.9585 - categorical_accuracy: 0.5108\n",
      "Epoch 1797/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.9123 - categorical_accuracy: 0.5480\n",
      "Epoch 1798/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.8000 - categorical_accuracy: 0.5812\n",
      "Epoch 1799/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.7512 - categorical_accuracy: 0.5969\n",
      "Epoch 1800/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6888 - categorical_accuracy: 0.6400\n",
      "Epoch 1801/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6528 - categorical_accuracy: 0.6849\n",
      "Epoch 1802/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6594 - categorical_accuracy: 0.6614\n",
      "Epoch 1803/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6864 - categorical_accuracy: 0.6145\n",
      "Epoch 1804/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6233 - categorical_accuracy: 0.7181\n",
      "Epoch 1805/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6296 - categorical_accuracy: 0.7280\n",
      "Epoch 1806/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6286 - categorical_accuracy: 0.7730\n",
      "Epoch 1807/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5993 - categorical_accuracy: 0.7182\n",
      "Epoch 1808/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5884 - categorical_accuracy: 0.7201\n",
      "Epoch 1809/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5986 - categorical_accuracy: 0.7671\n",
      "Epoch 1810/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6176 - categorical_accuracy: 0.7574\n",
      "Epoch 1811/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5720 - categorical_accuracy: 0.7729\n",
      "Epoch 1812/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5866 - categorical_accuracy: 0.7182\n",
      "Epoch 1813/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6039 - categorical_accuracy: 0.7691\n",
      "Epoch 1814/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5755 - categorical_accuracy: 0.7339\n",
      "Epoch 1815/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6032 - categorical_accuracy: 0.7593\n",
      "Epoch 1816/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5705 - categorical_accuracy: 0.7124\n",
      "Epoch 1817/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5447 - categorical_accuracy: 0.7632\n",
      "Epoch 1818/2000\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 0.4997 - categorical_accuracy: 0.7886\n",
      "Epoch 1819/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5587 - categorical_accuracy: 0.7983\n",
      "Epoch 1820/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4986 - categorical_accuracy: 0.8042\n",
      "Epoch 1821/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.5106 - categorical_accuracy: 0.7201\n",
      "Epoch 1822/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5860 - categorical_accuracy: 0.7280\n",
      "Epoch 1823/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5546 - categorical_accuracy: 0.7494\n",
      "Epoch 1824/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5583 - categorical_accuracy: 0.7319\n",
      "Epoch 1825/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5044 - categorical_accuracy: 0.8004\n",
      "Epoch 1826/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5356 - categorical_accuracy: 0.8023\n",
      "Epoch 1827/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5083 - categorical_accuracy: 0.7574\n",
      "Epoch 1828/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5696 - categorical_accuracy: 0.7573\n",
      "Epoch 1829/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5448 - categorical_accuracy: 0.7574\n",
      "Epoch 1830/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5680 - categorical_accuracy: 0.7065\n",
      "Epoch 1831/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5120 - categorical_accuracy: 0.8023\n",
      "Epoch 1832/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5102 - categorical_accuracy: 0.7828\n",
      "Epoch 1833/2000\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 0.5146 - categorical_accuracy: 0.7182\n",
      "Epoch 1834/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4711 - categorical_accuracy: 0.7925\n",
      "Epoch 1835/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5114 - categorical_accuracy: 0.7808\n",
      "Epoch 1836/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6855 - categorical_accuracy: 0.6908\n",
      "Epoch 1837/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.7151 - categorical_accuracy: 0.5949\n",
      "Epoch 1838/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.7670 - categorical_accuracy: 0.6536\n",
      "Epoch 1839/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5441 - categorical_accuracy: 0.7592\n",
      "Epoch 1840/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5775 - categorical_accuracy: 0.7515\n",
      "Epoch 1841/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.5870 - categorical_accuracy: 0.7416\n",
      "Epoch 1842/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5213 - categorical_accuracy: 0.7651\n",
      "Epoch 1843/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6425 - categorical_accuracy: 0.6674\n",
      "Epoch 1844/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5146 - categorical_accuracy: 0.7456\n",
      "Epoch 1845/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5585 - categorical_accuracy: 0.7025\n",
      "Epoch 1846/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.6880 - categorical_accuracy: 0.6790\n",
      "Epoch 1847/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6516 - categorical_accuracy: 0.6556\n",
      "Epoch 1848/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.5648 - categorical_accuracy: 0.7142\n",
      "Epoch 1849/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6171 - categorical_accuracy: 0.6673\n",
      "Epoch 1850/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5672 - categorical_accuracy: 0.7025\n",
      "Epoch 1851/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.7002 - categorical_accuracy: 0.6420\n",
      "Epoch 1852/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5902 - categorical_accuracy: 0.7338\n",
      "Epoch 1853/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.5591 - categorical_accuracy: 0.7671\n",
      "Epoch 1854/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5427 - categorical_accuracy: 0.7574\n",
      "Epoch 1855/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5202 - categorical_accuracy: 0.7867\n",
      "Epoch 1856/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5190 - categorical_accuracy: 0.7574\n",
      "Epoch 1857/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.5206 - categorical_accuracy: 0.7398\n",
      "Epoch 1858/2000\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 0.5519 - categorical_accuracy: 0.7437\n",
      "Epoch 1859/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5890 - categorical_accuracy: 0.7182\n",
      "Epoch 1860/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4965 - categorical_accuracy: 0.7867\n",
      "Epoch 1861/2000\n",
      "3/3 [==============================] - 0s 99ms/step - loss: 0.4878 - categorical_accuracy: 0.8082\n",
      "Epoch 1862/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.4779 - categorical_accuracy: 0.8141\n",
      "Epoch 1863/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.4453 - categorical_accuracy: 0.8043\n",
      "Epoch 1864/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.4839 - categorical_accuracy: 0.7867\n",
      "Epoch 1865/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4624 - categorical_accuracy: 0.8102\n",
      "Epoch 1866/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5238 - categorical_accuracy: 0.7476\n",
      "Epoch 1867/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5181 - categorical_accuracy: 0.7828\n",
      "Epoch 1868/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5559 - categorical_accuracy: 0.7495\n",
      "Epoch 1869/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4538 - categorical_accuracy: 0.8493\n",
      "Epoch 1870/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5433 - categorical_accuracy: 0.7613\n",
      "Epoch 1871/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4883 - categorical_accuracy: 0.7417\n",
      "Epoch 1872/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4763 - categorical_accuracy: 0.7711\n",
      "Epoch 1873/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.4762 - categorical_accuracy: 0.8239\n",
      "Epoch 1874/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5052 - categorical_accuracy: 0.7594\n",
      "Epoch 1875/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5103 - categorical_accuracy: 0.7926\n",
      "Epoch 1876/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5458 - categorical_accuracy: 0.7241\n",
      "Epoch 1877/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5457 - categorical_accuracy: 0.7202\n",
      "Epoch 1878/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5567 - categorical_accuracy: 0.7730\n",
      "Epoch 1879/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5607 - categorical_accuracy: 0.7221\n",
      "Epoch 1880/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4913 - categorical_accuracy: 0.7906\n",
      "Epoch 1881/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5630 - categorical_accuracy: 0.7456\n",
      "Epoch 1882/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5284 - categorical_accuracy: 0.7886\n",
      "Epoch 1883/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.7266 - categorical_accuracy: 0.6047\n",
      "Epoch 1884/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6122 - categorical_accuracy: 0.6986\n",
      "Epoch 1885/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.7095 - categorical_accuracy: 0.6537\n",
      "Epoch 1886/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5932 - categorical_accuracy: 0.7691\n",
      "Epoch 1887/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5668 - categorical_accuracy: 0.7456\n",
      "Epoch 1888/2000\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 0.4465 - categorical_accuracy: 0.7867\n",
      "Epoch 1889/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.4896 - categorical_accuracy: 0.8004\n",
      "Epoch 1890/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5150 - categorical_accuracy: 0.8141\n",
      "Epoch 1891/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4883 - categorical_accuracy: 0.7260\n",
      "Epoch 1892/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.4889 - categorical_accuracy: 0.8004\n",
      "Epoch 1893/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.4458 - categorical_accuracy: 0.8199\n",
      "Epoch 1894/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.4541 - categorical_accuracy: 0.8239\n",
      "Epoch 1895/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4740 - categorical_accuracy: 0.7847\n",
      "Epoch 1896/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4761 - categorical_accuracy: 0.7632\n",
      "Epoch 1897/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5181 - categorical_accuracy: 0.7457\n",
      "Epoch 1898/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.4631 - categorical_accuracy: 0.7710\n",
      "Epoch 1899/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.4392 - categorical_accuracy: 0.8003\n",
      "Epoch 1900/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.4863 - categorical_accuracy: 0.7729\n",
      "Epoch 1901/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5068 - categorical_accuracy: 0.7750\n",
      "Epoch 1902/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.7582 - categorical_accuracy: 0.7241\n",
      "Epoch 1903/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5602 - categorical_accuracy: 0.7242\n",
      "Epoch 1904/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4917 - categorical_accuracy: 0.7691\n",
      "Epoch 1905/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5019 - categorical_accuracy: 0.7926\n",
      "Epoch 1906/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4088 - categorical_accuracy: 0.8278\n",
      "Epoch 1907/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.4292 - categorical_accuracy: 0.8551\n",
      "Epoch 1908/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.4594 - categorical_accuracy: 0.7594\n",
      "Epoch 1909/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4589 - categorical_accuracy: 0.8043\n",
      "Epoch 1910/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.4781 - categorical_accuracy: 0.7828\n",
      "Epoch 1911/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5776 - categorical_accuracy: 0.7358\n",
      "Epoch 1912/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.7581 - categorical_accuracy: 0.7065\n",
      "Epoch 1913/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.8014 - categorical_accuracy: 0.6614\n",
      "Epoch 1914/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.7756 - categorical_accuracy: 0.6809\n",
      "Epoch 1915/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5726 - categorical_accuracy: 0.6986\n",
      "Epoch 1916/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4787 - categorical_accuracy: 0.7984\n",
      "Epoch 1917/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5448 - categorical_accuracy: 0.7476\n",
      "Epoch 1918/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.5307 - categorical_accuracy: 0.7730\n",
      "Epoch 1919/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5646 - categorical_accuracy: 0.7221\n",
      "Epoch 1920/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5307 - categorical_accuracy: 0.7652\n",
      "Epoch 1921/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5483 - categorical_accuracy: 0.8102\n",
      "Epoch 1922/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4969 - categorical_accuracy: 0.7613\n",
      "Epoch 1923/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.5821 - categorical_accuracy: 0.7006\n",
      "Epoch 1924/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4577 - categorical_accuracy: 0.8473\n",
      "Epoch 1925/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4711 - categorical_accuracy: 0.8219\n",
      "Epoch 1926/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4637 - categorical_accuracy: 0.8512\n",
      "Epoch 1927/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4982 - categorical_accuracy: 0.7651\n",
      "Epoch 1928/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.4124 - categorical_accuracy: 0.8376\n",
      "Epoch 1929/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4342 - categorical_accuracy: 0.8239\n",
      "Epoch 1930/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4387 - categorical_accuracy: 0.7984\n",
      "Epoch 1931/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.4321 - categorical_accuracy: 0.8395\n",
      "Epoch 1932/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.3992 - categorical_accuracy: 0.8376\n",
      "Epoch 1933/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.4175 - categorical_accuracy: 0.8493\n",
      "Epoch 1934/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.4514 - categorical_accuracy: 0.7945\n",
      "Epoch 1935/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.6848 - categorical_accuracy: 0.7162\n",
      "Epoch 1936/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4836 - categorical_accuracy: 0.7984\n",
      "Epoch 1937/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.4199 - categorical_accuracy: 0.8532\n",
      "Epoch 1938/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.4919 - categorical_accuracy: 0.8003\n",
      "Epoch 1939/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.4078 - categorical_accuracy: 0.8551\n",
      "Epoch 1940/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5356 - categorical_accuracy: 0.7808\n",
      "Epoch 1941/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.4106 - categorical_accuracy: 0.8552\n",
      "Epoch 1942/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5413 - categorical_accuracy: 0.7182\n",
      "Epoch 1943/2000\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 0.4422 - categorical_accuracy: 0.8062\n",
      "Epoch 1944/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.7053 - categorical_accuracy: 0.6733\n",
      "Epoch 1945/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.6186 - categorical_accuracy: 0.7554\n",
      "Epoch 1946/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6433 - categorical_accuracy: 0.6888\n",
      "Epoch 1947/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6444 - categorical_accuracy: 0.6927\n",
      "Epoch 1948/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6215 - categorical_accuracy: 0.7769\n",
      "Epoch 1949/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.5888 - categorical_accuracy: 0.7632\n",
      "Epoch 1950/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5109 - categorical_accuracy: 0.7907\n",
      "Epoch 1951/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5041 - categorical_accuracy: 0.7612\n",
      "Epoch 1952/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.4787 - categorical_accuracy: 0.7926\n",
      "Epoch 1953/2000\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 0.4985 - categorical_accuracy: 0.7691\n",
      "Epoch 1954/2000\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 0.5288 - categorical_accuracy: 0.7945\n",
      "Epoch 1955/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5304 - categorical_accuracy: 0.7437\n",
      "Epoch 1956/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.4753 - categorical_accuracy: 0.7652\n",
      "Epoch 1957/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.5261 - categorical_accuracy: 0.7574\n",
      "Epoch 1958/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.7336 - categorical_accuracy: 0.6948\n",
      "Epoch 1959/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5221 - categorical_accuracy: 0.7691\n",
      "Epoch 1960/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.7378 - categorical_accuracy: 0.6596\n",
      "Epoch 1961/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.7095 - categorical_accuracy: 0.6185\n",
      "Epoch 1962/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.8044 - categorical_accuracy: 0.6242\n",
      "Epoch 1963/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.8233 - categorical_accuracy: 0.6066\n",
      "Epoch 1964/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.6391 - categorical_accuracy: 0.7064\n",
      "Epoch 1965/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.8835 - categorical_accuracy: 0.5304\n",
      "Epoch 1966/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.7887 - categorical_accuracy: 0.5792\n",
      "Epoch 1967/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.8095 - categorical_accuracy: 0.5930\n",
      "Epoch 1968/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.7175 - categorical_accuracy: 0.6007\n",
      "Epoch 1969/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.7272 - categorical_accuracy: 0.5949\n",
      "Epoch 1970/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.6985 - categorical_accuracy: 0.6125\n",
      "Epoch 1971/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.6695 - categorical_accuracy: 0.5890\n",
      "Epoch 1972/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.6442 - categorical_accuracy: 0.6243\n",
      "Epoch 1973/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6141 - categorical_accuracy: 0.7065\n",
      "Epoch 1974/2000\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 0.6586 - categorical_accuracy: 0.6850\n",
      "Epoch 1975/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.6544 - categorical_accuracy: 0.6849\n",
      "Epoch 1976/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6372 - categorical_accuracy: 0.7378\n",
      "Epoch 1977/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5893 - categorical_accuracy: 0.7534\n",
      "Epoch 1978/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6029 - categorical_accuracy: 0.6830\n",
      "Epoch 1979/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5751 - categorical_accuracy: 0.7319\n",
      "Epoch 1980/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5666 - categorical_accuracy: 0.7729\n",
      "Epoch 1981/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6021 - categorical_accuracy: 0.7339\n",
      "Epoch 1982/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5719 - categorical_accuracy: 0.7828\n",
      "Epoch 1983/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5850 - categorical_accuracy: 0.7437\n",
      "Epoch 1984/2000\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 0.5031 - categorical_accuracy: 0.8062\n",
      "Epoch 1985/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5171 - categorical_accuracy: 0.7475\n",
      "Epoch 1986/2000\n",
      "3/3 [==============================] - 0s 102ms/step - loss: 0.5368 - categorical_accuracy: 0.7495\n",
      "Epoch 1987/2000\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 0.5211 - categorical_accuracy: 0.7632\n",
      "Epoch 1988/2000\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 0.5681 - categorical_accuracy: 0.7789\n",
      "Epoch 1989/2000\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 0.5665 - categorical_accuracy: 0.7632\n",
      "Epoch 1990/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5057 - categorical_accuracy: 0.8121\n",
      "Epoch 1991/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.6042 - categorical_accuracy: 0.7867\n",
      "Epoch 1992/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5793 - categorical_accuracy: 0.7730\n",
      "Epoch 1993/2000\n",
      "3/3 [==============================] - 0s 101ms/step - loss: 0.5411 - categorical_accuracy: 0.7534\n",
      "Epoch 1994/2000\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 0.5965 - categorical_accuracy: 0.6595\n",
      "Epoch 1995/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5287 - categorical_accuracy: 0.7319\n",
      "Epoch 1996/2000\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5031 - categorical_accuracy: 0.7671\n",
      "Epoch 1997/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 105ms/step - loss: 0.5730 - categorical_accuracy: 0.7143\n",
      "Epoch 1998/2000\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 0.6528 - categorical_accuracy: 0.6733\n",
      "Epoch 1999/2000\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 0.5453 - categorical_accuracy: 0.7378\n",
      "Epoch 2000/2000\n",
      "3/3 [==============================] - 0s 103ms/step - loss: 0.5129 - categorical_accuracy: 0.8004\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x1e4851c3be0>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, y_train, epochs=2000, callbacks=[tb_callback])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "1db0e01a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 30, 64)            442112    \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 30, 128)           98816     \n",
      "_________________________________________________________________\n",
      "lstm_2 (LSTM)                (None, 64)                49408     \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                4160      \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 32)                2080      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 3)                 99        \n",
      "=================================================================\n",
      "Total params: 596,675\n",
      "Trainable params: 596,675\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7bb59f70",
   "metadata": {},
   "source": [
    "# Make Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "282c5a1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "77746ded",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'thanks'"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actions[np.argmax(res[4])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "9af0bb39",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'thanks'"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actions[np.argmax(y_test[4])]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3588ba56",
   "metadata": {},
   "source": [
    "#  Save Weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "09fff5fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('action.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "f394583f",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.load_weights('action.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21c0516c",
   "metadata": {},
   "source": [
    "# Evaluation using Confusion Matrix and Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "0dd447a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import multilabel_confusion_matrix, accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "f504e5d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "yhat = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "a8cfe684",
   "metadata": {},
   "outputs": [],
   "source": [
    "ytrue = np.argmax(y_test, axis=1).tolist()\n",
    "yhat = np.argmax(yhat, axis=1).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "fac7dc52",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[4, 1],\n",
       "        [0, 0]],\n",
       "\n",
       "       [[1, 0],\n",
       "        [1, 3]],\n",
       "\n",
       "       [[4, 0],\n",
       "        [0, 1]]], dtype=int64)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "multilabel_confusion_matrix(ytrue, yhat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "5b8914cc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(ytrue, yhat)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "103e8c5c",
   "metadata": {},
   "source": [
    "# Test in Real Time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "33b43507",
   "metadata": {},
   "outputs": [],
   "source": [
    "colors = [(245,117,16), (117,245,16), (16,117,245)]\n",
    "def prob_viz(res, actions, input_frame, colors):\n",
    "    output_frame = input_frame.copy()\n",
    "    for num, prob in enumerate(res):\n",
    "        cv2.rectangle(output_frame, (0,60+num*40), (int(prob*100), 90+num*40), colors[num], -1)\n",
    "        cv2.putText(output_frame, actions[num], (0, 85+num*40), cv2.FONT_HERSHEY_SIMPLEX, 1, (255,255,255), 2, cv2.LINE_AA)\n",
    "        \n",
    "    return output_frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "f512f8f7",
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "only size-1 arrays can be converted to Python scalars",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-65-0248883a8ffa>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m18\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m18\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mimshow\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprob_viz\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mres\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mimage\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcolors\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m<ipython-input-64-e9bc22b256b7>\u001b[0m in \u001b[0;36mprob_viz\u001b[1;34m(res, actions, input_frame, colors)\u001b[0m\n\u001b[0;32m      3\u001b[0m     \u001b[0moutput_frame\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0minput_frame\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m     \u001b[1;32mfor\u001b[0m \u001b[0mnum\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprob\u001b[0m \u001b[1;32min\u001b[0m \u001b[0menumerate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mres\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 5\u001b[1;33m         \u001b[0mcv2\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mrectangle\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0moutput_frame\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m60\u001b[0m\u001b[1;33m+\u001b[0m\u001b[0mnum\u001b[0m\u001b[1;33m*\u001b[0m\u001b[1;36m40\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprob\u001b[0m\u001b[1;33m*\u001b[0m\u001b[1;36m100\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m90\u001b[0m\u001b[1;33m+\u001b[0m\u001b[0mnum\u001b[0m\u001b[1;33m*\u001b[0m\u001b[1;36m40\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcolors\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mnum\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      6\u001b[0m         \u001b[0mcv2\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mputText\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0moutput_frame\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mnum\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m85\u001b[0m\u001b[1;33m+\u001b[0m\u001b[0mnum\u001b[0m\u001b[1;33m*\u001b[0m\u001b[1;36m40\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcv2\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mFONT_HERSHEY_SIMPLEX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;36m255\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m255\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m255\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcv2\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mLINE_AA\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mTypeError\u001b[0m: only size-1 arrays can be converted to Python scalars"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1296x1296 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(18,18))\n",
    "plt.imshow(prob_viz(res, actions, image, colors))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "8d7428f9",
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'int' object has no attribute 'reverse'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-66-bcb169611ca5>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0msequence\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreverse\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'int' object has no attribute 'reverse'"
     ]
    }
   ],
   "source": [
    "sequence.reverse()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "f81f8d02",
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "object of type 'int' has no len()",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-67-63815c6569bf>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0msequence\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mTypeError\u001b[0m: object of type 'int' has no len()"
     ]
    }
   ],
   "source": [
    "len(sequence)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "334e4360",
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'int' object has no attribute 'append'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-68-6fafd5216796>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0msequence\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'def'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'int' object has no attribute 'append'"
     ]
    }
   ],
   "source": [
    "sequence.append('def')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "ae6185cf",
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'int' object has no attribute 'reverse'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-69-bcb169611ca5>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0msequence\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreverse\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'int' object has no attribute 'reverse'"
     ]
    }
   ],
   "source": [
    "sequence.reverse()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "110feefe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "hello\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "thanks\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n",
      "<class 'mediapipe.python.solution_base.SolutionOutputs'>\n",
      "iloveyou\n"
     ]
    }
   ],
   "source": [
    "# 1. New detection variables\n",
    "sequence = []\n",
    "sentence = []\n",
    "threshold = 0.8\n",
    "\n",
    "cap = cv2.VideoCapture(0)\n",
    "# Set mediapipe model \n",
    "with mp_holistic.Holistic(min_detection_confidence=0.5, min_tracking_confidence=0.5) as holistic:\n",
    "    while cap.isOpened():\n",
    "\n",
    "        # Read feed\n",
    "        ret, frame = cap.read()\n",
    "\n",
    "        # Make detections\n",
    "        image, results = mediapipe_detection(frame, holistic)\n",
    "        print(results)\n",
    "        \n",
    "        # Draw landmarks\n",
    "        draw_styled_landmarks(image, results)\n",
    "        \n",
    "        # 2. Prediction logic\n",
    "        keypoints = extract_keypoints(results)\n",
    "#         sequence.insert(0,keypoints)\n",
    "#         sequence = sequence[:30]\n",
    "        sequence.append(keypoints)\n",
    "        sequence = sequence[-30:]\n",
    "        \n",
    "        if len(sequence) == 30:\n",
    "            res = model.predict(np.expand_dims(sequence, axis=0))[0]\n",
    "            print(actions[np.argmax(res)])\n",
    "            \n",
    "            \n",
    "        #3. Viz logic\n",
    "            if res[np.argmax(res)] > threshold: \n",
    "                if len(sentence) > 0: \n",
    "                    if actions[np.argmax(res)] != sentence[-1]:\n",
    "                        sentence.append(actions[np.argmax(res)])\n",
    "                else:\n",
    "                    sentence.append(actions[np.argmax(res)])\n",
    "\n",
    "            if len(sentence) > 5: \n",
    "                sentence = sentence[-5:]\n",
    "\n",
    "            # Viz probabilities\n",
    "            image = prob_viz(res, actions, image, colors)\n",
    "            \n",
    "        cv2.rectangle(image, (0,0), (640, 40), (245, 117, 16), -1)\n",
    "        cv2.putText(image, ' '.join(sentence), (3,30), \n",
    "                       cv2.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 2, cv2.LINE_AA)\n",
    "        \n",
    "        # Show to screen\n",
    "        cv2.imshow('OpenCV Feed', image)\n",
    "\n",
    "        # Break gracefully\n",
    "        if cv2.waitKey(10) & 0xFF == ord('q'):\n",
    "            break\n",
    "    cap.release()\n",
    "    cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "7ac76ea3",
   "metadata": {},
   "outputs": [],
   "source": [
    "cap.release()\n",
    "cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "2435abf0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res[np.argmax(res)] > threshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "f3434f3a",
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'num_sequences' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-73-925db0742f6b>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[1;33m(\u001b[0m\u001b[0mnum_sequences\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m30\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m1662\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'num_sequences' is not defined"
     ]
    }
   ],
   "source": [
    "(num_sequences,30,1662)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "054a2b96",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.29763258, 0.00202154, 0.7003458 ]], dtype=float32)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(np.expand_dims(X_test[0], axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dc9fb2ab",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}