LAB6_RNN_Exercise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "R3MkAcSjlHjN"
   },
   "source": [
    "# Recurrent Neural Network (RNN)\n",
    "\n",
    "RNN is a special kind of neural network that allows us capturing temporal dynamic behavior. \n",
    "  \n",
    "![](https://miro.medium.com/max/724/1*dp-NFSRVZCvvldrhnZaQow.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 296
    },
    "id": "2GqLXrTY1-gC",
    "outputId": "e2df3c31-b913-4edc-e764-e823bf3e8a1a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f2314f70cd0>]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEGCAYAAABiq/5QAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9aZQt2VkduE8Md8g535BvqCrxalRVUbJKokoIKLAGBFhiEBjT0LSsboSFV6NlbEPb0N0LYa/VxuAGjE2DLSEh0RaS1RZrSQhhLCRBaVaVZFVJqirVPL4p33s53imm0z9OfBEnTpwTEXeIm5kv716rVr2XL2/euHnjfrFjf/vbH+OcY4YZZphhhsMDa68PYIYZZphhhuliVvhnmGGGGQ4ZZoV/hhlmmOGQYVb4Z5hhhhkOGWaFf4YZZpjhkMHZ6wOogmPHjvEzZ87s9WHMMMMMMxwofPnLX77EOT+ufv1AFP4zZ87g/vvv3+vDmGGGGWY4UGCMPaP7+kzqmWGGGWY4ZJgV/hlmmGGGQ4ZZ4Z9hhhlmOGSYFf4ZZphhhkOG2go/Y6zFGPsSY+wBxtg3GGP/Iv769YyxLzLGHmeM/WfGWKOuY5hhhhlmmCGPOhn/AMBrOOcvBXAngB9gjL0SwG8A+B3O+U0ANgC8pcZjmGGGGWaYQUFthZ8L7MZ/deP/OIDXAPgv8dffC+CNdR3DDDPMMMMMedSq8TPGbMbYVwFcBPBxAE8A2OScB/G3PA/gGsNj38oYu58xdv/6+vpIz/+Jhy/g9//68ZEeO8MMM8xwtaLWws85DznndwK4FsArANw6xGPfwTm/i3N+1/HjucGzSvj0Y5fw+596YqTHzjDDDDNcrZiKq4dzvgngUwC+A8AKY4wmhq8F8EJdz7u21MTuIEDXC8q/eYYZZpjhkKBOV89xxthK/Oc2gNcBeBjiAvDj8be9GcCH6zqGtcUWAODi9qCup5hhhqmg54V7fQgzXEWok/GfAvApxtiDAO4D8HHO+UcB/HMA/5Qx9jiAowDeVdcBnFhqAgAu7swK/wwHFw+f28Ydv/aXeOpSZ68PZd/h+Y0uPvXIxb0+jAOH2kLaOOcPAniZ5utPQuj9tSNh/Dv9aTzdDDPUgmevdBFGHBe2+7j+2PxeH86+wns++zTe9dmn8Gdvuwd3XLO814dzYHBVT+6uLcaMfyb1zHCA0RmIHlUY8T0+kv2H3UEAzoF/+dGHwPns91MVV3XhX5lz0bCtmdQzw4FGJ9b3gz0u/H4Y4dxWb0+PQUXfF7+bLz11BX/x9fN7fDQHB1d14WeM4fhiExe3Z1LPDAcXxPiDMNrT4/jTrzyP1/7W3yTFdj+g70e4aW0Bt55cxL/62MP76tj2M67qwg9AFP4Z45/hAKNLhX+PGf+lXQ9dL8ROf//Yo3t+iPmGjV/9wdvx/EYP7/7sU3t9SAcCV33hX1tszpq7Mxxo7A5iqSfc28LvBeKOYz+x6r4founa+M6bjuF7bzuB3//UE7NeSAVc/YV/acb4ZzjYoAHEINpbqccPqxX+aTZZ+36ItmsDAO4+s4rdQYBBsH8uTPsVV3/hX2xhs+tXOhmCMMKHv/rCzB0ww77C7j5x9ZDU1Cso/E+u7+L2X/1LPHZhZyrH1PcjtFxRxpqO+P/A39sL5EHAISj8wtK5XoH1f+6Jy/iFD3wVX39hu+7DmmGGyuh6+03qMRfWrzy7iZ4f4rmN7lSOqR+kjL8Z/38QTKfw3//0FTw6pQvcpHH1F/4hpneJyRQxmhlmmDZ290lzl6Seos/H4xdFErsXTOdYe16IFhV+YvxTkHrObfXwpnd9Cb/93x6t/bnqwNVf+IfI66ETe6+11BlmkEEaf7hPNP6i3CAq/P6UrKd9Xy78dvy1+p/71z/2CHp+eGD7CYeg8JPUU+7soVvpvb6lnmEGGZ3Y1ePv8XlJz19U7J5cJ8Y/rcIfJYWftP66i/EXn7yMjzxwFsDe34WNitqyevYLji40YbFqUg+xlL1uos0wg4z9EtnglTB+L4jwzBWh7U+D8YcRhxfKzd36Nf4gjPD2j3wDp5dbOLrQnNqdzaRx1TN+22I4ttCsJPXQ1fugvpkzXJ2gwu/vtdQTFGv8T1/uJBenaXyGiNmnzd36XT3vv+85PHJ+B//HG27HQtPZ84vxqLjqCz8gGrwXKkk9pPEfzDcTALb7Pn72vffj6y9s7fWhzDABRBFHNy604Z5LPcWunicu7iZ/9qZwrHTnMc3m7jvvfRJ3n1nF619yEo7N9lx+GxWHo/AvtioxfjpZD3Lh/92/egx/9fAFfO6JS3t9KDNMAD0/BI2V7PV5SUXOxPgflwv/FDT+fvwcqtRTV3OXc47z23287EWrYIzBsdiM8e9nrFXM60kY/wGVer55fgfv+dzTAIDLu97eHswME0FHWhs6CbeZH0Yjn99eyeTu4+u7yfKjaUg9KuOvu7nb80N4QYSVORcA4NjWgZWFD03hv9wZlJ7wxKgOoquHc463f+TrWGw5WJ1zcWlW+K8KkKMHmAzj/5n33Idf/cg3RnpsWWTDE+u7uOXEIiw2ncJPx6HaOetq7m52fQDA6lwDAODaTPuevP9Lz+LX/+LhWo5hUjgUhf/4UgucA5c7xcXQn7LG/74vPoNf/OADE/lZH33wHL7w5BX80ve9GNeuzuFyZ5ZPdDWAGrvAZAjJCxs93PfUlZEeW+TjjyKOJy52cNPaAhqONRWpJ9fcTSIb6mH8G11RP1Zjxm9bllbq+etvXsSffOHZytEvz1zu4B+9/79PzQILHJLCX3UTV+Ljn5J74pMPX8R/e2j85RF9P8T/9ecP445rlvBTr3gRji40cKXkIjfDwYBc+CehJw+CCE9d6oxUZPzArPGf3eqh54e48fgCXNtKZKE60fNI41dcPTUz/hVi/BbT3tkEIcfOIMD5intA7n10HR954Cyen1LMBXDYCn+Js4fexGl16l/Y7GGnH4zdU3jqUgfnt/v4B999A2yL4ch8Y6bxXyXoerLUM35BGwQRgojjmcvDL24vcvU8sS5+3k1rC2hMSftOpZ7p+PhTxi8Kv21o7vrx1x67sJv7Nx2udMQFZVoZQ8BhKfxLtHS9mPFTwZ/WaPy5LXEh2ur5Y/0cOmEWW2Ie79hCE5d2B7OU0asAmebuBAiJF8sjj1YsSpnHFmj85Oi5aU0wfn8KWT1050FSj20xuDarbV/ARqLxy83d/Ouk+lE1wI0uKLPCP2EcX6go9UTTY/ydQZAUfDqhRgXdtjds8QE4Ot/AIIgybHGG/YndQYCHzprTYEnqsdhwvaePPngWz17OSwdUXB67OHyqZFFI2+MXd7Ey5+LofENo/HvQ3AUE669N6onl05VMczf/XFQ/ZHtrEUiWras3ocOhKPwNx8LqnFtB6iHGX3/hl5dWb3bHk2WSwh83t47GF7qZ3LP/8Z++8Ax+7A8+azznyNWz1HaHOi//6QcfwPu+9Ezma5zzpCA/VrEoyaDPh45RP7G+ixuPL4AxwbqnUviDrMYPiAZvXXbOja6P+YadfM5si2mH6oIhf8fE+Pszxj95rC22cKG0uTs9H/8Lm+lFaHNcxh+KEz0p/POCkVyaOXv2Pa50PPT9yKiJE+NfarmVdXPOObwgykUXBBFPhsFGWZRSFNnwxMVd3HR8AQBiqWcKhd/LavxAXPhrGuDa7HoJ2wfi16lh/EGi8e9UkltnjL9GrC01SxM6k+buFBj/2c2U8W9MivHbxPjFyTlj/PsfnZKs/Y4XouFYaLl666AOphRNOk+ajoWnLnWGbsAmGr8iIW50PFzueLhpTRT+hjPt5q7E+N36pJ6NrofVeTf5u7G5G//+t/tBpcHRjc5VpPEzxq5jjH2KMfYQY+wbjLFfiL/+a4yxFxhjX43/e31dxyBjbbFV3tyNpij1bMpSz2Sau6rUc2XG+Pc9yBNvYsidQYCFpgPb0jcSdaCiqxYS+vutp5bgh8M7exJXj/Jzn4ijmG9cmwcgCMhU7Jx+CMdicO0s459Ec3ej4+XqwEbXTxw9ANk5eY7VB2E63Vvm7OGcJ/NFV0XhBxAA+EXO+e0AXgng5xljt8f/9juc8zvj/z5W4zEkOL7YLF2/GCR2zulIPSeXWnAsNjHG31Slnhnj3/cg144pebPjBZhr2HBtVtltFiSMP/v9dJ586+klANXthoAgQ1QH1QEuklBPr7QBYGquHjmLnzAJxr/V8/Hdv/kpfOC+ZzNfV6UeJ77gqDwxjDhuOxn/jkua6GKZC12orwKph3N+jnP+lfjPOwAeBnBNXc9XhrZrI4h4IZuf5iKWc1s9nF5pYWXOxeaYdk5iV8T4W66N+YY9k3oOAMh5ZWLznUGA+YYD29LHA+hA54OqdVPhv/3UEhgbrsFLZKhhW3FwXHosJFctNIWd2J2WqycI84V/As3dzz5+CbuDIHdhFIw/K/UAeaLoRxFOLYvPdpltVh60nOaS+Klo/IyxMwBeBuCL8Zfexhh7kDH2bsbYquExb2WM3c8Yu399fX3sY3Bs/Zskg07WaUQ2nN3s4fRKGytzjcm5eqRb3qMLzVlswwFAukhdf152vRDzTRuOxSoTEjrH1eJLBXG57eLa1fZQi8LpZ9GsiMyqd+LCv9gURbFh6ydaJ42+F2YauwAV/vGe+95HRb05v5X2BMOIY7vvK81dUVPUehGEHI7NcPPaAh4vYfwbnZT0XS1SDwCAMbYA4EMA/jHnfBvAHwC4EcCdAM4B+C3d4zjn7+Cc38U5v+v48eNjHwcVxaITMmX89b4BnHOc3erjmpU2Vtpu5s0fBaqdE8AstuGAIFmyYjjndgcB5psOHEMujA6Jxq9o3XIv6Ja1xco+cyDtQSy1RXGXdfTdvngN8007+fnTiWXWMX57LObMOU8K/zkpcmGr54NzKIxffN5US6cfctiWhZtPLOLRC7uFzp4rEum7KqQeAGCMuRBF/32c8z8FAM75Bc55yDmPALwTwCvqPAaCmzD+Aqknms7qxcsdD15At4ONybl65MI/35xp/AcAZI00nZfdQYj5hiOWflTU+H2Txh+mvaCbTizgyfVOZZJDP3MpZvyypbPjBWi5VqJ5u1OKbOh5YTK1S2i5FvpjFNAn1ndxdquPlmvh/FbeebeqYfzq+xJEEdyY8W/1/MLP4YYs9VwNjJ8xxgC8C8DDnPPflr5+Svq2HwXw9bqOQQadlEUnOp3cdds5ycp5eqWN1Tl37MgGL4zAGODEmiMgGryXd2dSz35HukjdzPjnYqlnWMavsm5PYfxemO7IrfozifHLDd6dfoCFZsqEXUOUwaQhmruq1DMe4/+bR8UCo9ffcQrrO2mUO8mxKxLjd4jx66Qey8LNa4sAimcm6K6cMXPcdR2ok/F/F4A3AXiNYt38TcbY1xhjDwJ4NYB/UuMxJCCpp6jpRIy/bqnnbDy8dXqljdX5yTD+hm1BXGsFSOqZ5fXsb3TJ1WNi/F5q5xxW41elg4Hk/rr5hPDcV3X2qBq/HNQmLKcp855WOqdW6nHH0/g//dg6bjg2j287s4qIA+sxeSI5Vmb8jqG5S4z/FvodF0hqVzoebIthda4xVcbv1PWDOeefAcA0/zQV+6YKau4WfXjIglZ3c1dm/MttF30/Qt/Pn8RVMQiijMwDiOZuEHFs9wIsSyxlhv2DKOKJZGIiG51BiLmGI5q7laUevY9fznS64bjw3D92YQc/cMfJyj9zqRUzflnjHwRYaKWlpDkljb/nhTgWz6zIzz2qVt73Q3zhycv4ybtfhFPLItjx3FYfp5bbWqnHVFOouXt8sYmlllNo6bzS9bA65459pzIsDs3krluhuetPifGf2+qhGecH0Yk0Duv3wijx8BOOLRzs2IbPPX4pGQy6WtEP0n26OsbvBRG8MMJC04Zj2PakA/0sk9TTdC3MNx1cu9rGA89vVborJFJkau6SlRMQ2vc0NP5BoPHxjxHSdv/TG+j7Eb7nlmM4uSRmEi7Ezp4ki38+b+eU3xfOOYJISD2MsaTBa8JGx8PqXKPWjCEdDlHhr9DcndKy9bObwtHDGEtcAuM4e0jqkXFk/mDHNvyTD34VP/zvP4NPPXJxrw+lNsjpqbrGLclAxPiHdvXkJnfjTKf4XHntrWv4q4cv4B994KvY6Reff4nU0ySpR2H8mcI/zeZu9rxvueJuIxrhM3zvY+to2BZeecPRDOMHBDFzLJa8fiAlk/KdGNUOkoFuObFQmNlzpeNhNU40vSqau/sNVRh/GtJWc+Hf6iVTjuQL3uyNwfiDCE2F+RydP9ixDT0vRMcL8Zb33of3xgvkrzZ0pX26usiGTnxhmJTGr7q/3v5D34r/7ftfjI997Rx+8N9/Bt84u1X6M5Pmbmnhz0cZTBomOydQ3Msz4d5H13HXmVXMNRyszLloOFayRWuj62Nlzs300RLGL70v9Gcyk7z4xCI2ur4xNWCj6+HIXKPWjCEdDl3hL9JJyc1T9+rFs5u9hFGQS2CcvB4d40+kngPK+P2Q46e//UV4za0n8PaPfAPv/sxTe31IE0dmyYqGoZLHf65pG7PfdfCCVOqRi69s5wQAy2L4+VffhA+89ZXoeiF+5U+/ZvyZaeGP7ZyeWeOnC0vdDd6+n7dzpnt3h3vujY6HR87v4J6bjwEAGGM4tdxKhrjUuAZAP8BFd270b7edEtEND53T71y40vGxOh9LPYqrp++HePziTnLnN0kcmsJPjRivIENkGqsXvSDCxZ1BwvgnpfGrzd3VCUg9v/lfH8Gb3vXF3Nc/eP9zuOc3Plkro/PioKv/+KZvw/XH5vH5Jy/X9lx7hYzUoymSVPjnm3Fkw5CMP+LZoqSb9wCAu88cwWtevFa4qMhPpJ5Y45fYKQ2ZEdJhyfrOD845+n7+TjfduzucXk6W6pPxtj7683lJ6llVTBJk55R7gjTMRVLPrXFmzyPn8w1ezrlg/PMuWhrG/9C5bXzvb9+LLz51ZajXUgWHpvAPM7lb5wDXhe0+OAdOr0yY8SsfZte2sNx2x4ptePpyB09orGhPrO/i+Y1ebR/sMM5Uatg2bIthseVMxSUybchMTve7JI//fMOBa1uVe0/ynYFcTNQUVxlLbQfbBTq/pzZ344vWIAjhBZGifcf9tBrfM3otOh+//O9VQUNfsnR0crmFc9vCgbfZ9XOM39E0d4nxk9SzPOfi9HILj2gY/3Y/QBhxHJlvaqMm6Hes3tVMAoem8DtVpJ4ppHPKVk5AnGht1x4rr0cn9QDCy395jNgGL+DarUB0G13XrTz9/l1HfLAa9nTsgdNGKeP30igEU/a7DnIypqcr/JpzZbHlouuFxnM/YfzK5G5ycZILv1NOssZFz1AU6UIw7DAUndPyheTkcgsXtgYJM88x/mQoVKPxS8OUt55a0jJ+Gt46Mu9qpR51p/AkcWgKv1tB6qErd52Mn1wCVPgBwfrH2bs70Eg9AHBsvjnW9G4QRbkIXiD90NW1MUhOggSmt9hj2uhmFqkXSD2xq6fq70C+IMuSB90Zyg1KAkUx7PT1ejIRprmGmCLuJ4U/m8wJpO9bnc1KHUMHxmD88euhxwPAqaUWvDDC5Y6Xy+IHZDun5OpRmrsAcOtJkYukkhcq/MLOmZd6iBjMNWaFf2SUNXcjKbK5TjvnC8T4l+XCP15Cp07qAYSlcxyNPwh57DXP/j7oQ1cX46cPiCsV/mlMgk4bHcnV4+mkHi9l0449vJ0TyDY5B0GIpobtA6mEY7J10l2Ea1tou3bCRulCsahp7tZ5se5rGDogNXeH1Pj7GunoZGzAePpSB14QmZu7obm5CwjGH0Q8N5eykTD+RjxxrGf8ow52FuHQFX7jrax0QajzhD231cPqnIu2dBVfLWD8D53dxlvec5+WeRO8INQW/nETOv0wAuf5Ak/sqC75hfRuek1Xq9Qjv6eFjL9pCztnVM0iKRcimUWaCAKQTuRu9/SMn84B17bQatgp4/fSBjTBnUJz1yT1jOrqGWgY/8mYnD0c6/PG5q6O8Vvp7/m2kyKz55HzWZ3/Sldm/PldwfQ7bs8Y/+goG+CSPyx1Sj1nN8UIuIzVAsZ/39NX8IlHLuL+Z8ydfS+MtEzu6EITV7r5FXJVQXc+fS97QvZ8/YDQpOCHWcbvTikCYNoos3N2BwEYE8XNjWWFKu+lfKH2lMKvTngTiPGbGryy/NZ27YRxUySz6uOXH1MH6K4z7+rJSz1RxPFNjcae/Xl5xk+W64fOicfmmrsaO2eQNHdTxn/9sXk0bAuPnMseQ4bxa6Se3kzqGR9lJ2Pmdq1GpkILWGQsz7lGV89uzPruK7B0mZjcsYUGOB/dKkosVI25rZvxp+xSfHiaUwr9mja6XpgUYt3vcjeOZGaMwTYs/dDBN2n8hl4QkPrztw1JsXLDveVaSVFKlrBopJ5aNf4hmrt//ehFfP+/vRfPb5iTSHWM/9hCE7bFEqaeZ/zmAS5Z6nFsEYqnevmvdD00HAtzDVvkG4XZiWPS+FvOrPCPjKTwG05GknoYq3eA64XNHq5ZaWW+thqvX9TdxlMDsMjLqwtpA8aPbSDdWXVI0Iekrg+2ulFsWos9pg1K3jQFsHW9IFluorMOmpAt/Fm9v1TqMTL+VMKQNX551oDgVth2Ny6Gae7STuCi+HNi/E2J8dsWw9piM7lboNkYQir15Bm/bWV/z7eezDt7NjpiapcxljyvTHD6viAGlpVvxo+LQ1P4dbdlMuhK3XLs3EadSWG772OnH+QY/+pcI17tltdXqQH41ec2jQ0ro50zjm0Y1ctPjL9nsJnVFSqVyAqOVPivRsY/CNFuiAA2rY/fE4wfkLLfK5ybMgPNSD0FjJ8Yu1HjD9K7sJabavw6qafKzMy4GKa5S8dYRB4GhkbqyeVWwrxXcnZOYvxyfzBm/Eqxvu3UItZ3BrgkueyudLyEnBGrl3X+nh/WIvMAh6jwl+Xx00nabtiVNx0Ni3ObeSsnkGqHWxq5hxj/IIjw9Rf0WSpFUg8wOuNPNP5c0ylKnrcOqBr/1drc7caF3RRq1pEmYlPiUv57KLJzNg2ywXzDgcWKNX7XZmAsW/h3JMspgd63Ot+z0uauvBM4fk1FEq68q0CGPMm70q6g8WvsnEA6wSv3GuTCr5s47mo2jE0Kh6bw6/Q4GUnhd+3aQtrU4S1CktCp0eI7gxBH45PjS09t5P49ikQMrN7VEzP+Eb389DtRpZ4qGv/uIMCvf+zhkbYKeZJ1ELi6m7sih8dc+Inx6SKATTBKPUGovTMERG7PYss1+vhF4RePVaWe+YadkSOmYuc0ST1unjnvlOw1BsQ5zVh+uI0snQtNJ/cZ00U2+JrmLgDceko4ex6WdP6Nrp/IR3RB7iuMvzVj/OPBthgYK2juxh+ollvdNjcsyMN/TY7xFxR+L8A1q23ctLaALz2Vz6vxFFkk83PbLmyLjRzUFhg0/qTwF3yQPvPYOv7jvU/iK8/mL1ZlSF9TOrkbRHykqN1p4Pc++Rj+4mvnhn5c1xO38q6tz+HpeGnqpavRk02QJ3er2jmBOLbB2NzlaeFvZF09ckAbIDH+Kdg51cLfcvLN3UpSTxCh5di54TY1TFGGjvGHidSjhiY2cXyxmdH5r3Q8HIl/rk6i6s8Y//hgjBXuApWlHqAeS+fZzR4cS2zmkZFEM+ukntjZcfeZI7j/mY3ccRWN4VsWw9H5hjEStgxmxh/bOQu80hfj5zQxyMLnlTZFAdNLexwV/+kLz+KjIxd+B46l72F0ByHm4sKfRgCX/w78KNJKHoMCOycgGrwmqceTGH/LtRLGrwa0AZLGvwdZPY5twbZY5nWTM67o/On7YaaxSzgRSz3q1C6gb7jr7JyEW08uJg6hIIyw1ZMZf/79ImJQBw5N4QdEw8WcRSLePLrC1jG9e3azh5PLreRDTChK6OzEzo5XXL+KnX6Q8yN7Bm2ScHyxmewNHRZp4U9/Z2HEkw9Q0QeJLjYmBlnleSmrp7nPC3/HCwoH7EzoekLKaTj6rP1daZdtmTlBhh/ypFkrx2oUNXeBuPAbmrt+EKERH0PLtRM75e4gyAS0Aen7Vuf7ZZJmgPz6RSIfZVKPzjZJMzdaxq+TejRZPYTbTy/hm+d38PSlDjbjz0Wq8ZMbKT3u3hjrWMtwuAp/QeYLvXmtOgv/Vj+n7wPActuc0Cl0XgevuP4oAOTkniKpB4gL/4iMP23upiej2iw0gSJ+dU6lMshTooDE+CsyyCfXd/HB+54b+nlHAeccXS8cKTNd3qerOy/pjgCQXD2VpJ4oYeGZRm+BnRMoTuj0wygJX2u7dqKxq1n8wHRcPb1YBtHlDqlJl6Txl0k9OsZPUs/wjD//s978HWfQcm380v/3QOLuoZ+rmzjW7RuYFA5V4XesIqlHYfw1nLRnN3s5fR8Qt/FLLUc7vdvxQsw3bVyz0sY1K23c93RWMzdlrBOOL4xR+OPfiWznlJltkZ2T7jJGYfyqj39Yl8h/vv85/LMPPYi//mb9axu9MEIY8WSaeRj0PNEYdTQSJOc8vtvLSj1ViqkfRonLRi4kut3MMhZbbiWNv+Xa8EMOP4zi5q7K+Ot39ei2bxHUxeW7iatneMa/tiRkWXV4CxBSqsX0w586xn96pY1f+6Fvxf3PbOD//stvApAYv0bqmdk5J4RGwRJoulLXxfjDiOP8Vj/J4VexOt/Q5vV0pQ/W3WdW8aWnr2S3Kil6uIq1pSYu7Q6GboxyzhOHgiz19JVmoQkXd4R1tSjj3QRdVo/4erVCQh/6X/vIN2pfYE3rE3tDMn7OObrxB7uh2a7V88Ui9vn4g09DUZUYf+zyathWrrlrsnMCpPGbs3pkVw8giuWOprk7jUUsPS8ysuGWa2WmzVONv9jOqfYLAHERedurb8IPvvS09nGObWXs3+nkrr60/tjLr8H33X4Cf/WwICVHFFdPzs45K/zjw7EtI5MPcox/sift+s4AQcS1Ug8gGryqxh9FcXGIWd/d1x/B+s4Az13pJd9ThfEHEU80xaoIIw66vshSj/znosKfavwjNHfHlHr8MILFgKcvd/HOe58c+vmHAeXtdIfU+Pu+CMCbazox48++tl1lInYoO2c80C/AD5AAACAASURBVKdOPA8MYX6EpbaD3UGgvbj4oaTxN1LrobpvF5iOj78f6JuxQJ7x71Rw9YgpWX2R/aXvfzHuPnNE+2+OxTJDdWFBcxcQJpN/9WMvSQp+MsDl5hl/35tp/BOBa5iQBFItlK6wk9YnXzB4+Amrc25upFxEIqesj6Kc5UlcLxQFx6zxizuMYeUeucD0jVKP/ncURTyxkI7G+NMpUWD47Bc/jHBiqYXXv+Qkfu9Tj+O5K+aMlnFBv49hm7tJqmVs51TPy26y4IQYPzUSq/n4XYflmpymCW8CxTbsali/6uMHxGvuaAq/bTHYJfsDPv3YemaKdVgMDNIMgEzEcRjx5KJcLPXoNf4yiLiNvNSj2jllHFto4nf+hzvxfbefwLF41kY3f3AgpR7G2HWMsU8xxh5ijH2DMfYL8dePMMY+zhh7LP7/al3HoMI0KANIkQ1uPXbOswYPP2Gl7eYY/26ybFt8sOiipCu+pg80WUfVwv8P/98v4w8/bWbD8u/J1Nw1FWI5EXQUjZ9+rjy5C1R3iZAe/X++4XYwMPzLjz409DFUBWXmq7EWZaDC3jZM7u4qE7G6pR8mUJGWm5xBGCHiZoIAFCd0+gHP2DkBYLPnIYh4TuoBiGTpj7Xvh3jzu7+E93/x2dLXYkLPN8sg8uum36N4DSXN3RHC0BzbysYyU1aPgfET/vYtx/GOv39X8r6qPn4viBBE/EA2dwMAv8g5vx3AKwH8PGPsdgC/DOATnPObAXwi/vtUUFj4o3RyV/77pECFn1wCKlbmGtjsZD9wVBzI0pcwLY3cUuTqAYD13X7ytSji+OQ3L+LB5/UREECWWWY0fqVZqANdZEQEwOhST2NEqceL4wVOr7Tx5u88g48/dGEk100VdOPCIjT56mSh68uMP2/nJJaaRDaUTJ7L8EMOx8pKPfReFfv4xXPpwsw8xdUDIGHsKuMHxGet6PyIeHrRHAV9X6/JA8hEHMuFv4g4DPzQ+POK4FiscnO3CPS+0OerziUsQI2Fn3N+jnP+lfjPOwAeBnANgB8B8N74294L4I11HYOKIqkncfU0qk9IqggLJn7Pbvaw1HKw2Mq7AwBh6dxR9FWSA8jSR7d9sp5cxccPZBn/+u4AXhAVNj7lhpXJ1WMqxDS8dd2RuZF9/I7FkhiAYSMA/CCVJU7GroxRfPZVQO8F5/lMoyJ0EsZva+2cauol2QMrNXfDCA2HxQWQ1mQWEwSghPFLGn9S+HfEHaqu8BflK1Hjf5weQK9gqrXlWmmWkPRaCgt/EI1UZNWh0LLmrglqc7fOJSzAlDR+xtgZAC8D8EUAJzjnNOZ4HsCJaRwDAG0TjZDz8Y/Q3H3Dv/s0fu+Tj2v/7YVNvYefsKxZfZewvrjwt3SMv8THP98Qy9zJVw8g0byLPnh+qNf4+xWkHrrI3Hh8YWRXj/zBaQzZLPSlQaVWMhhTT6NRXqYyzF1FV9pcpZsvSS/6o8Uyu7YVa91Zxl9U+IsSOmWNn/To9QLGX7Qnmc5F6k+NAtHcLbBzEuOX7jjLm7vDl0PbYklDFxBKAWPIDWmWQQTgpedpnft2gSkUfsbYAoAPAfjHnPPMJgIu6LH2TGaMvZUxdj9j7P719fWJHEujoPD7SuGXv49zjj/67FPJxhwdNrseHjm/g2+c3db+u24Bi4yEbUkfulTjF8c0p9H4Vc+7CsZYbnr3+Q0hOxUxINn9JFs4idXONWx4hjsGYnQ3Hp/H7iAY2krqBVFmkcXwrh4uFanhGsPDQr77Gkbnlz/YYqJcieKg2GFHCWmr5OMXr19m3emdYbGdEzAxfp5r7tIF3iT1mO6u6Y5wHMY/8CNzc1dqau/IGn8NjN+xGfzMABcvbOyawBjL9CZM6aOTQq2FnzHmQhT993HO/zT+8gXG2Kn4308B0E7ZcM7fwTm/i3N+1/HjxydyPI7NjIxJHeCSb6mfudzFv/izh/C7n3jM+LMfuygWKV/Y6Wv//exWz+jhB1J9Vf7QpRq/0twdQuMH8tO7QzN+TXFbbrvGx6/vDLDQdHBiqQXOgd0h9XURLZCe8O6QzV3S+AG9P3qSyBT+IeSkriTj6WzGSUywS0NswzJ+lmH89PorST06jV+Sz+g8TDR+Q3PX9H4RMRjnYiyauwaN303318pZUXJ4nYpRGb9q5wzCyGjlLIOwoWbNAgdO42dilvpdAB7mnP+29E8fAfDm+M9vBvDhuo5BhVugO6rN3Uzhiz80H7z/OW1mPgA8ekFk6MiSCqEzCLDZ9Ssxfrmxpt7uE8PJaPwVbuHV6d3n4hV0RR88ubmdkXfiE3Kp5Roff3FngOOLTWmB93Byj5wLA+inGgsfL8sSIy7froruQJZ6hmf81NxVh4voboru5OxhIhsSV4+k8ZfcGQLAYtMBY/pgPeobAKmrhwq/GtIGFH/WEqlnjMJvmrQF9FJPo6DZHITCQTMS47eyrh4/5EPLPASZ8ZO8OtfI/24ngToZ/3cBeBOA1zDGvhr/93oA/xrA6xhjjwH43vjvU4FbhfE38q4eT9Ld3velZ7SPf+yCYPwXd/q5Bu+5rWIrJ5Bq/HKR7CqWPstiIgvdy+uWpYx/V2b8vcxjdQgkd4JugKuM8R9fbEp7XIdj/HIuDDBCczdM/eppznk9jF92pgxV+OO7ubmmE5+XxYzfGSqyIZV6BsrSnCJWa1kMC019Xo/Ox0+zGmpIGz2PUeMfU+rhnBfbOSUf/+4gDUMzFX5T0mcVqDUliKKhG7uElmvnNP66pJ56LicAOOefAWC69L22ructQjUff97VQyfoQtPBez/3NH72nhtyhfaxi4Lx+yHHRtdPJvIA0dgFzMNbgN5RQUWFNH5AXJh6GV99OZNbW2xis+tjEIjpxOc3y6Ue+pAstpysq8cPYTFxTFcMPY9LOwPcdnqpdI9r0XOP1dzVeM7r0vjli/AwFxe6m2u7ejun2rtxhohsoN9f001ZbpXCD5gTOtWsHkDS+LVST4XCP+KQpBeKqeeirB4xGc2x2w/AmEjXNPn46X0bxcdvK3bOIORDWznT404vWL3E1VMPNz9kk7uWedl6POZPmTe6vaVv+o5vwYXtAf7sgbO5xz96YTfR6S9sZ3V+0+YtGUsaR0VnEAitVjoh266ttXMWFX6ydF7e9RCEEc5ulmus9PoXWk7Ox992bXGSGuSTizsDrC02CzXjIngBz7yeUXz89Bjd8u1JYlTG3/OEb9y2GByNHj4IovjfSOqJGX+Fwh+Qxi+9R4MKd4aAICC6C7WnyGeMCVnSYnpWKj5r+mNdH1Pjp/PRdBGTY7y3+wEWGg6aBXubx2H8qlNQdaQNA7k3QX219gGUevYdXKUDL8OPIji2pV2gPIj//L23reGWEwt456efzMg5m10P6zsDfOeNxwDoC7/FgBPKAhYZC02x81TW+OVYXoLYfpTV+F2bZVbfqZC9/Oe2+ggjngvwUkGvf7Hp5qSelmuj4dj65SFegN1BkNX4hxziUqWeYbNf5EakbpfpJNH1gqTQDGPn7HhB8t42NM1dL8zGK7jJsvXi30EYcUQcOY2/iiQIiDs89ULNOc/4+Bljib4+33S00ciuYyWfGxlBGOFyfKc4euEv9rjLPaHdQYDFlnmvsfzzRprctVjmLiyMxmvuUj+tK90R1oFDVviLpR7XYlr3hGyF+9l7bsAj53fwuSfSXPxHY33/u24WhV9t8L6w2cPJpZY2o5vAGMuxLdpnKmOukWf8RWwfyBZ+snKeOTZntGMCKbNcbDnZrJ648DcNe3Dp9n9tsSVp/EM2d8Nsc5fek6E0fmWJS13N3c4gTPJWhrJzDkLJo28h4lkZZ6BshLIrunrkgDv95G5xIdEldFJgn8xkqejqrJxAbJ3WnB+Xdr0k/G9UjZ/OR3NWT5p7Q6shi+5A6A5iZMYvZ/VEY0o9yeRu1mwyaRyqwu9Y+k1HANmwrMQ9ke3Up2zph+88jcWmg498NZV7SN+/5yYz4z9VIPMQlpQ8dLGMO/vBarl2zsdfxuKo8F/cGSSOnhuOLVTy8S+2HPghT4rSIB6VbygBYAQq/McXm0lRGFrjD7IaP2NMPN8QhT+VJeqVenpeiGMLjeTPVdH1wqRpTxur5AubyvirDnClS2yyvvAqdk5Av3c3CR6THksFyVj4HX1WD1k5BXEY7S6sV8L4W056l0fpoUXnD/1uTANhRXDVAa4wSpbmDIuMjz+5C5lp/GPDdczeYj8S2pwuE0XW0Vuujb/94uP4xCMXkmL42IVdzDdsnDk6hyPzjZyX/5nLXXzLkbnS4xMbkGSNP8xZ5eaU5m6Vwn90XmL8V7qwGHDm2Dy8IDJGTNCHlj7YxLISqccgFV1MGH8Tjm0Jl8gIrh71NTUL7IH5x+ebu/W5eoKkkT+Mxt/xgqRwkYwjF8qBkhZZdQOXHBnQdGwEkbhoV5V6dHt31Y1oQCqh6Rq79L3awh/fDV+72h65uVvG0KmA9/0IO30fiy23sL9X1jMogra5O46PX4psaLt2oYQ7Dg5X4bfMefx+PC2axN9qpB5iPK+7/QQu7Xr46nObAISH/6YTi2CMYW2xiQuS1NPzQpzb6uPMsfnS41tuu4rGr5d6eoqPv+zD3HAsrM65WN/t47mNHk4ttzHfsBFxM4MklkfZQj1psKSK1EN3GUst8zo/EzxNg0zNli9+/PQYf9cLsdByMwvIq6AXb1YDpH26UgEZGBh/mdylSj2AOH+rmAAA0dxVp63VmGygnPGbfPxEDK5dnatP6pEY/068GtJ0B0LfB4w2LKVe4PyIF0q6RZAH7roSMagDh6vw23ktlRBE4kqtG41PhqTiN/RVL16DYzF8/KELAITGf/PaAgDgxFILFyWp55krHQCoVPhzUs8g39xtaVw9ZR9mIJ3efX6ji2tX26VOGZK6KL9FZvzk6hG2uuzvcn1nANtiOBLvEl1qm9f5meCH+ddUNBAkQ21ENqQiUAfo4izcVsM0d0O03Vjq0WwYG/jZmGBa81fG+BOSEks9gHjt6lyACUstJzdtrS7GAapIPfmhNEBIPYwJh9uoF+NkqrVKc7cvlsEXDXCNy/jl9yQII7iT0PgLNoxNAoer8Gu0VIIfRnAtq7C5S0Vkue3i2284go8/dB4bHQ+Xdge45YQo/Crjf2pdFP7rj1Ys/P2sxj/f1DB+xcffqOBGoML/3JUerjsyV1r4fcnOCaQfjr6k8XOeX693caePYwuN5BZVJx2UQc3qAYpDv2SojUg7btjXxvgHYpBoruGg55mfYxCESgBfIC1ZyVs1dXdyYkq0WnO34ViZnKIq8x6APraBmqJy4W+VFX6T1LMzwJG5BuYb9siMf1DK+KXmbqzxF0k94zB+R0n8JQI5CsQAl0SwZox/MtBpqQTS5myNxp+wJemD+LrbTuCJ9U7C+m8+sQhAMP713UHCAp66TIy/msa/pTB+VeNvq83dClIPIGIbXtjs4cJOH9eutkslkECRemTG33Tt9MKh/C5pald+TZPQ+BsFPuzsY/ONSHUV36SQLERvOPFgnfl1/vrHHsHf+w+fT/4uW3UTxh/IjD+fHSP05OLXQRcGx7Iyg2+VpR7NPImnkXqoSOriGuh7tVLPtjg/qkh3H/ry8/jY187lvl7a3HVTe23XC4XGb7gDAaRAvJGau1ae8Y8q9TiK1DNj/JNBagvU7xR1bSv9EGoiG+QPzffeLtKk/8PfPAEAktTTRBjxZD3i05c6OLbQNObwy1huu+j7aU6+TuNvN8QkLWmwXhCiWVHqubA9AOfAdavljD+RejTNXSH12NrHi+GtNIxuFMbvaz48RfnuMnSNSHX59qQwCMRWq7mmnbPZqnhhs4dHzu8k087dQZDaOZO+kuLqURl/QeRI8jhZ6kkiqcPEJVTWLNRNW9Nxyec/Fd3FIZu76zt9rC210HDE3UtRcus7P/0k/kSzpau0uRufm5fjSImFViz1GM4BOjdGknqUuI0gGierx864emaFf0JIPmA6qSduylATTU7c88LsYhBANKduO7WEJy91MN+wkxyetSVR9Mi98PSlLq6vwPaB9DZ7py+aa9oBLiVfvoqrB0CmGGekHkMmOhWQRVXqCVKpRxxH9vHrOwMcX5AZ//Aav2rnBBDbR6swfipScshbPYw/iVZ27VzvJf+9gkE/8PwmOOfo+mFyUW9oCImq8QP5YSEdEj3esTJbnaqeJ5WlnvhnmRi/qbDTVLfpjlHGZtfXXuzLffziZ1+KyddibOc0xUSPx/izUg9tPxsFTUfcPfhhhJ4fzaSeSaFodys1ZXSj8aYPzeti1k+OHkBIPUDq5X/qcgdnKuj7ADJpll2fVu/lNX4gLSSVpR5Jfrl2tZ38LoxST5SVehJXjydSEXX5OWHEcWl3gLUlqfC3HOwMmcmve01VGb+uEakuHZ8UOtJO5Dllojr/veLfHnxuK86RSXcpOxoJ0gujHAN1CjLuCfTvDdnVE4q7yEqFXzNt7UkXE0LZAJfuzjmKONbjwi/r8CZs9jyt975M6qHfm8z4y3YAy48bBrZW6hmR8Us9md5M6pkcdLY5Amn8jLGYWWWlHm3hv00U/ltimQcQUg8AXNgeYHcQYH1nUMnRAyCZdN3q+Wkyp6rxK5n8w7h6ACEBnFhqlS4oobsi2cfPOUc/EE0nerxcjK90PEQcisbvDp3J74c895qqNndTqSN9fNW7hWFB78F8wymVemTGr8ZtU0HNMP4gr/Gr56UO9L45luTqIcZf4Tyhc1BuRBfZOU1Sj44YbHTFcnaZ8Q8Md5x9P0Tfj7QN2fKsnuy+gMWWg4ZtG6WlfhDGJoDhy6F6QQnGsXMmF8OwMH10EjhUhV9nmyP4UpyqY6sLlPUfmjuuWcLfffm1+OE7TydfO7bQBGOC8T99KXb0VCz8STRzP0jCv+YNUg81eKvewlMxvmalDdtiSV+gzNUj2znlVETdHUPi4ZelniEz+cN44Ejr46/U3M0PKjWluNtJImH8DRtt1ymc3CXG/8Bzm2kkMzV3NR593fuqDgvp4GWkHknjD6JSKyeQXujl5m4qn2V/p0D+/CTo+mnJcN9Sq/T824z3Xuje8368GF2XEQSk2n/C+JtO4ujT/TyxzWu0Upizc0bj2TkBYvyjbQSritpimfcj0sJvbu4C4tZb/h5TcWWM4bd+4qW55zg638TFnT6eJkfPCFKPXFRkqAvXKxf+uBhfF08QV2nu2hZLTr6+H6LvpUyr4eQLP43jZ6QeOZN/tfQwJY1asXNWbe4a9OjBEMNVVSGvT2w3rEIfP4W5Xe54ydKeXHNXOedURuvaVeycqdQj/6xBRcafTFtrGb/Gx29i/E5+hak81a2uhFSx2fOM/07T4yboGX9K+tTHFu3vLYMTvyecczDGxpvcle7C+35Y275d4NAxfrOPX87RduzsLfWgoo5OOLEkHDTE+KtYOYHsFq6OSepRFq5X1fhX5lw0bAvXroomdBUfv2szKfIgStwP7YatfTw5Vo7Maxh/RWePOixHqDq5mzJ+qbnr2pm9wZNCskWr6Qgff5HG74W4+8wRAEgC/hKpR3NeDgyMv3Jz17YyDLIqQQDyCZ26i2mbIhsK7JzisVLhj/tea4ut0ubuRsfPPZ5QtH2LnpuxdFHMQtMtPN/HYfzE7umCLDZwjSf19P1w/9g5GWPVqtc+RqHUk2H8LN/cHUK3O7HUwoXtPp661MWJpWbl9WnL0jIWuajISDR+L12pV+XYGGP4N3/vb+Et99wAoDzKgAbaMoxfclMkdk7pd0nylFwMhs3kJ01XLVJVJ3eNzd1aGH8cnRtP7vb9SKsh+6EovC9/0QoajoXPPykKP723uvNSMP68q6dqZIMj7XEYBJG2WWyCasHVXUzp2OmcVaEr7KnU0yxNTd2KGb/u9dIQoQm0uPxKJ2X8RXf7/SAamfFTaipdkMUGrlEHuMQx7g4CRNzcvJ4ESisSY+w7AfwhgAUAL2KMvRTAz3HO/9fajqomOAU+fnnizlEyfYZhS4Bg/A8+v4WWW93RA8QSim1huxckDcC8jz/f3K36gf6RO69J/lxm56RbVgqu68UNJwBJVg+ATEFNG9LpMQ+byZ8MYGk1/mrbp9THm3KFxgXp9tTcBcT7ol6s6SK+1HZxx+klfOVZkfFEjE4tSmHEEURc6+MvZ/x5Vw9FNlQ9h9WhO93F9O+85BRsiyXSoQrdxWx9Z4DFlhPvcyhm/InGr3nfqnweaQsXY+LOqoj0jbpoHcgOhbZcO1YOxmP89Nr3mvH/DoDvB3AZADjnDwD4ntqOqEY0Ct78QAoGUwdlhmX8a4stXO4M8MT6buXGLkCZ/EJflXeyypiL8126nnDZVJV6VFTR+El7bsVsltwU7Yal/eB2BmLNnXwbPmwmv67IANVjfHUXDnkwZpIgxk8DXOJr+WOk75tvOnjpdSvJ1+kCkbjNoqzunZd6qkc20M5d+nlVoz0AM+OXf6cLTQc/9vJrjT8j+awFcnO3j7XYZFC2TnMjLn46O2eVc74lSVEU6w3o73AHweiNVLJ/h4nUM4adMz7Gja6429lzVw/n/DnlS/UkXtUM3YQkwZPeMEdxTwxbXE8stcC5uHJXtXISlloioXN3oGf8rXgHZ88P46ZS+Ri+DmUfPD9eTAPEhT9QpZ784zteiDklSnbYTH76YOqyeiq5ejRT1i23Hh+/OsAF6OOfO4O0CfzSa6XCnwxwZX+XA8MkqWvll7KrkK2XcrNwGPKiLgTyDHdhRXB1Us92OtVdRjw2JalHDQIcBGHpayH2TJPnjYL+3liMX1ERwjGyelLGHxf+PWb8z8VyD2eMuYyxXwLwcG1HVCPShpPOx58uUBCd+vGkHsIwjB9IJ10TNqn0B+jvPS+onLGuQ5mPX15/2HIt9L1U6pGzeuTHi+Cx7PEOm8lPH0z1gzj0AJfc3I1v+yeNjheg4VhwbCt5XwoZfyPL+NuJqyfbIDQz/nI7JxUgR2L8wscfVrJzAnGUtiz1VMz5kaFt7u6kw31FDBwAtmLGzzXR4br+hwo6f2gAsUjqGYfxE5kMY2ePiGwYUeqJ359E6tljxv8PAfw8gGsAvADgzvjvBw6NAsYv27ByjH+E5i5hpMIf+/hlnZaQ+vijsQp/6eSu5HIixk96ftvVT+7uakLlgOEy+U1SD0Vql4WUmZaG1MH4e14au6BOVMtIGH9TLOuhIDQ1pI1eWxoKqDR3K2T1yJ57sVGOwQuFxl8l0wlIGX8kyRdA3mJbBFVW5ZxnpB45MloHkjuA/F1BJY1fWRRT7OoZnfHb0gwGvTfj+vg39oPGzzm/xDn/ac75Cc75Guf8f+KcXy573H6EU3DVzw1wlUTkFoFYDWPAiyps3pKx1HKwE0/uzjXzb7xtCb2y6wep9XGMwm+WetLfR1vR+OWsHvnxcvBY5jUNkddjKvz0fMNEFhCacU5LWWN0WMj7ElS3lQyZ8TPG8NLrVtB0rKRoUIOQJJWB4YJeKZZZkcoo432Yu9ZlZdra9J4UIX2/xGN3BgH6fpRIPaaQPwKxXt33VGlU089fUJxT2gGuMRi/HOMeSHdbo4COmRxNe+3q+SMAubONc/4ztRxRjUj0OK3UwyWN38o3d4corkfnm7AthpNLraFPqKU2afyhcSqStnBVjdrVwYpz6k26uexyarkWep7q6kmnQgkUUZx7TUMkdKYav77we0FxeJXezpkWmUl+mLpeeqFT5ytkJFPY8YX87778WhyN1zUCcpRIscZfJbLBD4WThS4qFPU7DHlJ5km6PpZabnJBGmaJuCqtbCQzHuJ1V3X1yD+DMFRzt6VaZjV2zngSeBTYyUrMKMklGre5Ow1XTxWD+UelP7cA/CiAs4bv3dfQBUcBSLS5RONXcs+H8UAD4kN3bKFReXBLxnJ8m90Z5JewECiTv+oCbROKUit9qefRcm3sDoKkcdl2ba2G2/XCTEEjLLUdnN3s576uf96YsasafybbxRxxrcuVkWWFyRb+MHFdFbp6Btl+zRtfdg3e+LLUWqsWybE0/ojDtdI4Axp8G8b2SxbcrZ6P65BGlpgiEnRwFSmRitnKnPjZZXecmz1PfA4jnpMjvQqyFV3sSVbTmREIork7IuO30uZuOMIFUsY0Nf7Sws85/5D8d8bY+wF8puxxjLF3A/hBABc553fEX/s1AP8AwHr8bf875/xjQx7zyNAtvABkC2A6uTuOxg8Ab3vNzTglaf1VsdRy4YccVzqecfCr3bDR9aV1eiMWfuGU0Wuscj5R07FxadeTpB47scjJdrvdQaD1dS+1XDzS36l0TKZGYkNzodFB3Y9MxwtMfu9u1wswF//sIqnHlLtEUJ0hpve1UmRDECkXPXtoH788SKj7mVXQUBg2afYrcwrjL5B6ji82cW6rn7srqCb1pHZOoEpzd7TPkNzcJUI5rtSzX1w9Km4GsFbh+94D4Ac0X/8dzvmd8X9TK/pA3j1BCJQ3zB3T1QMAb3rltyTLWoYB+d7PbvUKGX9flnpGLfwFThm52d2OI4cTV49jJf/Pavxhzn4qXtMIGr+a1VNR49dFPqS59JNt8IoNadTcJVdPvrlLjN/E4CgRVmX8o2zgkt1Y9DO6Xogw4mjY1QrJsjJtrf7MKlA1ftoslzD+AlePuJuNkkawrrlbRnbSwp99PrXwc87HYvyO3Nwdk/HTmtDN3j5o7jLGdhhj2/R/AH8G4J+XPY5zfi+AKxM4xonBlMfvK2+YbY3X3B0HdJt9YbtfqPF3Mxr/aCdIUf4NLaYBRMhZ3w8T94Ol6MeEjsbOKV5T9Ux+nSsHSF9jGePXLQ0ps66Oip4vNXfjD2nXoPHrHFoyZEPBOK4eT5kcbThWMhNS2c4pB+vFP3PYyGI1f4jki9WY8Tvx8njd+0ke/uNxIzin8Q/T3E00fnE86jkQRBwRN2/zKoMzweYuII6b5MI9tXNyfoBgRwAAIABJREFUzhc550vS/29R5Z8h8TbG2IOMsXczxox5jYyxtzLG7meM3b++vm76tqGQ2uYUxq+4Y+R0TlNMcF0gtuWH3LjdiNYvjuPqAfKFWwYtpgFocjfMpSLKdwyci41h2ubuEJn8poZ1mTRA8EORKiqvv6uy9GMUyH0YERMM9A2uHp1DS4acRWS6k6uygSsII2X7mJVINlXlymUpLBAwx5IXwVXeL5J6SHMnqVDX3KWLBLnjcox/iOYuxYqbpvb7kmFhFMg7usdt7gLpnQpjo0u4VWD8yYyxlxf9N+Lz/QGAGyFmAc4B+C3TN3LO38E5v4tzftfx48dHfLosbEuk9qlvfsr4SepJ3RPjyinDYkkKvTLFsrZjl83YUk8B41elHsrqkZlR07WladMIYcS1BW6YTP6y5q6pJ5E+Pq9Hl3nGR0XXC9F200LWNqxf7BQ4tAiyvGhy9dhWlQ1cqtRjYzfOSap6nsw3HFgsW/hH1fg9ifEvtpwMGzZJjXSR0Ek9RMTK7nIpdC2Z3CWpJ1AL/3h9soRMRrLUMw7jTy3UwzTTh0XR2WgsyhD2ztcM+2Sc8wv0Z8bYO5F1DE0Frp1nGXKiIZB1T4xjmRwFS1K+uYnxUwTwuMdWFIMg+/iF1BPvAVUYPxWpJEZay/irZ/KbB7j0t+oqvDC/r7eO5i7nPJa20t/HXNx0V9H1zA4tgmuzRKYyXdBlQmKCr8gyTdfCzpCF37JYJrbB1/xOy5AaKcRr2ux6icxDaBgylGhqlzz/soGgKtlJNH7FzplvFKfT6KOA5OEg4lKvcAzG72YHAuuCsfBzzl896SdjjJ3inJ+L//qjAL4+6ecog6uxxCUTd/Eb5tpWcttGq+HqvO2SITN+E0ukxd7jSj2icJs0/vTDTifjds/PSj3SHYMpRhoYLpPf9MGm338VxqteCOto7g6CeG+u9B61Yputio4XlkZzO5Z0zhk0frX3pIMXRpnmYsNOC/8w5/ByPE8CiIiTYXVrktsSjb/nJ41dgmkX8kZS+AXjl1n60IU/5+rJ/v7GZfzE7oMw1fgnIfXUuX0LqLiBizF2B4DbIXz8AADO+R+XPOb9AF4F4Bhj7HkAbwfwKsbYnRB3DE8D+LmRjnoMuE5+d2u6p1T80m2LJZ7cqUs9LanwG1iiGOAKxrZzNl07+XCrkCMbiOVvdr0MM5LvGEwx0kB2wUwZ0uauuoGrenM3l+xZQ3NXtyGNButUdAtmMgiurbnL1Nk5Sy58gaJ/N107+Z0Oc55QWCBAF9Phi5k8ILjZ9RMrZ3JsBqmRmruJxp9ZUFONiCUhbUlWj94OTD9v9KyedPiOGP+oWT1AVuqpE1Umd98OUcBvB/AxAH8HwsdfWPg55z+l+fK7hj/EyUJdqwjknSSunS5imXbhbziWGNDyzSyR/r1OO6cfSq4eKvw9H6eW09kEigMApEKoYfxqs7AIidSjfHiGae6qVlDdlPG4kNcuEtoNx+jqWdUMtslw7ZSQmDX+ann86i4CwjDnybJkwR1F6gGyDevNrpeLLzH1mLa6PpqOlZAg+XtMcRYqSOKhc48xJs73XHM3nU0ZBbLUk8wDjWjnBNJztU5HD1DNx//jAF4L4Dzn/H8B8FIAy7UeVY1o2PktRuotmmNZUr42Zb/U+0bIIE3c6ONv2Ig4EpveqBp/Uca9vEmIGrobHS8v9RDjT5aS5I+ZbvG3utUKv2OxTLQzPRdQ3tzVafxl256q4on13UQu0klbc664E1PR9QLt70WGY6eExAtE7ILqBxeb4ar0OCSpR2n0VoUs9Yxa+JvS3bVO6jH1mDbifoDOe1/17uWHXnoa7/6f78LxxTQpV/RRFMbvjyflylHvE7FzutNh/FWOsM85jwAEjLElABcBXFfrUdUI17FyQzDqAJesTU6b8QMpSyny8QOp7W0sV4+puRtkQ9oAsUVLXrIi3zHIy0ZULDQd2BZLbuGLYPJom27Vc8et0fh1zV3OOc5u9kqPhxCEEX7w330G77j3SQCptNXOMH6zq0d3JyRDJiSDeEBJdXU4lgXOUTgPoRbpURn/UtvBluzjH+Eco7uYMOLY6uWlHtMdp5CF3LQhq9P4S4rrQtPBa27NDlA2NDIvnRNjM/5Qntw9wIyfMfb/MMbuAfAlxtgKgHcC+DKArwD4fK1HVSPEhGT2g6PeomX01nC8PJxRQLe4Ju83FWJiZGP5+E1ZPZmQtmxxSx4v2Tl3pTWEKhhjWGm7meAtE1SpgpAy/uGkDkBv5/z8k5fxXb/xSTzw3GbpMQFiMKvnh/jvz26Iv2teL9lec4+tyPgDqbmrK2ymyXMZQU7qyV6oqyLj6glG1fhFYd/p++AcWGnnGb+u70KFXzfdOw4R0zn6+mMz/vQ9CZM6sv8ZfxENeRTAvwFwGkAHwPsBvA7AEuf8wVqPqkbIWiohkXpogMtOpZ5BRYYxSSyVMH4qvls9TysJVEUR4xcDXOTqSV+77OOXXUHyGkIdZOmgCDqpBgCaFZu7niZXJnX1pI89u9kH58CHv3o2sxzFBBrM+vrZbQDS65WbuxpXTxSJwbYyV49s5xwE+uXfybBQFKFh4Gyq574xIuNfbrvwggh9PxxD4xckSw1oIzSd1HEkY7Pn4YZjC2mwmkbqGbnwK8m84zN+cvVEk7FzUnN3rxg/5/x3OeffAbFf9zKAdwP4rwB+lDF2c61HVSN0hT/x8VvpIhZ/jwa4gNTLX5TVAwjGP2xqogzTrXYUj7EnA1zShyLDICXG1ilg/ACwPFet8JvYZdXmru7CQQtJZMZPjcuPfe1cpSgJknDWdwa4uN3XNnfnNFIP3QGUu3qszDmnZfxWOeNXX7/MZId19QDi9zRq4W84wlFEA1l5H381qUeOUR8k9svhC2NTI/Wkk7uj2jnzzd1RiZg4xmzMd10ofbWc82c457/BOX8ZgJ8C8EYAj9R6VDWCWIgMdWiItNQw4lMf4AIkjb9ggAsQH5Bx5gtMH7x09DzvKc5IPVJzuOsFsJj5A1RV6jGN4w/j6tE9XpW1iGme3+7jK7F8UwRZwvnG2e3Uviq9Ry2NnbNjWKGpQnX16HJ1ZD3ZhElp/LITyyS/laFhM3hBlISOLeeau3bujpNznlg/aRZAbuiPI726GqKTSj3j2jl5rlc4CqZl5yw9QsaYwxj7IcbY+wD8BYBvAvixWo+qRjg6qSeibjzL/D+IorGHpEYBST3mWGZxLFs9H40RT1hAnOxBlN9Mpbqc5MIvN3flrJ/dQZBsmNJhZa5RqblrYpc2hXpViGzQXaRbbnZKdKfvJ8FpH33wXO77VWQL/1ZS4LNSjwMvjDLmgaQXUML45XWfRsYvOUhMkBcKAeM0d9PCLy7Go2n8fhglMcM5xq8pxJRBJef2j9Lc1R6Pk3f0pVLPmANck2L88XHUPblb1Nx9XZyp/zxEhv6fA7iRc/6TnPMP13pUNaJRKPUQ40+v4nsh9Xz79UfxPbccT3JGVFA+zNYEGD+QZ9Hq70NmHxmNP+4RcM7RHYSFQWTLVRm/ZgBLfr7SyV3D49Up0Z1+gJU5F69+8fFKco/M5L/+wna6R1e6ONOHVb5IDMP4vYTx6zV+Oi+LvPx5xq+X6cogZ/KP4+MXhT/W+LXN3eyFXP1e9a60qo9fhyIf//iMP5II0ziMPw7920Mf/68A+ByA2zjnP8w5/xPOeafWo5kCXE20bZB49dPmLiCu4ntR+O+5+Rj++GdekfOyE0hu2RkEYx2XufCrjD99DlnqadhCEgsiboxkJqzMudjpB0PnyWeOt2DgrOzxzThviLAz8LHUdvGGv3UaF3cGuO/p4gRxKvwvOjKHr5/dQtcLMntzAf0ylm7JEhaCOrmr2zBlV5B6vEAp/HKo3qhSTzCqxm/Bi5u7jGXjSOh4VFePurBFXBCzS5GGfS0EndQzCEK4djbNdRjIGn84geZua699/Jzz13DO/5BzXi6AHiA4mjdfDWlLGb8k9UxR4y/DnFJ8R0Vql8uyLlWrNEo9UhSCKZKZsJIwyOJo5qK1eqZQr8zjDUmStImKsNMPsNhy8Npb19ByLfz514rlHprIvevMKp7f6OHcVj93odOtX0wnmivYOWl2JIy0Gr8rERIT/CKpZxg7Zyu9qxxV46dCu9n1sNRyc8VVF9mwpTiA1O8ZR3rV+fj7fjQy2weEVVlkKEWSLXx8xr9nUs/VioadX1/nGzX+vWH8Zciw7jGOK/W36+2tdAHMOEPcfCHxggi7g6DwZCUGR3qvCbrIBfl4RxngouOWX+d2z8diy8V808Frbl3Dx752vlBCITvn3WeOAADue/pKjpXpFq5XZ/wpsx0EofY12InUU6DxR1l2nu6YyE9DFyHV+AOh8Y/AYhuxpq6b2qVjI6mQsKHu5lUsx4MxQtUadl4qHASjL1onUH8myeo5ACFt+6eaTQmONrIhmw/jSg2bUQKu6kbbnWzhN8VU089mjGlvQamx7AVRHD1sLm5V83qK9GQ59Mv8eD07bSmL5YnxA8AbXnIal3YH+OUPPWiMlaBifvcZkSstGL9S+IsYfwmDExJkaufUM/7sbl4VnHNNVo8d/3+488S1Lcw37ER6GUfj39AEtAFZqZBABgBqBAt5T3b1jOvjnyzjB5AshZ+MnXOfuHquNri2lcvrSDM20jx+8fVoT+ycZXBtKykCY0k9mpF4IL+YBkgZSCvj6U+lou4gLC78MYPbLCn8RWv+Gk7+vVNh0qMF45d8/P0gkTN+4I6T+LnvuQEf+srzeO1v/w3+XOPyoWJ+zcpcElSnNmzp71qNvySyQT4vTZO7lPpoujPRLbGh92iUQrnUdrG+OxDHNyrDDiJsdb1cY1c+JvlOjJq7RBSEEyev8Y/m6tFIPQbr7DAgmW4izV13/4S0XVWQkzcJqm/dkZiVF+hDw/YaVIAn0dxVpR615wGkDEQNaQPEh7FTEktAH/yyoDYvCI2vqWjSOHm8QSpSG4k7fT8ZUrIthl95/W34yNvuwcnlJn7+T76CLzx5OfN4edH8t54WGYUqi081/rSP0dFM+OrgKImwut+BvNhbB3UQEUjfo1HOk+W2i8tU+EcstF6B1NN08sRjs+uh7drp+a2xc1psNK+8ztUz8KNM32oUkGEkiES43qiNYiD9ncw0/glDO7mrLOim/9MA137S9wl0Yown9ehjENTFNEBa8HUy0yCIRBBZUXO3ssbPjWyuqqtHq/E7djKs4wURBkGUSD2EO65Zxh/89LcBAJ693M38W98P0XZtWBbDHdcsAcgz/pZO4x+EsC1WKrXINmMR0qaxc9rFdk4d40ylnuELyVLLxaVdLz6+ETT++P3a6OS3bwFZqZBAU7vp9+Sbu6OHEjKtq2dcxm8nGj8fq7ELAN/2Lat4452ncevJpbF+Thn2X0WrGbplFuqV2paY1TgnWp1oT5Dx5wq/4uMH5KZTvnE4IMZf4FwhWaVM6ina71q0IxgQBTHienYqM/6dOHxssZVnobSxqaPEK3e9ILn9Jsavvt45jZ2z44mmd1mshjotrntfbck6qEOyV2KCUs9YjN9m6AcRtvtBIt3I0J1/mz0/873qdK9puK3a8WjSOSfA+Gl7WhBGY1k5AeDYQhP/9idfVioNjov9V9FqhqNpEPph9krtKq6ecTS7utCO2abJ+lgFpox7X8McizT+7Z5IXyw6WR3bwmLLKR3iKm7uFks9VED0Gn9qBSVLqcr4gfQ1UFOW0PPSfcMp4zdJPVnGX+boAWR5MRIsVCv1xKYDQ3M3acprQtpGKZbLbTdx2Yzs449/56sGVw+QPf/U3bwU+0AYBOHI0+o6V09/Eq4eWyzI8UM+lswzTey/ilYzGrYmj19hmWkTLRqLYdSJSUg99LrUaOY0u6hY46fiRM6PsujhlQpBbYMCaa3Mzmla20iPpaUbRYy/4YjGeScXtpYy/pNLLdx2agk3ry1mvqetmdzd9YJSDz8gvReB8INrNX4pSkQHdeIaGJfxpxesUfP4CSZXD5BNTd3sqoxfsXPGuwpGgas5f/p+OBlXT2zn3I8kUYd67yf2IRzLQhTfUsu3znKzyLXS5u4gHP1EqxOTkHpI21RZtC5silhRNpZZHMOVjij8ZbEEK+1GJR+/UeMvae5S4dO9X7KPf6eA8QPidXRzjD9MfueMMfzFL3x3/vhsCxZTXD1xhlEZqKhTY1ir8ZdENiR3apm5i9HsnACyBXjEPP7kZ+mau5rzb0tpBOuauyMXfjudGyDpbRBEYzN+17aSDVzjWDmnif1X0WoGOT7UdW4yS6SCt5+bu8QuJ2HnzLt68n5krdTjKoy/hNlWyeQvmhIta+6qKasymvHUL+c8YfxLGsYPCJ2fFssQen5Y6q1mjImLRkbjDys5NOiYSWLSu3rE10w+fq3UY4/n6lGPbxjI56auuUsypfyebvf9TLSDqsuP83mkC4bcI5kE46fm7qgTznuBg3GUE4SbfHjSkykIo8ztsdzcNcX87jUmwviNzd28xt92bTgWy06F2lT4RSEta0gtz7mFzd0wTgot8vEXFv7A7KOWp5SLNH6AcvU1jL9CARdbuNLHlg22EYh40AVHq/GXuHp0Fz7XZmBsNMYvXxhH1fgJRT5+ek8HQYi+HyVGAPqeSbl6dOs7J8H4HduCH2f1jNvcnRb2X0WrGbrpxyDkmTcsae7GPv6rVuM3RTZEea28KXmr1cdvVJZ63EIfvzoxrELnypChc7UQ5L27JPWYGP9c09Fo/OWMH8gvY+kOJsf45Q1cOqTzF+ljGRNW0r1g/G4J41cLP70vOY1fkXrGcfUAWdLX98Ox4xEciyGMIrGu9IBIPYdP46egK+nN96Osd1zO2N6vUs8kBrhMPn6dO+YNLzmFtcVm5vvouUnjL5N6VmLGL2usmectaM7S81WRenR6tDxlTNu3FgyMf6Fp5109frUC3lbWL4rBtioavzi+3QFp/Pn31S119eTnLwBxZzaKnLGc8dOPovGLxzCmv7tSiQe9L0tK4R9MqLmbuojEz+Ocx5ENYzL+eI+3qhzsZxy6wp/EFKhST0bjzy5iMRWIvcTcJDR+k9SjhNYBIir6npuPZb6PPjBk0SwrcCvtBsKIY3cQaB01FFlQNLk7qNDcLZR6fMH45xu20Xo313BweTc7wNXzwkoZ6ccXmzi31U/+XrangEAXq05B4bcrSj3qObE852bkk6qYlNSz3Ha1k+8J8YjtnNT/kZ+XBtuILHhBNNJrAdLXQOd7ksM1LuO3Gfp+lFMO9jP2X0WrGdTclVmTr1yp1UUs+1nqGYet6FbbAfoBLh0SqSdp7pZr/IC4UGgLf0nWSTNu7pruGAoLfyL1hNjp65+fsNB08rtzvWpSz60nF/HHn38GQRjBtlh1xm9lpZ4iV49fIvWor/8P//7dODKfl1rKIEsuozBZ+tzoZB4gTzyo9yLbSOUgNzf29I88uZtIPeI864+R9CnDsSwEYSCknn1YK3So7SgZY+9mjF1kjH1d+toRxtjHGWOPxf9frev5TXA0zV01w/ywSD2A+DDkffx6yUD3WCBl/GVSyEpJQqeJsRJc5YOrwqvQ3O3HjF8uLirmGlmph3OObkWp59aTSxgEEZ6+3MEgiBDx8ix+IL27ouZuUVaPifHT61dZ54tPLuK4ItNVQaYAj7h6EYB2ahfIhwRu6xi/k2fpIw9waZrJwPgRyElWTxglVvD9jjor2nsA/IDytV8G8AnO+c0APhH/fapwdVJPFGWu1PI6taKBor0ENVLHPbamm/fGB5GZOctgjCXe+ip5NGlej77wk9Zr3MDl5C/aMtLmcJHGH2FnUMz455tOJrJB2ECrFYjbTomp3ofP7SQXjyqMn4pgkdRTZuek921Sd6ht104u/qOGtAHQBrQBGo2/r9f4AanwT7C5S4Rn3MIvZ/UcFKmntorGOb8XgLrP7kcAvDf+83sBvLGu5zdBduwQcoxfjmwoGCjaS9DCdRqiGhU6b3zi469wEpMXe75CHg0xP9PS9aLmLGDuSaiP1+bxZ6SewGjlBESh7vtRwqwp3K2K1HPj2jwci+Hhc9uJXFTlTkFt7hZN7poWsRS9/lHAGEvjkUfM6gHMUo+6CGi7l3db5Yr1GEQssXPGP6vvm62zw0DYOaM4AWD/1Qodpn2UJzjnFHZ+HsAJ0zcyxt7KGLufMXb/+vr6xA5AZ+lS82HSTJRo30o9tHB9bKlH45RJCkgFXZeev4pXnZhfmdRT5OMH8pPGhMKsHtnH3ytj/KJQE+sfpoA3HRs3Hl/AI+d3kscP4+Mv0vjLQtqSOYYJnq9UhEcpaPQ7ryz19H24NjMGAYr/63OMqkAlDp0h3tciCDunYPyzrJ4ScLFvzbjrjnP+Ds75XZzzu44fPz6x55Wz9gnqqLW6enF/Mv4JafxOfuF1EHJYDJV2ENDzV/nwJIzfIPWU+fhNi2MIXlFzNy6k5OopYvwko3VjvZ2yd6oux7jt1CIePreNzqB6YUl8/F7R5G7xsvXk9U+w+JDsMtLik5LmrmWxzFa17TiZU75zVLfEjRPZ0FBIn7rYfVSI5q6Y3D0ods5pH+UFxtgpAIj/f3HKz5978+nP2ayetLm7Xyd3bzmxgBuOz+OmtYWxfk5Ts8DcV3oexY+vzvhbro2WaxkZ/6CAsQPmgTNCsoFK6+qhxwqpxzS8BaSMn2QX8uVX1YJvPbWEc1t9nN3sxT+vCuMnjd8sP5Qx/mDCUg+Q3YQ1LOg4TBo/kJUatzXvi3yHzjkfc3I3+9mn3KhRHE/Zn8virB5zpPh+w7Qr2kcAvDn+85sBfHjKz5+8+fL0Y6AMcNEHzAsi8W/7sPCfWm7jk7/4Klyz0h7r5+iCz4KCZSi6xwPVGphAcVBbmZ1Td9HOPt58xyBHSHthfgmLDHotFNtAjL+qJEAN3i8/s5H5eUVwFKlH9xoYY8mUqA66kLZxsTSGxr8614DFgOuOmM9RWWrc6vlYVGQh+S4viDg4H715rUo9Vzp+fJzmC1MVZJu7+69W6FCbj58x9n4ArwJwjDH2PIC3A/jXAD7IGHsLgGcA/ERdz28CfcDI+gbkB7joqk3a7n4s/JOC8MarefzVM0dSjb9aUVyZc81ST8k+1erN3fyxE1tfjzdKFQ0Bkf2S2Dcx/qoLsG87KeKak8I/RCxz0eQuIM7fUqlngqxzObZ0jhJFcHK5hXv/2asLyYmQGsXvd7vn594X+T1P9u2Oyfi9+Pe32fVgMXN0xzA/1w8jEct8QDT+2go/5/ynDP/02rqeswr0Uk9Wm2NMDDYR49uPGv+k0HCs3LapYbRK0s7LcnoIy21zUFupxl+1uVvA+C/FG6XKBriAlH0Pq/EfX2zi6HwDD53bBlDtd5Mw/gKNH4j15LKQtgnqzDevLeLa1XapY8uEa1fnCv9dZvzbfR/XrLZz/w5MpvAnn/2E8XtYmWuMvU/bpububBHL/oWjkXp06/5E4Z+M3Ws/Q7fcZBitkj5MwzB+U1BbaVZPSXO3UOOPL1DrO1T4y5u7VISHZfyMMdx6ajGxg1b53aSTu6HxNQAkK5jveGyLjV3IZLzpld+Cv/6lV03s56loSqsVt3tmjd+LZ2qAMQq/Qhw2u/7YMg8gLtp+vIHroEg9B+MoJ4gknVOWejSDF67M+K/iwq+zcw4ziDKsxi8Yv17jL7JjAimTH8XHT/HEVRg/FWq68A/L+AHgtnhZNmOotNNVlnoajmVk2DQlqkOgzKNMApbFai1m2eaun5uobmoY/6j5+Wkyb8r4TY6joX6uZcV2zllzd98i6eznGH/2V2FbLGVfV3vhz+0grj6IQh/MuYrLoVfmGgU+fp4ck/ZYNVPX2cdHsBi0t9sUT0yMvyiyQd27m/j43erK6K1xg3fOtSsx8LT3VGxXJFlBB+8ADRARyE7c98M4gM3N/TsgXhtlSo2s8SvEYaPrYXVMRw+Qvid+cHDSOQ/GUU4QrqLzAcSUrNz30S3+uNOx+xliF23e1VNVJ04Zf7Xf0XLbRd+PkqlJGWVZPabFMYSywtdy7UqMf87NNnfpWFuN6h+X206JBm/VC6J83EWF37HyC8MJw1yw9wuo8FNcgzrsJct7g5Lmf+lzKcRho+tNROohlt8PZox/38KVhrMIIp0z+4Y5Nks03oPyZo4CE+MfVuqpzvjN07tFzVn5uYxST1BsQ206VpLIWKTxO7aFpmOldk4vhMWGKzg3rS3AsVjlC2Jm6U2BlOHYBXbOYPJST92gHlMS16AUfjoX/FCWesb08QccnHNsdP0JMX7xc8NZVs/+hRrSxrnef+tYEuO/mqUe285HNgzhR6YP4ULV5m7bHNRWZRGL/H0q/DAq9LBTQWUMWCjpSYi9u6nUM9dwhnK2UHRDVbcTefSB4vPNtswavx8dPMZPhT/N4lfsnBLjH9fVQzHkfhih6wlpaSIav3S+2gdE6jl8efzJBi7x4aEPkeq/lRn/VV34JR81YZh42WHtnCtJJn++wVtmR9TlLKmPL2K8yUWq4ZTq7nNNO9PcHSXB8S33XJ/b3VsEJ27cFt1ZuHE8gA4Hadk3ge44dcmc9O+AmNYe19UDxM3kMEriGo5MoPDLasGh9/HvV9iWcHdQ8QgM047OIbJz+iFHFPGkGOp6HiaM4uoBoPXyU+E2FeUyqccLihkvxTYUyTyE+YaTNHf7FbP4VfzE3dcN9f2uLaSoZsHy70LGf4A0ZkLDtjLrMNXmbjORevjYUg+AZJnLRjy1WxQnURV2JtL9YNSKg3GUE4Zrp7o2uXtyGr9lpZO7V3FzVyefiKye4Xz8VZaNAMDaklgI8s3zO7l/Kyvc9FymrJ6yCG2yVaqsUgc5k7/rBZU9/OOAXnvRa3DiXBgdDmpz1wsi7fYtILsukc7RsRi/I6ZsNyaU0wNkWf5BufAerLNkQnCtdOyd3D3qB0Y00YrthVcD1PRDYDRP5oYJAAAVcklEQVQ750LF5u7aYgvfccNRfPD+5xApzLVMqigf4Joc4xdbuEjqiSrt2x0XRD6KGL9TYOf0owMq9cRR2UCe8cvrQb0xXT30WC+IJpbMScdIGCXaYi9wsM6SCcGNr/qAfrE4kH0Dr+bCn2iokqVTjamu8vhhpJCf+vYX4fmNHj7z+KXM18vsmGqMrwo/LA7Uo35EkZWTMN9wJFdPkFg860Qlxl+k8R9Aqafp2DHj99F0LG0vhYr1uM1dIP3sb3QmyPhnUs/BgAhVihm/YdpTfgOv5sI/LuM/sdRC27WHYk7f/60nsDrn4gP3PZv5ul8ha921rcwMRubxFZu7lTT+piMx/nCoqd1RQcdeZOcUGv9VJvWEws5pkuBcm8EPOQYTkHros3+l64Mx85KYYSCTxoPC+A9dcxcQUk+uuVvE+A/Yh2kY6BqmQVTdD/5DLz2N77zpaGWpBxCF7ce/7Vr80WefxvrOIFkE7lXICNLNHRBKm7tDFX47k9UzVY2/aIDLZugH5gGuKtn/+wmNuBBv9TxjYmoj3hkxoFWJY/TcRDM5wmbXw3LbnUioWnaJ08GoFQfjKCcM17GSoKsgae7mNX7C1cz4qXGdKfxDhE3ZFsPaYmvo5/3JV7wIQcTxX778fPK1KoxVtyNYfnzRe0UyQpUY3vmmk27g8qbD+Ol3Xjy5a45lPqh2TgC4tOMZGT95/emCX9QDKQNJPZPK6QGyteOgSG0H6yyZEByLJVIP5fLnGX+1EfqDjtQnnXr5qzDvcXHj8QW84voj+MB9zyZNXi8oL1y6UDlCWeFLGX8VjV+kRnpBJKSeKTB+WjJfzPiLY5kPSuEh0HuyvjswXpDpLm8yzV1xt7/Z9Sdi5QQAOyP1HIxacTCOcsKQ7Zwmxi9/gA4aixoGuvybIJxO2NT/+IoX4ZnLXXz+ycsAyhk7EA+cjTrA5VJzt4qrJ93CNS2NP2X8BZENhRu4Dp7Gn+xJ2BkUa/xxc9cZM3aaiMOVzv/f3rkGyVFdd/x3ZmZnZ7XalbRidwUrrYRBCMkyYCEoHAixeFlgl51K2VVOOcF2qPAhTkLs2C5SSVzJh1Q5FSqJk7JdVsmvpFKObeIkhDhgG+PCuBwetjEIEAaEQdqgB3qsBPvePfnQfWd7Znd2ejU9PdPd51e1pXn0qO/t23Pm3P8995ypSDZvQeWGQ0vZ0MZ05Oelnlrl6pz258LJ0sqiGv9sPDlHdm9fR3cxz71PvlpuQz1vrrjE4m69qKDlaPxuzeLMxAwT03MxafwhUzakUOo5MzmzhMY/7/E3Kru6almnxqYiCeWEynDOpMy4knWXRISLEoBAgeoFKRvqh9algbLUU7WBK45+lzryXHtRP9/ffwRV9XPtnP3i7nSdDVzO8IfR+N2GtON+2F88UT31Nf6OOlJP8SyKoreSoCGv5fGXwzkbKLTu8Gb7yomxKfq6o5F6kpirJxmtjBj3qw/zHv/CJG31va80UN4NWx3HH5Pncv3WQY6cnmTfyOlQUkWpkOehXxxj26fuY9un7uNj33ii/N50nTWC0jKkHpeC4jU/f//ZpGxYLmGTtNXcwBWTRBclwV3xtUIrnTwzOd24Q1Is5Dg9Ps3E9FxkHn/QdliunjamIz+fcrecsqHG4m7aDX+pozKOv5ytNCYDsmtLPyLwvWePMDW7dIIygDtu2MwPnjsKwA+ff40fBTaB1ZsxlD3+kCkbAI6/4Rn+s0nStlzCefzzocjVJFnqgdozsY58jjPTM0zNLp3HKNT58vPFeKLYvAXJDOfMqOGfT3RVjuOvsbibeqmnKpxzusa+hmaxdmUnO4bX8MD+I0zNzC6ZVhng6gvP4eoLzwHgrvuf43M/eMFbjPblgKUM39CaLro68gz4+waWwnn4r73uSz1tYvjrVuBKmNTTWSH1LG6OOgs5jvuLu41+H4M7v6MowgJVG7hM429fnJGAgMZf9YVxCzZpDuWEhYu75SinGH/wrt86wL6R0xw5PbmsL/bgqhJzOq/D10vStmvLAI/92Q2hpvjO4z8Wo9QTZueuV4FrcY9/pk7/25EwHr9b15mcmaN4lvV2FztfdHH8wbTMybj+yWhlxBSDGr/L1bPA48+G1FMdxz/v8cfX7xu2DgJeofHlzDQGfc/9yOkJoH44o4iE3mHcXb24G4PHXwhxz9VK0jY7p8xpcuLIHWEXd6dno1vcdURRfQsqr7l5/G1MISD1jJwcBxYuLGVlcbc6jr88A4rxBt48sJLhvhXA8q73ulXejuHDoxNlwxfVeFUv7saRnTOU1JNfPB//dI2Za7sTnKHUCufsKCdpm6Wz0cXdfBM8/gTm6mmJVRORX4rIUyLyhIg8Hvf5g4m+vvPMYd4ytKqcL8aRz4rGX2X4y1FOMXqOIsL1WweA5c00Bns9w3/kzGTNZHtni/PwXXH2OKWeeh7/YobfpRkOE6raTnSG8fhdVE8EcfzBz0e1c7fS40+GvWhlK3ep6mWqujPuE3fkc0zPKYdHJ/jZK6fYvX3dwmMyEtVT8CuSlQvTzC4e5dRsnNyznB/atd1FcgJHT0/Urde7XHJ+ofQ4pZ55j39pjX92zisWHsTNXNev6WpeA5tAsK9Lavz+4m6ja27uGveUCpE5CebxJwQXEvfdZw4DXprgavIZkXpEpJyxEAI1iGM2/Fds6mP9mi42ru0O/ZlCPkd/TyeHRyfKM7gox2tFZ6HsScej8Yfz+IEFOv/IqWQaftfXro58zX67OrlR7dyF6GQeqFrcTYjH36pwTgW+IyIKfEFV91QfICK3A7cDDA8PR3pyL2WDcv/TR3hTfzcXDvQscoyUj007LvshBDX+ePtdLOR46BO7lp2HZbC35Es90S9KdxfzHPPtaxw7d4shNH4nJczMKcGJwSHf4z9vdTINf61QTnfMVESLu+7zUS3sQqW8Y4u7S3ONqu4AbgY+IiLXVh+gqntUdaeq7uzv74/05IW8MDY1w48PHOcdb14o83jHZEPqgfl859Aajd9xNsm3BntLHD09EbnGD1Tkto/F4w8hLzrvslrnP3RynL7uYjm5XFIoG/4l1iaK+RyqMDY12/jOXd8wRxXDDxbOGRpVHfH/PQr8O3BlnOcv5nPMqTdd3l3L8Ls4/ox5/NMtiOpphMHeTg43QeOH+cieYj4Xy6JdoRzHv/QGLpifmTlGTo0nTuaB+VnOUrup3Y/D6xMzkXn8UWXmhErDn0/I9yZ2qyYi3SLS4x4DNwH74myD86zOXVXikvWrahyTDY0f/FTHfhx/KzZwNcJgT4lTY9O8PuGl4IgyCsslais1mCYgLMUQs0z3w1bt8Y+cHGMoYTIPeP0RqR3K6R3jXY/x6dnINP6o8vSAFVsPyyDwsIj8HHgU+G9VvS/OBrhY55u2DSKy+EBlSeqp9PhdCotk3MCDfiy/07ibIfXEoe/D/L22dM1d75jg4q6qMnJqPJGG3wUXhPH4YelrEwZ3f0SVmRO8PjiDn5Q1wdgFQVU9AFwa93mDOM/qHYuEcTrKHn9CBrIRVnYWGB2fBgK5ixLyg+di+Q+eHAOibXe3b/Dj0s1vfss6RFiwpySIk4OCaRuOvzHFxPQcQwmUesCriHbOytp9Dhr+qKSeKD1+8MZlVjUxtTuStRIUEb92UT+HTo5z5aa+msdkyePf0LeCR186AQSylSbkBl7X6zx+3/BHqLE6gx9HZk6AgZ4St75t05LHLBbOOR/Dv6JpbWsmX/nwFeVd2IvRWeHxR7NzN6rMnI6OXI65BJmKBDU1OjYP9vAX737zkjp2mF2UaWG4bwX/NzrO5MxsOR4+KVPWwV7PUzx4wjN+Uc7QXF6frpg0/jA4j3I6UIXLxfAnUeoB2D60akmPP3gvNjq+Tr4Lk6F1OeTzkphQTsio4Q9DVjZwgWf4VT3P0S0aJuUmXtXVQbGQC3j80S/utlOIpOtf0ON3fU+q1FOPoLFv9Pu4c+MavvyhK7h845pGm1VBIZdLjMwDGZV6wlCOqU6I59sIG9d6EsHLJ8bmUzYkJB5ZRFjXWyov7kb5Q+08/riknjCUwzkDBddHTo7T01moWcEq6USp8edywq6LBxpt0gIKOWEuQbYiOS2NmXIcf0Y8foCDJ8bKi7tJ+sEb7O0sb0CL1OP3Pf04ErSFpbyBq0rqSau3D9VRPe15Xxbykph1MTDDX5NChlI29Pd0UurI8crxsUAcf3JuYhfZA9H+YLmonjh27YYlmLLBcehkMjdvhaUjQqmnWRRykihbkZyWxkxWCrGAJ5cM963wpZ5kafxQafijzEcfdxx/GGpF9SR1YTcMQS+/XWeihXwuUd8Z0/hrkKXFXfDknoMBjT8pOUdgPrIHot7A5Xv8bWj4XcqG0fFpzkzOZEbqadfvYyEnqCbH8LfnVWwDslJs3THc180rwcXdBHkvFR5/EzT+9pJ6KlM2uBj+odXJjOEPQyKknrwkJiACzOOvyaa13Ww9t5eL1/W2uimxMNzXxdjULIdHvYpTSdIrm6Xxz8fxt4/hdykb3FpMUvPwL4dELO7mcmh+rv6BbUJ7XsU2YO3KTv7njl9leG16PakgrgDKgddeB5Jl+NdVePzRzVTWrizS113kgoHwxWGazWo/ZHP/4TNA+mP4ofLHvNFcPc2ikJPEJDYEM/yGzwY/pPPFY68jQqI2owz4Gn/U7V5RLPDTP7+R6y5eWKGtVWw6p5u3b+lnz0MHOD0xzcjJcUodOdZGnIKgnUiCxl/qyFNq07Ythkk9BuBJBSJeaGCSFnbBM9A9pQKTM3M1s62miY/ftIV3/ePD7H3oACOnxjlvdVeq+12MMGVDs7jz5oupKoPc1pjhNwDPY1nXW+LV0YlELew6XNuzwPahVbzzknPZ+/BL9Pd0LqtOcRJJgse/fWjxuh7tSnteRaMlOLknSfq+Y7C3lJiqYVHwsRsvYnJmjpePJ7MAy3LI5yRz4dXNxq6iUWZj2fAnz4Bu6Oti5RJVnNLGBf0ree+O9UC6I3ocYaqTGeGxq2iUcTl7khSP7PjojRfxhd/a2epmxModN2xmQ18XO4ajzTTZjnSEqEdshCc7LpJRFxe6mkSNf6CnxEBP7WIeaeS81V388JPXtboZsVAs5IGZtl3cTRp2FY0ywwnW+I1001nIUcznUh29FCf2DTfKDCdY4zfSTbGQM30/QuxKGmX6uous7CwkUuM30k1HXszwR4hdSaOMiLChb4V5/EbbUfSlHiMabHHXqOB9l68vV7MyjHahmM/R2UZF75OOGX6jgt+55vxWN8EwFtCRN48/SuxKGobR9nR25E3jj5CWePwishv4DJAH9qrqp1vRDsMwksGHr97EG5MzrW5Gaojd8ItIHvgscCNwCHhMRO5R1WfibothGMlg15aBVjchVbRi7nQl8IKqHlDVKeBfgfe0oB2GYRiZpBWGfwg4GHh+yH+tAhG5XUQeF5HHjx07FlvjDMMw0k7brpao6h5V3amqO/v7+1vdHMMwjNTQCsM/AmwIPF/vv2YYhmHEQCsM/2PAZhE5X0SKwPuBe1rQDsMwjEwSe1SPqs6IyO8D9+OFc35JVZ+Oux2GYRhZpSVx/Kr6beDbrTi3YRhG1mnbxV3DMAyjOYiqtroNdRGRY8DLZ/nxc4DXImxOUshiv7PYZ8hmv7PYZ1h+vzeq6oKwyEQY/kYQkcdVNVvFWMlmv7PYZ8hmv7PYZ4iu3yb1GIZhZAwz/IZhGBkjC4Z/T6sb0CKy2O8s9hmy2e8s9hki6nfqNX7DMAyjkix4/IZhGEYAM/yGYRgZI9WGX0R2i8hzIvKCiNzZ6vY0AxHZICIPisgzIvK0iNzhv94nIt8Vkef9f9e0uq1RIyJ5EfmZiNzrPz9fRB7xx/vrfi6oVCEiq0XkbhHZLyLPisjb0j7WIvJR/97eJyJfE5FSGsdaRL4kIkdFZF/gtUXHVjz+we//kyKyYznnSq3hD1T6uhnYBvymiGxrbauawgzwx6q6DbgK+IjfzzuBB1R1M/CA/zxt3AE8G3j+18DfqeqFwEngtpa0qrl8BrhPVS8GLsXrf2rHWkSGgD8Edqrqdrz8Xu8nnWP9FWB31Wu1xvZmYLP/dzvw+eWcKLWGn4xU+lLVV1X1p/7jM3iGYAivr1/1D/sq8OutaWFzEJH1wDuBvf5zAa4D7vYPSWOfVwHXAl8EUNUpVT1FyscaL6dYl4gUgBXAq6RwrFX1IeBE1cu1xvY9wD+px/8Cq0Xk3LDnSrPhD1XpK02IyCbgrcAjwKCqvuq/dRgYbFGzmsXfA58E5vzna4FTquoqcqdxvM8HjgFf9iWuvSLSTYrHWlVHgLuAV/AM/ijwE9I/1o5aY9uQfUuz4c8UIrIS+Dfgj1T1dPA99WJ2UxO3KyLvAo6q6k9a3ZaYKQA7gM+r6luBN6iSdVI41mvwvNvzgfOAbhbKIZkgyrFNs+HPTKUvEenAM/r/oqrf8l8+4qZ+/r9HW9W+JnA18G4R+SWehHcdnva92pcDIJ3jfQg4pKqP+M/vxvshSPNY3wC8pKrHVHUa+Bbe+Kd9rB21xrYh+5Zmw5+JSl++tv1F4FlV/dvAW/cAH/QffxD4z7jb1ixU9U9Udb2qbsIb1++r6geAB4H3+oelqs8AqnoYOCgiW/yXrgeeIcVjjSfxXCUiK/x73fU51WMdoNbY3gPc6kf3XAWMBiSh+qhqav+AW4BfAC8Cf9rq9jSpj9fgTf+eBJ7w/27B07wfAJ4Hvgf0tbqtTer/24F7/cdvAh4FXgC+CXS2un1N6O9lwOP+eP8HsCbtYw38JbAf2Af8M9CZxrEGvoa3jjGNN7u7rdbYAoIXtfgi8BRe1FPoc1nKBsMwjIyRZqnHMAzDWAQz/IZhGBnDDL9hGEbGMMNvGIaRMczwG4ZhZAwz/IZRhZ8B8/f8x+eJyN31PmMYScLCOQ2jCj/n0b3qZYM0jNRRqH+IYWSOTwMXiMgTeBtntqrqdhH5EF52xG68dLh3AUXgt4FJ4BZVPSEiF+BtrukHxoDfVdX98XfDMBbHpB7DWMidwIuqehnwiar3tgO/AVwB/BUwpl7CtB8Dt/rH7AH+QFUvBz4OfC6WVhtGSMzjN4zl8aB6dQ/OiMgo8F/+608Bl/hZUn8F+KaXWgbwUgwYRttght8wlsdk4PFc4Pkc3vcph5cr/rK4G2YYYTGpxzAWcgboOZsPqlcL4SUReR+Ua6NeGmXjDKNRzPAbRhWqehz4kV/0+m/O4r/4AHCbiPwceJoUlvw0ko2FcxqGYWQM8/gNwzAyhhl+wzCMjGGG3zAMI2OY4TcMw8gYZvgNwzAyhhl+wzCMjGGG3zAMI2P8P1++s61winlwAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt \n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "def generate_sequence(length):                         ## Generate a seqeunce of numbers \n",
    "    return np.array([round(random.uniform(0, 30), 2) for i in range(length)])\n",
    "\n",
    "def create_dataset(dataset, time_step):                     ## Build training data X and Y   \n",
    "    dataX, dataY = [], []\n",
    "    for i in range(len(dataset)-time_step-1):\n",
    "        a = dataset[i:(i+time_step)]        \n",
    "        dataX.append(a)\n",
    "        dataY.append(dataset[i + time_step])\n",
    "    return np.array(dataX), np.array(dataY)\n",
    "\n",
    "def normalize_data(data):\n",
    "    scaler = MinMaxScaler()\n",
    "    data = np.reshape(data, (-1, 1))\n",
    "    scaled_data = scaler.fit_transform(data)                      ## normalize data (0 - 1) \n",
    "    return scaled_data.reshape(-1)\n",
    "\n",
    "data = generate_sequence(100)                   \n",
    "plt.ylabel('Value')\n",
    "plt.xlabel('time')\n",
    "plt.plot(data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cfgREGwaeYlX"
   },
   "source": [
    "## RNN architectures \n",
    "\n",
    "![](https://static.javatpoint.com/tutorial/tensorflow/images/tensorflow-types-of-rnn.png)\n",
    "\n",
    "Given a sequence of data X = <x(1), x(2), x(3)..x(t),.., x(T)> be a T-length sequence where t is time step.\n",
    "\n",
    "One-to-One model\n",
    "\n",
    "           y(t) = f(x(t))               \n",
    "\n",
    "Many-to-One model \n",
    "\n",
    "           y(t) = f(x(t), x(t-1), x(t-2)) \n",
    "           \n",
    "Many-to-Many model\n",
    "\n",
    "           y(t), y(t+1), y(t+2) = f(x(t), x(t-1), x(t-2)) \n",
    "\n",
    "One-to-Many model \n",
    "\n",
    "           y(t), y(t+1), y(t+2) = f(x(t)) \n",
    "           "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "SLw7M4Pt60sU"
   },
   "source": [
    "# One-to-One RNN "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VKfYgEyCrWUt"
   },
   "source": [
    "## Generate dataset with time-step "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "INekZJf4pFy4",
    "outputId": "81a03be8-0aa0-4cfe-fc42-ee02254957b6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X's shape =  (98, 1)  Y's shape =  (98,)\n"
     ]
    }
   ],
   "source": [
    "time_step = 1                                                          ## time step = 1 \n",
    "norm_data = normalize_data(data)                                               \n",
    "X, Y = create_dataset(norm_data, time_step)                            ## y(t) = f(x(t))\n",
    "print('X\\'s shape = ', X.shape,' Y\\'s shape = ', Y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VC5V-JE2EU9k"
   },
   "source": [
    "# Reshape input data [samples, time step, n_feature] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "esRuZPpeENhG",
    "outputId": "7cc63f8f-306c-4702-f0e4-2ddb0b0bccae"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before X's reshape:  (98, 1, 1)\n",
      "After X's reshape:  (98, 1, 1)\n"
     ]
    }
   ],
   "source": [
    "# reshape X to be [samples, time steps, features]\n",
    "time_step = 1\n",
    "print('Before X\\'s reshape: ', X.shape)\n",
    "X = np.reshape(X, (98, 1, 1))                 ## [samples, time_step, feature]      \n",
    "print('After X\\'s reshape: ', X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 505
    },
    "id": "RmCIfUMbeYlZ",
    "outputId": "760e0afc-f6a8-4a47-908f-050ad97cfbd9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru (GRU)                    (None, 10)                390       \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 401\n",
      "Trainable params: 401\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "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": [
    "# One-to-One Model \n",
    "from keras.models import Sequential\n",
    "from keras.layers import GRU, Dense\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "####################### Define GRU #######################################################\n",
    "model = Sequential()\n",
    "model.add(GRU(10, input_shape=(1 , 1)))                                 # input_shape = (time_step, feature)\n",
    "model.add(Dense(1))                                                     # output y \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "print(model.summary())\n",
    "#########################  Training ######################################################\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, batch_size=8, epochs=100, verbose=0)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train loss', 'val loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "48XTNimQeYlg",
    "outputId": "e0b6a752-757d-40a4-8452-d804eb810c15"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 20ms/step - loss: 0.0996\n",
      "test loss =  0.09958979487419128\n",
      "input :  [0.89051343]\n",
      "Predict output :  [0.55022067]\n",
      "Actual Output:  0.7963277796667801\n"
     ]
    }
   ],
   "source": [
    "# Test model \n",
    "loss = model.evaluate(X_test, Y_test)\n",
    "print('test loss = ', loss)\n",
    "\n",
    "x_test = X_test[1]\n",
    "x_test = x_test.reshape(1, x_test.shape[0], x_test.shape[1])   ## reshape input from 2D to 3D \n",
    "\n",
    "predicted_y = model.predict(x_test)\n",
    "\n",
    "print('input : ', x_test.reshape(-1))\n",
    "print('Predict output : ', predicted_y.reshape(-1))\n",
    "print('Actual Output: ', Y_test[1])    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tWa407bT10sl"
   },
   "source": [
    "# Many-to-One RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "da6yGk10oAh_",
    "outputId": "25d4a26b-c152-493b-a0f4-05594dc60b35"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X's shape =  (96, 3)  Y's shape =  (96,)\n"
     ]
    }
   ],
   "source": [
    "time_step = 3                                                          ### time step = 3 \n",
    "X, Y = create_dataset(norm_data, time_step)                   \n",
    "print('X\\'s shape = ', X.shape,' Y\\'s shape = ', Y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "2h2XA4EL7nee"
   },
   "source": [
    "what is reshape's parameter of X ? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "R3115mKX7byL",
    "outputId": "7487d01e-c1e8-4c36-947d-682d943a332f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(96, 3, 1)\n"
     ]
    }
   ],
   "source": [
    "# X = np.reshape()\n",
    "X = np.reshape(X, (96, 3, 1))      \n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "NbHjjL3LeYll",
    "outputId": "0ade2338-edeb-4864-d85c-539a31d8986a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru_1 (GRU)                  (None, 10)                390       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 401\n",
      "Trainable params: 401\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/100\n",
      "3/3 [==============================] - 2s 212ms/step - loss: 0.9447 - val_loss: 0.8994\n",
      "Epoch 2/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.6366 - val_loss: 0.6457\n",
      "Epoch 3/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.4439 - val_loss: 0.4683\n",
      "Epoch 4/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.3133 - val_loss: 0.3509\n",
      "Epoch 5/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2302 - val_loss: 0.2798\n",
      "Epoch 6/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1810 - val_loss: 0.2321\n",
      "Epoch 7/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1483 - val_loss: 0.1966\n",
      "Epoch 8/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1258 - val_loss: 0.1750\n",
      "Epoch 9/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1125 - val_loss: 0.1594\n",
      "Epoch 10/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1037 - val_loss: 0.1496\n",
      "Epoch 11/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0983 - val_loss: 0.1406\n",
      "Epoch 12/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0938 - val_loss: 0.1354\n",
      "Epoch 13/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0914 - val_loss: 0.1315\n",
      "Epoch 14/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0899 - val_loss: 0.1296\n",
      "Epoch 15/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0890 - val_loss: 0.1274\n",
      "Epoch 16/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0881 - val_loss: 0.1248\n",
      "Epoch 17/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0873 - val_loss: 0.1224\n",
      "Epoch 18/100\n",
      "3/3 [==============================] - 0s 20ms/step - loss: 0.0870 - val_loss: 0.1219\n",
      "Epoch 19/100\n",
      "3/3 [==============================] - 0s 13ms/step - loss: 0.0866 - val_loss: 0.1219\n",
      "Epoch 20/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0865 - val_loss: 0.1216\n",
      "Epoch 21/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0864 - val_loss: 0.1203\n",
      "Epoch 22/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0862 - val_loss: 0.1203\n",
      "Epoch 23/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0861 - val_loss: 0.1204\n",
      "Epoch 24/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0868 - val_loss: 0.1203\n",
      "Epoch 25/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0860 - val_loss: 0.1196\n",
      "Epoch 26/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0861 - val_loss: 0.1198\n",
      "Epoch 27/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0859 - val_loss: 0.1193\n",
      "Epoch 28/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0857 - val_loss: 0.1194\n",
      "Epoch 29/100\n",
      "3/3 [==============================] - 0s 19ms/step - loss: 0.0857 - val_loss: 0.1192\n",
      "Epoch 30/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0858 - val_loss: 0.1196\n",
      "Epoch 31/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0856 - val_loss: 0.1196\n",
      "Epoch 32/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0857 - val_loss: 0.1193\n",
      "Epoch 33/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0854 - val_loss: 0.1191\n",
      "Epoch 34/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0854 - val_loss: 0.1197\n",
      "Epoch 35/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0854 - val_loss: 0.1189\n",
      "Epoch 36/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0852 - val_loss: 0.1191\n",
      "Epoch 37/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0852 - val_loss: 0.1193\n",
      "Epoch 38/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0853 - val_loss: 0.1194\n",
      "Epoch 39/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0852 - val_loss: 0.1195\n",
      "Epoch 40/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0851 - val_loss: 0.1198\n",
      "Epoch 41/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0850 - val_loss: 0.1199\n",
      "Epoch 42/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0851 - val_loss: 0.1195\n",
      "Epoch 43/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0848 - val_loss: 0.1196\n",
      "Epoch 44/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0848 - val_loss: 0.1190\n",
      "Epoch 45/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0849 - val_loss: 0.1196\n",
      "Epoch 46/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0847 - val_loss: 0.1197\n",
      "Epoch 47/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0849 - val_loss: 0.1197\n",
      "Epoch 48/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0846 - val_loss: 0.1199\n",
      "Epoch 49/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0845 - val_loss: 0.1196\n",
      "Epoch 50/100\n",
      "3/3 [==============================] - 0s 19ms/step - loss: 0.0845 - val_loss: 0.1198\n",
      "Epoch 51/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0844 - val_loss: 0.1197\n",
      "Epoch 52/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0845 - val_loss: 0.1188\n",
      "Epoch 53/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0846 - val_loss: 0.1200\n",
      "Epoch 54/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0847 - val_loss: 0.1198\n",
      "Epoch 55/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0843 - val_loss: 0.1192\n",
      "Epoch 56/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0842 - val_loss: 0.1189\n",
      "Epoch 57/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0843 - val_loss: 0.1193\n",
      "Epoch 58/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0841 - val_loss: 0.1194\n",
      "Epoch 59/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0842 - val_loss: 0.1191\n",
      "Epoch 60/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0840 - val_loss: 0.1194\n",
      "Epoch 61/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0840 - val_loss: 0.1191\n",
      "Epoch 62/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0841 - val_loss: 0.1193\n",
      "Epoch 63/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0839 - val_loss: 0.1196\n",
      "Epoch 64/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0839 - val_loss: 0.1193\n",
      "Epoch 65/100\n",
      "3/3 [==============================] - 0s 20ms/step - loss: 0.0838 - val_loss: 0.1188\n",
      "Epoch 66/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0837 - val_loss: 0.1192\n",
      "Epoch 67/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0837 - val_loss: 0.1194\n",
      "Epoch 68/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0838 - val_loss: 0.1203\n",
      "Epoch 69/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0840 - val_loss: 0.1204\n",
      "Epoch 70/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0839 - val_loss: 0.1212\n",
      "Epoch 71/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0838 - val_loss: 0.1209\n",
      "Epoch 72/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0837 - val_loss: 0.1202\n",
      "Epoch 73/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0836 - val_loss: 0.1200\n",
      "Epoch 74/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0834 - val_loss: 0.1194\n",
      "Epoch 75/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0835 - val_loss: 0.1185\n",
      "Epoch 76/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0833 - val_loss: 0.1184\n",
      "Epoch 77/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0833 - val_loss: 0.1179\n",
      "Epoch 78/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0834 - val_loss: 0.1175\n",
      "Epoch 79/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0834 - val_loss: 0.1180\n",
      "Epoch 80/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0832 - val_loss: 0.1186\n",
      "Epoch 81/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0833 - val_loss: 0.1182\n",
      "Epoch 82/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0831 - val_loss: 0.1181\n",
      "Epoch 83/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0831 - val_loss: 0.1176\n",
      "Epoch 84/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0831 - val_loss: 0.1175\n",
      "Epoch 85/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0830 - val_loss: 0.1175\n",
      "Epoch 86/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0831 - val_loss: 0.1177\n",
      "Epoch 87/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0830 - val_loss: 0.1174\n",
      "Epoch 88/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0832 - val_loss: 0.1171\n",
      "Epoch 89/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0832 - val_loss: 0.1168\n",
      "Epoch 90/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0831 - val_loss: 0.1171\n",
      "Epoch 91/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0831 - val_loss: 0.1168\n",
      "Epoch 92/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0830 - val_loss: 0.1171\n",
      "Epoch 93/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0828 - val_loss: 0.1176\n",
      "Epoch 94/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0827 - val_loss: 0.1182\n",
      "Epoch 95/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.0825 - val_loss: 0.1181\n",
      "Epoch 96/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0825 - val_loss: 0.1178\n",
      "Epoch 97/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.0826 - val_loss: 0.1179\n",
      "Epoch 98/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0824 - val_loss: 0.1182\n",
      "Epoch 99/100\n",
      "3/3 [==============================] - 0s 20ms/step - loss: 0.0826 - val_loss: 0.1176\n",
      "Epoch 100/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0824 - val_loss: 0.1181\n"
     ]
    },
    {
     "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": [
    "# Many-to-one LSTM \n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, GRU\n",
    "from keras.layers import Dense\n",
    "from numpy import array  \n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "######################################################################################################\n",
    "model = Sequential()\n",
    "model.add(GRU(10, input_shape = (3, 1)))           ## (3 GRU's cells , 1 feature) \n",
    "model.add(Dense(1))                                ##  output y \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "print(model.summary())\n",
    "########################################################################################################\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=100)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train loss, validation loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "_SC3TThaeYlp",
    "outputId": "3ea2dabd-64e4-43db-ae91-e563b41c6c63"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 18ms/step - loss: 0.0890\n",
      " loss:  0.08901672065258026\n",
      "input :  [[[0.98299898]\n",
      "  [0.07344441]\n",
      "  [0.04386263]]]\n",
      "Predict output :  [0.6929723]\n",
      "Actual Output:  0.17137028221693304\n"
     ]
    }
   ],
   "source": [
    "# Test model \n",
    "loss = model.evaluate(X_test, Y_test)\n",
    "print(' loss: ', loss)\n",
    "\n",
    "x_test = X_test[1]\n",
    "x_test = x_test.reshape(1, 3, 1)\n",
    "testPredict = model.predict(x_test)\n",
    "print('input : ', x_test)\n",
    "print('Predict output : ', testPredict.reshape(-1))\n",
    "print('Actual Output: ', Y_test[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XjI2w6Fv5QUo"
   },
   "source": [
    "## One-to-Many RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "0GDZF7ehqgV3",
    "outputId": "cd198769-9421-463a-a43e-9740da05f769"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X's shape :  (95, 1, 1) Y's shape :  (95, 5)\n"
     ]
    }
   ],
   "source": [
    "def create_dataset(dataset, look_back, look_forward):\n",
    "    dataX, dataY = [], []\n",
    "    np.array(dataY)\n",
    "    for i in range(len(dataset)-look_back-1):\n",
    "        a = dataset[i:(i+look_back)]\n",
    "        if len(dataset[i + look_back:i+look_back+look_forward]) == look_forward:\n",
    "            dataX.append(a)\n",
    "            dataY.append(dataset[i + look_back:i+look_back+look_forward])\n",
    "    return np.array(dataX), np.array(dataY)\n",
    "\n",
    "##########################################################################################################\n",
    "look_back = 1                                        #### time steps for X \n",
    "look_forward = 5                                     #### time steps for Y \n",
    "\n",
    "data = generate_sequence(100)\n",
    "norm_data = normalize_data(data)\n",
    "\n",
    "X, Y = create_dataset(norm_data, look_back, look_forward)  \n",
    "\n",
    "X = X.reshape(95, 1, 1)\n",
    "\n",
    "print('X\\'s shape : ', X.shape, 'Y\\'s shape : ', Y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "sJx90Oro-mIp",
    "outputId": "99100513-1eee-4a31-a074-5ed9521d53bf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru_4 (GRU)                  (None, 10)                390       \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 5)                 55        \n",
      "=================================================================\n",
      "Total params: 445\n",
      "Trainable params: 445\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/100\n",
      "3/3 [==============================] - 2s 233ms/step - loss: 0.4262 - val_loss: 0.3757\n",
      "Epoch 2/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.4141 - val_loss: 0.3651\n",
      "Epoch 3/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.4025 - val_loss: 0.3549\n",
      "Epoch 4/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3912 - val_loss: 0.3449\n",
      "Epoch 5/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3803 - val_loss: 0.3356\n",
      "Epoch 6/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3700 - val_loss: 0.3266\n",
      "Epoch 7/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3601 - val_loss: 0.3178\n",
      "Epoch 8/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.3504 - val_loss: 0.3092\n",
      "Epoch 9/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3409 - val_loss: 0.3011\n",
      "Epoch 10/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.3319 - val_loss: 0.2931\n",
      "Epoch 11/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.3230 - val_loss: 0.2856\n",
      "Epoch 12/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.3146 - val_loss: 0.2783\n",
      "Epoch 13/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.3066 - val_loss: 0.2713\n",
      "Epoch 14/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.2988 - val_loss: 0.2644\n",
      "Epoch 15/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.2911 - val_loss: 0.2579\n",
      "Epoch 16/100\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.2839 - val_loss: 0.2518\n",
      "Epoch 17/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2771 - val_loss: 0.2459\n",
      "Epoch 18/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.2704 - val_loss: 0.2400\n",
      "Epoch 19/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.2639 - val_loss: 0.2348\n",
      "Epoch 20/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.2580 - val_loss: 0.2295\n",
      "Epoch 21/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.2521 - val_loss: 0.2243\n",
      "Epoch 22/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.2462 - val_loss: 0.2193\n",
      "Epoch 23/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2406 - val_loss: 0.2145\n",
      "Epoch 24/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2352 - val_loss: 0.2099\n",
      "Epoch 25/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2300 - val_loss: 0.2055\n",
      "Epoch 26/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.2250 - val_loss: 0.2013\n",
      "Epoch 27/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2202 - val_loss: 0.1970\n",
      "Epoch 28/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.2155 - val_loss: 0.1931\n",
      "Epoch 29/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.2111 - val_loss: 0.1895\n",
      "Epoch 30/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.2070 - val_loss: 0.1859\n",
      "Epoch 31/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.2029 - val_loss: 0.1823\n",
      "Epoch 32/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.1988 - val_loss: 0.1787\n",
      "Epoch 33/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1948 - val_loss: 0.1754\n",
      "Epoch 34/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1910 - val_loss: 0.1722\n",
      "Epoch 35/100\n",
      "3/3 [==============================] - 0s 13ms/step - loss: 0.1873 - val_loss: 0.1692\n",
      "Epoch 36/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1839 - val_loss: 0.1663\n",
      "Epoch 37/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1806 - val_loss: 0.1634\n",
      "Epoch 38/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1773 - val_loss: 0.1605\n",
      "Epoch 39/100\n",
      "3/3 [==============================] - 0s 19ms/step - loss: 0.1741 - val_loss: 0.1579\n",
      "Epoch 40/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.1711 - val_loss: 0.1554\n",
      "Epoch 41/100\n",
      "3/3 [==============================] - 0s 20ms/step - loss: 0.1682 - val_loss: 0.1530\n",
      "Epoch 42/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1654 - val_loss: 0.1506\n",
      "Epoch 43/100\n",
      "3/3 [==============================] - 0s 26ms/step - loss: 0.1627 - val_loss: 0.1483\n",
      "Epoch 44/100\n",
      "3/3 [==============================] - 0s 20ms/step - loss: 0.1601 - val_loss: 0.1461\n",
      "Epoch 45/100\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.1576 - val_loss: 0.1440\n",
      "Epoch 46/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.1551 - val_loss: 0.1419\n",
      "Epoch 47/100\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.1528 - val_loss: 0.1399\n",
      "Epoch 48/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.1504 - val_loss: 0.1379\n",
      "Epoch 49/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.1481 - val_loss: 0.1361\n",
      "Epoch 50/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1461 - val_loss: 0.1345\n",
      "Epoch 51/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1442 - val_loss: 0.1328\n",
      "Epoch 52/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1422 - val_loss: 0.1312\n",
      "Epoch 53/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1403 - val_loss: 0.1296\n",
      "Epoch 54/100\n",
      "3/3 [==============================] - 0s 13ms/step - loss: 0.1384 - val_loss: 0.1280\n",
      "Epoch 55/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1365 - val_loss: 0.1265\n",
      "Epoch 56/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1348 - val_loss: 0.1250\n",
      "Epoch 57/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1331 - val_loss: 0.1236\n",
      "Epoch 58/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1315 - val_loss: 0.1222\n",
      "Epoch 59/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1299 - val_loss: 0.1210\n",
      "Epoch 60/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1283 - val_loss: 0.1196\n",
      "Epoch 61/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1267 - val_loss: 0.1184\n",
      "Epoch 62/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1252 - val_loss: 0.1171\n",
      "Epoch 63/100\n",
      "3/3 [==============================] - 0s 30ms/step - loss: 0.1238 - val_loss: 0.1160\n",
      "Epoch 64/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1224 - val_loss: 0.1149\n",
      "Epoch 65/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1211 - val_loss: 0.1138\n",
      "Epoch 66/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1198 - val_loss: 0.1127\n",
      "Epoch 67/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1184 - val_loss: 0.1117\n",
      "Epoch 68/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1173 - val_loss: 0.1108\n",
      "Epoch 69/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.1162 - val_loss: 0.1098\n",
      "Epoch 70/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1150 - val_loss: 0.1089\n",
      "Epoch 71/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1139 - val_loss: 0.1080\n",
      "Epoch 72/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1128 - val_loss: 0.1070\n",
      "Epoch 73/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1117 - val_loss: 0.1063\n",
      "Epoch 74/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1108 - val_loss: 0.1055\n",
      "Epoch 75/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1098 - val_loss: 0.1048\n",
      "Epoch 76/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1089 - val_loss: 0.1040\n",
      "Epoch 77/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1080 - val_loss: 0.1033\n",
      "Epoch 78/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1071 - val_loss: 0.1026\n",
      "Epoch 79/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.1063 - val_loss: 0.1020\n",
      "Epoch 80/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1055 - val_loss: 0.1013\n",
      "Epoch 81/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1048 - val_loss: 0.1007\n",
      "Epoch 82/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1040 - val_loss: 0.1001\n",
      "Epoch 83/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1032 - val_loss: 0.0995\n",
      "Epoch 84/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.1025 - val_loss: 0.0990\n",
      "Epoch 85/100\n",
      "3/3 [==============================] - 0s 14ms/step - loss: 0.1018 - val_loss: 0.0984\n",
      "Epoch 86/100\n",
      "3/3 [==============================] - 0s 21ms/step - loss: 0.1011 - val_loss: 0.0979\n",
      "Epoch 87/100\n",
      "3/3 [==============================] - 0s 17ms/step - loss: 0.1006 - val_loss: 0.0974\n",
      "Epoch 88/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.1000 - val_loss: 0.0969\n",
      "Epoch 89/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0994 - val_loss: 0.0964\n",
      "Epoch 90/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0987 - val_loss: 0.0959\n",
      "Epoch 91/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0982 - val_loss: 0.0954\n",
      "Epoch 92/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0976 - val_loss: 0.0950\n",
      "Epoch 93/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0970 - val_loss: 0.0945\n",
      "Epoch 94/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0965 - val_loss: 0.0942\n",
      "Epoch 95/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0960 - val_loss: 0.0939\n",
      "Epoch 96/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0956 - val_loss: 0.0936\n",
      "Epoch 97/100\n",
      "3/3 [==============================] - 0s 15ms/step - loss: 0.0952 - val_loss: 0.0933\n",
      "Epoch 98/100\n",
      "3/3 [==============================] - 0s 24ms/step - loss: 0.0948 - val_loss: 0.0929\n",
      "Epoch 99/100\n",
      "3/3 [==============================] - 0s 16ms/step - loss: 0.0944 - val_loss: 0.0926\n",
      "Epoch 100/100\n",
      "3/3 [==============================] - 0s 18ms/step - loss: 0.0940 - val_loss: 0.0923\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, GRU\n",
    "from keras.layers import Dense\n",
    "from numpy import array  \n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "######################################################################################################\n",
    "model = Sequential()\n",
    "model.add(GRU(10, input_shape = (1, 1)))           ## (time step = 1 , feature = 1) \n",
    "model.add(Dense(5))                                ##  output y \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "print(model.summary())\n",
    "########################################################################################################\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=100)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train_loss, val_loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "M2u_L04uBSF3",
    "outputId": "a4037553-a2f5-47f5-ff45-f5591b2d8a58"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input :  [[[0.87778552]]]\n",
      "Predict output :  [0.37652576 0.37678432 0.41520366 0.37429613 0.43235084]\n",
      "Actual Output:  [0.57729805 0.09714485 0.69150418 0.84749304 0.65355153]\n"
     ]
    }
   ],
   "source": [
    "x_test = X_test[1]\n",
    "x_test = x_test.reshape(1, 1, 1)\n",
    "\n",
    "testPredict = model.predict(x_test)\n",
    "print('input : ', x_test)\n",
    "print('Predict output : ', testPredict.reshape(-1))\n",
    "print('Actual Output: ', Y_test[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mBZqm0qYBjsS"
   },
   "source": [
    "# Many-to-Many RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "25Op_3LjBdni",
    "outputId": "396dcb9e-32cb-4647-c9b7-229bb2105f83"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X's shape :  (94, 2, 1) Y's shape :  (94, 5)\n"
     ]
    }
   ],
   "source": [
    "def create_dataset(dataset, look_back, look_forward):\n",
    "    dataX, dataY = [], []\n",
    "    np.array(dataY)\n",
    "    for i in range(len(dataset)-look_back-1):\n",
    "        a = dataset[i:(i+look_back)]\n",
    "        if len(dataset[i + look_back:i+look_back+look_forward]) == look_forward:\n",
    "            dataX.append(a)\n",
    "            dataY.append(dataset[i + look_back:i+look_back+look_forward])\n",
    "    return np.array(dataX), np.array(dataY)\n",
    "\n",
    "##########################################################################################################\n",
    "look_back = 2                                        #### time steps for X \n",
    "look_forward = 5                                     #### time steps for Y \n",
    "\n",
    "data = generate_sequence(100)\n",
    "norm_data = normalize_data(data)\n",
    "\n",
    "X, Y = create_dataset(norm_data, look_back, look_forward)  \n",
    "\n",
    "X = X.reshape(94, 2, 1)\n",
    "print('X\\'s shape : ', X.shape, 'Y\\'s shape : ', Y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 574
    },
    "id": "BbjoSLmYeYlt",
    "outputId": "a6973fe1-deff-4c10-b570-8bd89b1a6605"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_7\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru_7 (GRU)                  (None, 10)                390       \n",
      "_________________________________________________________________\n",
      "repeat_vector_1 (RepeatVecto (None, 5, 10)             0         \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 5, 10)             840       \n",
      "_________________________________________________________________\n",
      "time_distributed_1 (TimeDist (None, 5, 1)              11        \n",
      "=================================================================\n",
      "Total params: 1,241\n",
      "Trainable params: 1,241\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "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": [
    "# Many-to-Many LSTM \n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, GRU, RepeatVector, TimeDistributed\n",
    "from keras.layers import Dense\n",
    "from numpy import array  \n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "######################################################################################################\n",
    "model = Sequential()\n",
    "\n",
    "model.add(GRU(10, input_shape=(2, 1)))         # encode layer \n",
    "\n",
    "model.add(RepeatVector(5))                     # repeat vector  \n",
    "\n",
    "model.add(LSTM(10, return_sequences=True))     # decode layer \n",
    "\n",
    "model.add(TimeDistributed(Dense(1)))\n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "\n",
    "print(model.summary())\n",
    "########################################################################################################\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=100, verbose=0)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train_loss, val_loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "eB2GVJmqeYly",
    "outputId": "3a84c0de-99cf-46c6-d0ab-c51693163a78"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input :  [[[0.21489289]\n",
      "  [0.25807548]]]\n",
      "Predict output :  [0.3638235  0.41362894 0.44663447 0.46853822 0.4830976 ]\n",
      "Actual Output:  [0.4083645  0.30873852 0.66575995 0.67460048 0.3199592 ]\n"
     ]
    }
   ],
   "source": [
    "x_test = X_test[1]\n",
    "x_test = x_test.reshape(1, 2, 1)\n",
    "\n",
    "testPredict = model.predict(x_test)\n",
    "print('input : ', x_test)\n",
    "print('Predict output : ', testPredict.reshape(-1))\n",
    "print('Actual Output: ', Y_test[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8QrW6_gZGII2"
   },
   "source": [
    "# Stock price prediction using RNN "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OpoajQNuGMSy",
    "outputId": "7734f4ad-dd45-470a-e22d-9b6dc9f468b4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pandas-datareader in /usr/local/lib/python3.7/dist-packages (0.9.0)\n",
      "Requirement already satisfied: lxml in /usr/local/lib/python3.7/dist-packages (from pandas-datareader) (4.2.6)\n",
      "Requirement already satisfied: pandas>=0.23 in /usr/local/lib/python3.7/dist-packages (from pandas-datareader) (1.1.5)\n",
      "Requirement already satisfied: requests>=2.19.0 in /usr/local/lib/python3.7/dist-packages (from pandas-datareader) (2.23.0)\n",
      "Requirement already satisfied: numpy>=1.15.4 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->pandas-datareader) (1.19.5)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->pandas-datareader) (2.8.2)\n",
      "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->pandas-datareader) (2018.9)\n",
      "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas>=0.23->pandas-datareader) (1.15.0)\n",
      "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests>=2.19.0->pandas-datareader) (3.0.4)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests>=2.19.0->pandas-datareader) (2.10)\n",
      "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests>=2.19.0->pandas-datareader) (1.24.3)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests>=2.19.0->pandas-datareader) (2021.5.30)\n",
      "Collecting yfinance\n",
      "  Downloading yfinance-0.1.63.tar.gz (26 kB)\n",
      "Requirement already satisfied: pandas>=0.24 in /usr/local/lib/python3.7/dist-packages (from yfinance) (1.1.5)\n",
      "Requirement already satisfied: numpy>=1.15 in /usr/local/lib/python3.7/dist-packages (from yfinance) (1.19.5)\n",
      "Requirement already satisfied: requests>=2.20 in /usr/local/lib/python3.7/dist-packages (from yfinance) (2.23.0)\n",
      "Requirement already satisfied: multitasking>=0.0.7 in /usr/local/lib/python3.7/dist-packages (from yfinance) (0.0.9)\n",
      "Collecting lxml>=4.5.1\n",
      "  Downloading lxml-4.6.3-cp37-cp37m-manylinux2014_x86_64.whl (6.3 MB)\n",
      "\u001b[K     |████████████████████████████████| 6.3 MB 8.2 MB/s \n",
      "\u001b[?25hRequirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24->yfinance) (2018.9)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24->yfinance) (2.8.2)\n",
      "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas>=0.24->yfinance) (1.15.0)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests>=2.20->yfinance) (2.10)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests>=2.20->yfinance) (2021.5.30)\n",
      "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests>=2.20->yfinance) (1.24.3)\n",
      "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests>=2.20->yfinance) (3.0.4)\n",
      "Building wheels for collected packages: yfinance\n",
      "  Building wheel for yfinance (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Created wheel for yfinance: filename=yfinance-0.1.63-py2.py3-none-any.whl size=23918 sha256=e1d779139b9a1fa0bbf8cf1d86bcbb97866164b94b4738abdf1b116a3c90f4e5\n",
      "  Stored in directory: /root/.cache/pip/wheels/fe/87/8b/7ec24486e001d3926537f5f7801f57a74d181be25b11157983\n",
      "Successfully built yfinance\n",
      "Installing collected packages: lxml, yfinance\n",
      "  Attempting uninstall: lxml\n",
      "    Found existing installation: lxml 4.2.6\n",
      "    Uninstalling lxml-4.2.6:\n",
      "      Successfully uninstalled lxml-4.2.6\n",
      "Successfully installed lxml-4.6.3 yfinance-0.1.63\n"
     ]
    }
   ],
   "source": [
    "!pip install yfinance                   ## install yahoo finance API"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "iKyVqKQKiWZG"
   },
   "source": [
    "## Retrive UREKA stock data  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 455
    },
    "id": "gIpkPXQnGdya",
    "outputId": "077c2a4a-183e-4a69-8641-cda5ee5b37f3"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Open</th>\n",
       "      <th>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2013-03-01</th>\n",
       "      <td>2.868080</td>\n",
       "      <td>2.868080</td>\n",
       "      <td>2.783100</td>\n",
       "      <td>2.868080</td>\n",
       "      <td>117838764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-03-04</th>\n",
       "      <td>3.314226</td>\n",
       "      <td>3.356716</td>\n",
       "      <td>2.549405</td>\n",
       "      <td>2.570650</td>\n",
       "      <td>251703086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-03-05</th>\n",
       "      <td>2.613140</td>\n",
       "      <td>2.613140</td>\n",
       "      <td>2.209484</td>\n",
       "      <td>2.273219</td>\n",
       "      <td>46487210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-03-06</th>\n",
       "      <td>2.273219</td>\n",
       "      <td>2.400689</td>\n",
       "      <td>2.116006</td>\n",
       "      <td>2.145749</td>\n",
       "      <td>55903800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-03-07</th>\n",
       "      <td>2.166994</td>\n",
       "      <td>2.230729</td>\n",
       "      <td>2.116006</td>\n",
       "      <td>2.124504</td>\n",
       "      <td>17275108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-10-08</th>\n",
       "      <td>3.760000</td>\n",
       "      <td>3.960000</td>\n",
       "      <td>3.760000</td>\n",
       "      <td>3.780000</td>\n",
       "      <td>37383000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-10-11</th>\n",
       "      <td>3.860000</td>\n",
       "      <td>3.900000</td>\n",
       "      <td>3.780000</td>\n",
       "      <td>3.840000</td>\n",
       "      <td>19537900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-10-12</th>\n",
       "      <td>3.900000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>3.840000</td>\n",
       "      <td>3.920000</td>\n",
       "      <td>24821900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-10-14</th>\n",
       "      <td>3.920000</td>\n",
       "      <td>4.020000</td>\n",
       "      <td>3.880000</td>\n",
       "      <td>3.940000</td>\n",
       "      <td>22285700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-10-15</th>\n",
       "      <td>3.940000</td>\n",
       "      <td>3.960000</td>\n",
       "      <td>2.760000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>79543800</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2100 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Open      High       Low     Close     Volume\n",
       "Date                                                         \n",
       "2013-03-01  2.868080  2.868080  2.783100  2.868080  117838764\n",
       "2013-03-04  3.314226  3.356716  2.549405  2.570650  251703086\n",
       "2013-03-05  2.613140  2.613140  2.209484  2.273219   46487210\n",
       "2013-03-06  2.273219  2.400689  2.116006  2.145749   55903800\n",
       "2013-03-07  2.166994  2.230729  2.116006  2.124504   17275108\n",
       "...              ...       ...       ...       ...        ...\n",
       "2021-10-08  3.760000  3.960000  3.760000  3.780000   37383000\n",
       "2021-10-11  3.860000  3.900000  3.780000  3.840000   19537900\n",
       "2021-10-12  3.900000  4.000000  3.840000  3.920000   24821900\n",
       "2021-10-14  3.920000  4.020000  3.880000  3.940000   22285700\n",
       "2021-10-15  3.940000  3.960000  2.760000  3.000000   79543800\n",
       "\n",
       "[2100 rows x 5 columns]"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import yfinance as yf \n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "u= yf.Ticker('UREKA.BK')\n",
    "stock_data = u.history(period=\"max\", actions=False) \n",
    "stock_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 531
    },
    "id": "SYredcBmmkGQ",
    "outputId": "7602e1fa-720c-46e0-ff83-294c4b43d156"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f25eb3c1f50>"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "stock_data['Close'].plot(figsize=(10, 9))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pZM07lVliwX9"
   },
   "source": [
    "## Many-to-One RNN using Close price   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "iH2fiR1PfTi8",
    "outputId": "e465feb2-d286-4b4e-e94e-5d6ee7b0d1d7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train data :  (1990, 10, 1) (1990,)\n",
      "test_data :  (90, 10, 1) (90,)\n"
     ]
    }
   ],
   "source": [
    "def generate_dataset(timesteps, train, test):\n",
    "    X_train = []\n",
    "    Y_train = []\n",
    "    X_val = []\n",
    "    Y_val = []\n",
    "    for i in range(timesteps,train.shape[0]):\n",
    "        X_train.append(train[i-timesteps:i])\n",
    "        Y_train.append(train[i][0])\n",
    "    X_train,Y_train = np.array(X_train),np.array(Y_train)\n",
    "  \n",
    "    for i in range(timesteps,test.shape[0]):\n",
    "        X_val.append(test[i-timesteps:i])\n",
    "        Y_val.append(test[i][0])\n",
    "    X_val,Y_val = np.array(X_val),np.array(Y_val)\n",
    "    return X_train, Y_train, X_val, Y_val\n",
    "\n",
    "sc = MinMaxScaler()\n",
    "dataset = sc.fit_transform(stock_data['Close'].values.reshape(-1, 1))  ## normalize closed price data \n",
    "\n",
    "train = dataset[:2000]                    # split training data / test data \n",
    "test = dataset[2000:]\n",
    "\n",
    "timestep = 10\n",
    "X_train, Y_train, X_test, Y_test = generate_dataset(timestep, train, test)\n",
    "print('train data : ', X_train.shape, Y_train.shape)\n",
    "print('test_data : ' , X_test.shape, Y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "jBYNVEwnhY4V",
    "outputId": "78d5ed9f-f880-473d-eac6-cb2b38b087cf"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import GRU\n",
    "from keras.layers import Dense\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "model = Sequential()\n",
    "model.add(GRU(50, input_shape = (10, 1)))          \n",
    "model.add(Dense(1))                               \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=100, verbose=0)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train_loss, val_loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oXAtxuzFmxfO"
   },
   "source": [
    "## Model's evaluation "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "etIwBXDbl_3Z",
    "outputId": "5cc2babb-113d-4c55-c439-e4e5b4b352a4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " MSE :  0.0035318839645154866\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "Y_hat = model.predict(X_test)                            \n",
    "\n",
    "print(' MSE : ', mean_squared_error(Y_hat, Y_test))          ## compute mean sqaure error on test data \n",
    "plt.plot(Y_test,c = 'r')\n",
    "plt.plot(Y_hat,c = 'y')\n",
    "plt.xlabel('Day')\n",
    "plt.ylabel('Closed Price')\n",
    "plt.title(\"Stock Price Prediction\")\n",
    "plt.legend(['Actual','Predicted'],loc = 'lower right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HealXAfkeYmG"
   },
   "source": [
    "## Homework\n",
    "\n",
    "1. Download airline passenger dataset from \n",
    "https://raw.githubusercontent.com/lazyprogrammer/machine_learning_examples/master/airline/international-airline-passengers.csv\n",
    "\n",
    "and then build many-to-one model to get the input number of passengers p(t-2),\n",
    "p(t-1), p(t) and predict p(t+1)\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "YKeivLQ0QzYt",
    "outputId": "b5dedce6-a7fd-46f5-d90c-d78fe4af903a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(141, 2) (141,)\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd \n",
    "\n",
    "def create_dataset(dataset, time_step):                     ## Build training data X and Y   \n",
    "    dataX, dataY = [], []\n",
    "    for i in range(len(dataset)-time_step-1):\n",
    "        a = dataset[i:(i+time_step)]        \n",
    "        dataX.append(a)\n",
    "        dataY.append(dataset[i + time_step])\n",
    "    return np.array(dataX), np.array(dataY)\n",
    "\n",
    "data = pd.read_csv('https://raw.githubusercontent.com/lazyprogrammer/machine_learning_examples/master/airline/international-airline-passengers.csv')\n",
    "data = data.dropna()                  ## remove n/a record\n",
    "YY = data['International airline passengers: monthly totals in thousands. Jan 49 ? Dec 60'].values\n",
    "\n",
    "YY = YY / 1000.0\n",
    "# norm_data = normalize_data(YY)\n",
    "X, Y = create_dataset(norm_data, 2)\n",
    "\n",
    "print(X.shape, Y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "id": "XP1YoVXyQYA3"
   },
   "outputs": [],
   "source": [
    "X = X.reshape(141, 2, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "cFzqENEmTJjD",
    "outputId": "b9592166-7ca7-49bc-f24d-553f8ccc7069"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_11\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "gru_11 (GRU)                 (None, 10)                390       \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 1)                 11        \n",
      "=================================================================\n",
      "Total params: 401\n",
      "Trainable params: 401\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/100\n",
      "4/4 [==============================] - 3s 148ms/step - loss: 0.0564 - val_loss: 0.0695\n",
      "Epoch 2/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0419 - val_loss: 0.0559\n",
      "Epoch 3/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0329 - val_loss: 0.0463\n",
      "Epoch 4/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0269 - val_loss: 0.0404\n",
      "Epoch 5/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0234 - val_loss: 0.0358\n",
      "Epoch 6/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0209 - val_loss: 0.0328\n",
      "Epoch 7/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0194 - val_loss: 0.0306\n",
      "Epoch 8/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0184 - val_loss: 0.0291\n",
      "Epoch 9/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0177 - val_loss: 0.0277\n",
      "Epoch 10/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0172 - val_loss: 0.0268\n",
      "Epoch 11/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0169 - val_loss: 0.0264\n",
      "Epoch 12/100\n",
      "4/4 [==============================] - 0s 17ms/step - loss: 0.0167 - val_loss: 0.0258\n",
      "Epoch 13/100\n",
      "4/4 [==============================] - 0s 19ms/step - loss: 0.0165 - val_loss: 0.0255\n",
      "Epoch 14/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0163 - val_loss: 0.0252\n",
      "Epoch 15/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0161 - val_loss: 0.0248\n",
      "Epoch 16/100\n",
      "4/4 [==============================] - 0s 15ms/step - loss: 0.0160 - val_loss: 0.0246\n",
      "Epoch 17/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0158 - val_loss: 0.0243\n",
      "Epoch 18/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0157 - val_loss: 0.0240\n",
      "Epoch 19/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0156 - val_loss: 0.0236\n",
      "Epoch 20/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0154 - val_loss: 0.0235\n",
      "Epoch 21/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0153 - val_loss: 0.0233\n",
      "Epoch 22/100\n",
      "4/4 [==============================] - 0s 17ms/step - loss: 0.0151 - val_loss: 0.0231\n",
      "Epoch 23/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0150 - val_loss: 0.0229\n",
      "Epoch 24/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0149 - val_loss: 0.0229\n",
      "Epoch 25/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0148 - val_loss: 0.0226\n",
      "Epoch 26/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0146 - val_loss: 0.0224\n",
      "Epoch 27/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0146 - val_loss: 0.0222\n",
      "Epoch 28/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0144 - val_loss: 0.0220\n",
      "Epoch 29/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0143 - val_loss: 0.0219\n",
      "Epoch 30/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0142 - val_loss: 0.0218\n",
      "Epoch 31/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0140 - val_loss: 0.0216\n",
      "Epoch 32/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0139 - val_loss: 0.0215\n",
      "Epoch 33/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0138 - val_loss: 0.0213\n",
      "Epoch 34/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0137 - val_loss: 0.0212\n",
      "Epoch 35/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0136 - val_loss: 0.0209\n",
      "Epoch 36/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0135 - val_loss: 0.0207\n",
      "Epoch 37/100\n",
      "4/4 [==============================] - 0s 15ms/step - loss: 0.0134 - val_loss: 0.0205\n",
      "Epoch 38/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0133 - val_loss: 0.0204\n",
      "Epoch 39/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0132 - val_loss: 0.0201\n",
      "Epoch 40/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0131 - val_loss: 0.0200\n",
      "Epoch 41/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0129 - val_loss: 0.0199\n",
      "Epoch 42/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0128 - val_loss: 0.0197\n",
      "Epoch 43/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0128 - val_loss: 0.0196\n",
      "Epoch 44/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0126 - val_loss: 0.0196\n",
      "Epoch 45/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0125 - val_loss: 0.0195\n",
      "Epoch 46/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0124 - val_loss: 0.0193\n",
      "Epoch 47/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0124 - val_loss: 0.0192\n",
      "Epoch 48/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0122 - val_loss: 0.0190\n",
      "Epoch 49/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0121 - val_loss: 0.0188\n",
      "Epoch 50/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0121 - val_loss: 0.0186\n",
      "Epoch 51/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0120 - val_loss: 0.0184\n",
      "Epoch 52/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0119 - val_loss: 0.0183\n",
      "Epoch 53/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0118 - val_loss: 0.0182\n",
      "Epoch 54/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0117 - val_loss: 0.0181\n",
      "Epoch 55/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0116 - val_loss: 0.0180\n",
      "Epoch 56/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0115 - val_loss: 0.0179\n",
      "Epoch 57/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0114 - val_loss: 0.0176\n",
      "Epoch 58/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0113 - val_loss: 0.0175\n",
      "Epoch 59/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0113 - val_loss: 0.0174\n",
      "Epoch 60/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0112 - val_loss: 0.0172\n",
      "Epoch 61/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0111 - val_loss: 0.0171\n",
      "Epoch 62/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0110 - val_loss: 0.0170\n",
      "Epoch 63/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0109 - val_loss: 0.0170\n",
      "Epoch 64/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0109 - val_loss: 0.0167\n",
      "Epoch 65/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0108 - val_loss: 0.0168\n",
      "Epoch 66/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0107 - val_loss: 0.0167\n",
      "Epoch 67/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0106 - val_loss: 0.0166\n",
      "Epoch 68/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0105 - val_loss: 0.0165\n",
      "Epoch 69/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0105 - val_loss: 0.0164\n",
      "Epoch 70/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0104 - val_loss: 0.0162\n",
      "Epoch 71/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0103 - val_loss: 0.0161\n",
      "Epoch 72/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0103 - val_loss: 0.0160\n",
      "Epoch 73/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0102 - val_loss: 0.0159\n",
      "Epoch 74/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0101 - val_loss: 0.0159\n",
      "Epoch 75/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0100 - val_loss: 0.0158\n",
      "Epoch 76/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0100 - val_loss: 0.0157\n",
      "Epoch 77/100\n",
      "4/4 [==============================] - 0s 16ms/step - loss: 0.0099 - val_loss: 0.0155\n",
      "Epoch 78/100\n",
      "4/4 [==============================] - 0s 17ms/step - loss: 0.0098 - val_loss: 0.0155\n",
      "Epoch 79/100\n",
      "4/4 [==============================] - 0s 19ms/step - loss: 0.0098 - val_loss: 0.0154\n",
      "Epoch 80/100\n",
      "4/4 [==============================] - 0s 14ms/step - loss: 0.0097 - val_loss: 0.0153\n",
      "Epoch 81/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0097 - val_loss: 0.0153\n",
      "Epoch 82/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0096 - val_loss: 0.0152\n",
      "Epoch 83/100\n",
      "4/4 [==============================] - 0s 10ms/step - loss: 0.0095 - val_loss: 0.0151\n",
      "Epoch 84/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0095 - val_loss: 0.0149\n",
      "Epoch 85/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0094 - val_loss: 0.0149\n",
      "Epoch 86/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0093 - val_loss: 0.0148\n",
      "Epoch 87/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0093 - val_loss: 0.0147\n",
      "Epoch 88/100\n",
      "4/4 [==============================] - 0s 14ms/step - loss: 0.0093 - val_loss: 0.0146\n",
      "Epoch 89/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0092 - val_loss: 0.0145\n",
      "Epoch 90/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0091 - val_loss: 0.0145\n",
      "Epoch 91/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0091 - val_loss: 0.0144\n",
      "Epoch 92/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0090 - val_loss: 0.0143\n",
      "Epoch 93/100\n",
      "4/4 [==============================] - 0s 11ms/step - loss: 0.0090 - val_loss: 0.0142\n",
      "Epoch 94/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0089 - val_loss: 0.0140\n",
      "Epoch 95/100\n",
      "4/4 [==============================] - 0s 14ms/step - loss: 0.0089 - val_loss: 0.0140\n",
      "Epoch 96/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0088 - val_loss: 0.0139\n",
      "Epoch 97/100\n",
      "4/4 [==============================] - 0s 14ms/step - loss: 0.0088 - val_loss: 0.0139\n",
      "Epoch 98/100\n",
      "4/4 [==============================] - 0s 18ms/step - loss: 0.0087 - val_loss: 0.0138\n",
      "Epoch 99/100\n",
      "4/4 [==============================] - 0s 13ms/step - loss: 0.0087 - val_loss: 0.0137\n",
      "Epoch 100/100\n",
      "4/4 [==============================] - 0s 12ms/step - loss: 0.0086 - val_loss: 0.0136\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEGCAYAAAB/+QKOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3ic9Zn/+/c9RTOjLqtYtmVbboArBhtTTG9rysZwQjGBBNgkHK6EQ/Zkl4UkG5ZlT7Lhl5xkw4YNsBAgQChxQuIE/0I2oSdgbLAJcQF3S67qfVTv3x/fR9JYHsmyrdHImvt1XXPNzPM8M/N9PLI++tZHVBVjjDGmL1+yC2CMMWZksoAwxhgTlwWEMcaYuCwgjDHGxGUBYYwxJq5AsgswVAoKCrS0tDTZxTDGmOPK+++/X6mqhfH2jZqAKC0tZc2aNckuhjHGHFdEZGd/+6yJyRhjTFwWEMYYY+KygDDGGBNXQvsgRGQJ8EPADzymqt/psz8E/BRYAFQB16vqDhG5Ebgr5tB5wKmquu5IPr+9vZ3y8nKi0eixnIYxRywcDlNSUkIwGEx2UYw5agkLCBHxAw8BlwDlwGoRWaGqG2IO+zxQo6rTRWQZ8AAuJJ4FnvXeZy7wqyMNB4Dy8nKysrIoLS1FRI71lIwZFFWlqqqK8vJypkyZkuziGHPUEtnEtAjYoqrbVLUNeB5Y2ueYpcBT3uPlwEVy6G/yG7zXHrFoNEp+fr6FgxlWIkJ+fr7VXM1xL5EBMQEoi3le7m2Le4yqdgB1QH6fY64HnjvaQlg4mGSwnzszGozoTmoROR1oVtW/9rP/NhFZIyJrKioqju5DOtqgfi902F97xhgTK5EBsRuYGPO8xNsW9xgRCQA5uM7qbssYoPagqo+q6kJVXVhYGHci4OF1dUDjPmgf+oCora3lv/7rv47qtZdffjm1tbWDPv6+++7je9/73lF9ViKVlpZSWVkJwFlnnRX3mFtuuYXly5cP+D5PPvkke/bs6Xn+hS98gQ0bNgzwisF58sknueOOO475fYwZjRIZEKuBGSIyRUTScL/sV/Q5ZgVws/f4GuBV9a5gJCI+4DqOsv9h0Hx+d6+dQ/7WAwVER0fHgK9duXIlubm5Q16mZPrzn/981K/tGxCPPfYYs2bNGopiGWP6kbCA8PoU7gBeATYCL6rqehG5X0Q+5R32OJAvIluArwL3xLzFuUCZqm5LVBkB8HkDuboG/oV9NO655x62bt3K/Pnzueuuu3j99dc555xz+NSnPtXzy+2qq65iwYIFzJ49m0cffbTntd1/ee/YsYOZM2fyxS9+kdmzZ3PppZfS0tIy4OeuW7eOM844g3nz5nH11VdTU1MDwIMPPsisWbOYN28ey5YtA+CNN95g/vz5zJ8/n1NOOYWGhoZ+3/fhhx/mrrt6Rx/H/vXd33nEyszMBNwonzvuuIMTTzyRiy++mAMHDvQcc//993PaaacxZ84cbrvtNlSV5cuXs2bNGm688Ubmz59PS0sL559/fs/SKs899xxz585lzpw53H333Qd93je+8Q1OPvlkzjjjDPbv3z/gv9uOHTu48MILmTdvHhdddBG7du0C4Oc//zlz5szh5JNP5txzzwVg/fr1LFq0iPnz5zNv3jw2b9484Hsbc1xS1VFxW7Bggfa1YcOGnsf3rfirXvfwn+PfHvy9XvfQ6/3v7+d234q/HvKZsbZv366zZ8/uef7aa69penq6btu2rWdbVVWVqqo2Nzfr7NmztbKyUlVVJ0+erBUVFbp9+3b1+/26du1aVVW99tpr9emnnz7ks/7lX/5Fv/vd76qq6ty5c/X1119XVdVvfvOb+pWvfEVVVceNG6fRaFRVVWtqalRV9corr9S3335bVVUbGhq0vb293/M5cOCATps2ref5kiVL9K233hrUeaiqZmRkqKrqL37xC7344ou1o6NDd+/erTk5Ofrzn//8oPdRVb3pppt0xYoVqqp63nnn6erVq3v2dT/fvXu3Tpw4UQ8cOKDt7e16wQUX6EsvvaSqqkDP6++66y79t3/7t0PO6YknntAvf/nLPf8WTz75pKqqPv7447p06VJVVZ0zZ46Wl5cf9O92xx136DPPPKOqqq2trdrc3HzIe8f+/BkzUgFrtJ/fqyO6k3r4CAzTtbkXLVp00Nj4Bx98sOcv3LKysrh/iU6ZMoX58+cDsGDBAnbs2NHv+9fV1VFbW8t5550HwM0338ybb74JwLx587jxxht55plnCARczWnx4sV89atf5cEHH6S2trZnezyFhYVMnTqVd999l6qqKjZt2sTixYsHfR7d3nzzTW644Qb8fj/jx4/nwgsv7Nn32muvcfrppzN37lxeffVV1q9f3+/7AKxevZrzzz+fwsJCAoEAN954Y8/5pqWlceWVVw7q3w3gnXfe4TOf+QwAn/3sZ3n77bd7/o1uueUW/vu//5vOTtcUeeaZZ/Ltb3+bBx54gJ07dxKJRAZ8b2OOR6NmNdfD+Ze/nd3/zv0bIBiBMYmf1JSRkdHz+PXXX+cPf/gD77zzDunp6Zx//vlxx86HQqGex36//7BNTP15+eWXefPNN/nNb37Dt771LT766CPuuecerrjiClauXMnixYt55ZVXOOmkk/p9j2XLlvHiiy9y0kkncfXVVyMigz6Pw4lGo3zpS19izZo1TJw4kfvuu++Y5hIEg8Ge4aZ+v/+w/T79efjhh1m1ahUvv/wyCxYs4P333+czn/kMp59+Oi+//DKXX345jzzyyEFBZ8xoYDUIcB3VXUPfSZ2VlTVgm35dXR15eXmkp6ezadMm3n333WP+zJycHPLy8njrrbcAePrppznvvPPo6uqirKyMCy64gAceeIC6ujoaGxvZunUrc+fO5e677+a0005j06ZNAP2GxNVXX82vf/1rnnvuuZ5+jCM9j3PPPZcXXniBzs5O9u7dy2uvvQbQEwYFBQU0NjYeNLKpv3/LRYsW8cYbb1BZWUlnZyfPPfdcT+3pSJ111lk8/7wbE/Hss89yzjnnALB161ZOP/107r//fgoLCykrK2Pbtm1MnTqVO++8k6VLl/KXv/zlqD7TmJEsZWoQA/IFoKt9yN82Pz+fxYsXM2fOHC677DKuuOKKg/YvWbKEhx9+mJkzZ3LiiSdyxhlnDMnnPvXUU9x+++00NzczdepUnnjiCTo7O7npppuoq6tDVbnzzjvJzc3lm9/8Jq+99ho+n4/Zs2dz2WWXUVlZifbT5JaXl8fMmTPZsGEDixYtOqrzuPrqq3n11VeZNWsWkyZN4swzzwQgNzeXL37xi8yZM4fi4mJOO+20ntfccsst3H777UQiEd55552e7ePGjeM73/kOF1xwAarKFVdcwdKlfSfsD85//ud/cuutt/Ld736XwsJCnnjiCQDuuusuNm/ejKpy0UUXcfLJJ/PAAw/w9NNPEwwGKS4u5utf//pRfaYxI5n094vgeLNw4ULte8GgjRs3MnPmzMO/uGYHtDXB2AGaoVLIb3/7W7Zt28add96Z7KIc1wb982dMEonI+6q6MN4+q0GAV4MY+iam41V3x64xJrVZHwS4PgjtHLaRTMYYczwY9QExqCY08WZTWy3CDJHR0nRrUtuoDohwOExVVdXh/7N2z6bWoZ9NbVKPeteDCIfDyS6KMcdkVPdBlJSUUF5ezmFXem1vgaYKqBIIhAY+1phB6L6inDHHs1EdEMFgcHBX9CpbDb+8Dm5cDjMuSXzBjDHmODCqm5gGLZLn7lsGv7y2McaMdhYQABFvWe2WmuSWwxhjRhALCICwBYQxxvRlAQHgD0Ao2wLCGGNiWEB0i+RaQBhjTAwLiG6RPIhaJ7UxxnSzgOgWybMahDHGxLCA6Ba2JiZjjIllAdHNahDGGHMQC4hu3QFhi6wZYwxgAdErkgddHe7CQcYYYywgevQst2HNTMYYAxYQvWy5DWOMOUhCA0JElojIxyKyRUTuibM/JCIvePtXiUhpzL55IvKOiKwXkY9EJLGL61sNwhhjDpKwgBARP/AQcBkwC7hBRGb1OezzQI2qTgd+ADzgvTYAPAPcrqqzgfOB9kSVFbCAMMaYPhJZg1gEbFHVbaraBjwPLO1zzFLgKe/xcuAiERHgUuAvqvohgKpWqWpirwfaHRA2m9oYY4DEBsQEoCzmebm3Le4xqtoB1AH5wAmAisgrIvKBiPxTvA8QkdtEZI2IrDnsVeMOx2oQxhhzkJHaSR0AzgZu9O6vFpGL+h6kqo+q6kJVXVhYWHhsnxiMgD9kAWGMMZ5EBsRuYGLM8xJvW9xjvH6HHKAKV9t4U1UrVbUZWAmcmsCyOjab2hhjeiQyIFYDM0RkioikAcuAFX2OWQHc7D2+BnhVVRV4BZgrIulecJwHbEhgWR0LCGOM6RFI1BuraoeI3IH7Ze8HfqKq60XkfmCNqq4AHgeeFpEtQDUuRFDVGhH5Pi5kFFipqi8nqqw9Inl2XWpjjPEkLCAAVHUlrnkodtu9MY+jwLX9vPYZ3FDX4RPJg9pdw/qRxhgzUo3UTurksCYmY4zpYQERyy47aowxPSwgYkVyob0JOtqSXRJjjEk6C4hYNpvaGGN6WEDEstnUxhjTI+UD4q+76zj923/gz1srLSCMMSZGygdEWsDH/vpWqpvaIGzXhDDGmG4pHxDZ4SAAdS3tkD7GbWyuSmKJjDFmZEj5gMiJuICob+mAzLFuY+P+JJbIGGNGhpQPiHDQR5rf52oQwQiEcqDBAsIYY1I+IESE7EjQBQRAVjE07E1uoYwxZgRI+YAAyI4EqO8JiLHWxGSMMVhAAK4foj7qBURmMTTsS26BjDFmBLCAwAXEwU1M+0A1uYUyxpgks4AgTkB0ttpyG8aYlGcBgZsL0RMQ3UNdbSSTMSbFWUDg9UG0tKOqrgYB0Gj9EMaY1GYBgQuILoXG1g7IGuc2Wke1MSbFWUDghrmCt9xGTxOTBYQxJrVZQNBnuY1QJqRl2lwIY0zKs4AAsiMxC/aBq0XYbGpjTIqzgKC3BtE71HWcjWIyxqQ8Cwh6l/w+eLkN64MwxqS2hAaEiCwRkY9FZIuI3BNnf0hEXvD2rxKRUm97qYi0iMg67/ZwIsuZk+4FxEHLbey32dTGmJQWSNQbi4gfeAi4BCgHVovIClXdEHPY54EaVZ0uIsuAB4DrvX1bVXV+osoXKzMtgE84eDZ1exO0NkA4eziKYIwxI04iaxCLgC2quk1V24DngaV9jlkKPOU9Xg5cJCKSwDLF5fPFWfIbbCSTMSalJTIgJgBlMc/LvW1xj1HVDqAOyPf2TRGRtSLyhoicE+8DROQ2EVkjImsqKiqOqbDZ4WBvH4TNhTDGmBHbSb0XmKSqpwBfBX4mIoe09ajqo6q6UFUXFhYWHtMHHrJgH1hAGGNSWiIDYjcwMeZ5ibct7jEiEgBygCpVbVXVKgBVfR/YCpyQwLLGDwgbyWSMSWGJDIjVwAwRmSIiacAyYEWfY1YAN3uPrwFeVVUVkUKvkxsRmQrMALYlsKwHB0QoGwIRq0EYY1JawkYxqWqHiNwBvAL4gZ+o6noRuR9Yo6orgMeBp0VkC1CNCxGAc4H7RaQd6AJuV9XqRJUVvMuORjvcExE3F8ICwhiTwhIWEACquhJY2WfbvTGPo8C1cV73C+AXiSxbXweNYgI3F8JGMRljUthI7aQedjmRIG0dXUTbO92GLLs2tTEmtVlAeLqX2zioo9pqEMaYFGYB4eld8jtmLkRrPbQ1JbFUxhiTPBYQnkNXdLW5EMaY1GYB4ek3IKyZyRiToiwgPN0XDTpoRVewCwcZY1KWBYSnpwbR3KcGUW8BYYxJTRYQnuywmxJS1+JNlovkuVvV5iSWyhhjkscCwhPw+8gMBXr7IESg4ESo+CS5BTPGmCSxgIiRHQ709kEAFJ4AlR8nr0DGGJNEFhAxDlluo+BEaK6CpqrkFcoYY5LEAiJGTt+AKDzR3VstwhiTgiwgYuREYq4qB1DgXYKiwgLCGJN6LCBiZPcNiJyJEEyHSuuoNsakHguIGIc0Mfl8kD/dahDGmJRkAREjJxKkqa2T9s6u3o2FJ0KlzYUwxqQeC4gY3bOpG7qvLAduJFPdLlvV1RiTciwgYmRHumdTx3ZUz3D3VoswxqQYC4gYh6zoCjFDXa2j2hiTWiwgYsQNiDHTQPzWUW2MSTkWEDEOuewoQCANxkyxyXLGmJRjARGjIDMEQGVDa58dtmifMSb1WEDEyE0Pkhbwsb8+evCOwhOgeit0tsd/oTHGjEIWEDFEhLHZIfb1DYiCE6GrA6q3J6dgxhiTBIMKCBH5iohki/O4iHwgIpcO4nVLRORjEdkiIvfE2R8SkRe8/atEpLTP/kki0igi/zjYEzpWxdlh9tXFqUGA9UMYY1LKYGsQf6eq9cClQB7wWeA7A71ARPzAQ8BlwCzgBhGZ1eewzwM1qjod+AHwQJ/93wf+9yDLOCTGZocPbWLqWbRv03AWxRhjkmqwASHe/eXA06q6PmZbfxYBW1R1m6q2Ac8DS/scsxR4ynu8HLhIRARARK4CtgPrB1nGIVGcHWZffRRV7d0YyoK8KbD3w+EsijHGJNVgA+J9Efk9LiBeEZEsoOswr5kAlMU8L/e2xT1GVTuAOiBfRDKBu4F/HegDROQ2EVkjImsqKioGeSoDK84JE23vor6l4+AdE06F3WuH5DOMMeZ4MNiA+DxwD3CaqjYDQeDWhJUK7gN+oKqNAx2kqo+q6kJVXVhYWDgkHzw2OwxwaEf1hAVQXw4N+4fkc4wxZqQbbECcCXysqrUichPwz7i/9geyG5gY87zE2xb3GBEJADlAFXA68L9EZAfw98DXReSOQZb1mBTn9BMQ409193s+GI5iGGNM0g02IH4MNIvIycA/AFuBnx7mNauBGSIyRUTSgGXAij7HrABu9h5fA7yqzjmqWqqqpcB/AN9W1R8NsqzHpNirQezvO5Jp3DwQH+y2gDDGpIbBBkSHul7bpcCPVPUhIGugF3h9CncArwAbgRdVdb2I3C8in/IOexzX57AF+CquGSupirLdbOpDRjKlZUDRLNj9fhJKZYwxwy8wyOMaRORruOGt54iID9cPMSBVXQms7LPt3pjHUeDaw7zHfYMs45AIBfyMyUg7tIkJYPwpsOm3oApyuEFcxhhzfBtsDeJ6oBU3H2Ifrj/huwkrVZLFnQsBbiRTSw3U7Bj2MhljzHAbVEB4ofAskCMiVwJRVT1cH8RxqzjechvgRjKBNTMZY1LCYJfauA54D9ccdB2wSkSuSWTBkmlsdph9da2H7iiaBYEw7LH5EMaY0W+wfRDfwM2BOAAgIoXAH3Czn0edsdlhqppaae/sIuiPyVB/EIrn2kgmY0xKGGwfhK87HDxVR/Da405xThhVOND3uhDgmpn2roPOjkP3GWPMKDLYX/K/E5FXROQWEbkFeJk+o5NGk+65EIes6gpuwlx7s63saowZ9QbVxKSqd4nIp4HF3qZHVfWlxBUrubqX2+h3JBO4Zqaxs4exVMYYM7wG2weBqv4C+EUCyzJi9Cy3Ea8GMWYahHOg7F049bPDXDJjjBk+AwaEiDQAGm8XoKqanZBSJVlef5ceBfD5YPrF8PHvoKsTfP7hL6AxxgyDAfsgVDVLVbPj3LJGazjAAJce7XbSldBcCbveHd6CGWPMMBq1I5GOVdxLj3abcQn4Q27ZDWOMGaUsIPrR73Ib4K4wN+0C2Oity2SMMaOQBUQ/4l56NNZJV0LdLrsMqTFm1LKA6Ee/lx7tduLl7voQ1sxkjBmlLCD6UdTfpUe7ZeTD5MWumckYY0YhC4h+FB8uIMA1M1VshMotw1QqY4wZPhYQ/SgeaDZ1t5lXuvtNvxmGEhljzPCygOhHcU4Yv08oq27u/6CcEig5Ddb9zEYzGWNGHQuIfqQFfEwak86WA40DH3jaF6HyE9j22vAUzBhjhokFxACmFWayteIwATH7KsgohFWPDE+hjDFmmFhADGB6USbbK5vo6Ozq/6BACBb+HXzyClRvG77CGWNMgllADGBaYQbtncqugfohwAWEzw/vPTY8BTPGmGFgATGA6UWZAIfvh8gqhtlXw9qnofUwxxpjzHEioQEhIktE5GMR2SIi98TZHxKRF7z9q0Sk1Nu+SETWebcPReTqRJazP9O6A+Jw/RAAp98OrfXw4XMJLpUxxgyPhAWEiPiBh4DLgFnADSIyq89hnwdqVHU68APgAW/7X4GFqjofWAI8IiKDvrjRUMkOBxmbHTp8DQKgZCFMWAhv/we0NSW+cMYYk2CJrEEsArao6jZVbQOeB5b2OWYp8JT3eDlwkYiIqjaravciSGHiX7RoWEwvymRrxSB/4f/Nt6C+HN74X4ktlDHGDINEBsQEoCzmebm3Le4xXiDUAfkAInK6iKwHPgJujwmMHiJym4isEZE1FRUVCTgFmF6YydYDjf2v6hpr0hkw/yZ450dQ8XFCymOMMcNlxHZSq+oqVZ0NnAZ8TUTCcY55VFUXqurCwsLChJRjelEmja0d7K9vHdwLLvlXSMuEl//BZlcbY45riQyI3cDEmOcl3ra4x3h9DDlAVewBqroRaATmJKykA5g22JFM3TIK4KJ7Ycdb8NHyBJbMGGMSK5EBsRqYISJTRCQNWAas6HPMCuBm7/E1wKuqqt5rAgAiMhk4CdiRwLL2q3eoa8PgX7TgFhh/KvzuHmjYn5iCGWNMgiUsILw+gzuAV4CNwIuqul5E7heRT3mHPQ7ki8gW4KtA91DYs4EPRWQd8BLwJVWtTFRZB1KYGSIrHBjcUNduPj9c9WNoa4Rffwm6BpiJbYwxI1RCh46q6kpgZZ9t98Y8jgLXxnnd08DTiSzbYIkI04syB9/E1K3oJLj0/4OV/wjvPQpn3J6YAhpjTIKM2E7qkWR64REMdY112hfghCXwP/fC/vVDXzBjjEkgC4hBmF6USUVDK3Ut7Uf2QhH41I8gnAMvfs76I4wxxxULiEEY9JpM8WQWwrVPQv0eePJyd2+MMccBC4hB6A6IrUcTEACli+GmX7oaxBOXQ23Z4V9jjDFJZgExCCV56USCftbvqTv6N5l8JnzuV9BcDf99Ibz7MLS3DF0hjTFmiFlADILfJyyaMoa3txzjSNuShXDrSig4AX53N/zHPPjzj6CjbWgKaowxQ8gCYpDOmVHA1oom9tYd41/9xXPg1pfhlpUwdjb8/hvwyLmw692hKagxxgwRC4hBOntGAQBvbR6i+Xqli12T0w0vuAl1P/kb+PWXoew9m1hnjBkRhv0aC8erE8dmUZgV4u3NlVy3cOLhXzDoN14CpWfD6/8Oqx6Btc9A1jg46QqYdiFMXgyR3KH7PGOMGSQLiEESEc6eXsCbn1TQ1aX4fDJ0bx7KdNeSOO+f4JPfw8YVsO5nsPoxEJ9b1+nUz8G86yAYGbrPNcaYAVhAHIGzpxfw0trdbNxXz+zxOUP/AeEcmHetu3W0Qflq2P4GbHoZfnMn/PFf3UKApedA0UzIHOsm4xljTAJYQByB7n6ItzdXJiYgYgXSXD9F6WI4/2uw421498fw1vfhrf/fHRPJg6kXwMy/hRmXQCgrsWUyxqQUC4gjMDY7zAljM3l7SyX/93nThu+DRWDKOe7WVOnWdarYBHv/AptfgfW/BH/IhcScT7v1n9LSh698xphRyQLiCJ09vZBnV+0k2t5JOOgf/gJkFMDU89wNoKsTylbBhhWw/iXY9FsIZsDsq2DBrW7uhTVDGWOOggXEETpnRgE/+dN21uyo6WlySiqfHyaf5W5/8y3Y+Sf4y4suLNY9C0WzYcq5kJEP6QUwZgqMm28jo4wxh2UBcYROnzqGoF9445MDIyMgYvn8LgymnAtL/t1d8nTt027obFufK+KNmQYTFsCk02Hi6VA0y73eGGM8FhBHKD0twLkzClnx4R7uuWwm/qEc7jqUQlmw8FZ3A2iPQnMVVH4Muz+APWvdCKmPXnT707KgZAGULHKzvSN5blRV1jjILEreeRhjksYC4ihcu7CEPz5zgDc3V3DBicfJL89gGHImuNu0C902VajdCbtWuX6M8vfgre+B9pnJXXCCGy019TyYeIZrrjLGjHoWEEfhwpPGMiYjjZ+vKTt+AiIeEcgrdbeTr3fbWhuhehu01kO0zj3e9oZrqnrvEXdMwQmuWWr8Ka4/Y+wsm8BnzChkAXEU0gI+rpo/gWfe3UlNUxt5GWnJLtLQCWXCuHkHbzvr/3ET93a/D7vecQsLbvqtCw0A8bs+jAleYOROhuxxrnkqkmejqIw5TllAHKVrF5bwkz9t59frdnPL4inJLk7iBdLcNS0mn+meq0JdGez9EPasgz0fwMbfwAc/Pfh1oRzInwpjproax6SzXAD5g8N/DsaYI2IBcZRmjstmzoRsXlxTnhoB0ZcI5E5yt5l/67apQl051O+Ghr3u8qrV26F6K5Sthr/+wh0XiEDeZMgodMuFjJnqOsbHznHNXTaaypgRwQLiGFy3cCL3/no96/fUJX7pjeOBCOROdLd4Gva55qmyVa720Vjhmq3W/7K3Y9wXcE1T2RPc9TJOvNzNIA+Ehu88jDEAiKomuwxDYuHChbpmzZph/cza5jYWfeuPXLOwhG9fPXdYP3tUaW+BAxth/19djaN+j6uJ7PkA2pshLROmng9TvBnkBSdYv4YxQ0RE3lfVhfH2JbQGISJLgB8CfuAxVf1On/0h4KfAAqAKuF5Vd4jIJcB3gDSgDbhLVV9NZFmPRm56GtedVsLPVu3ihtMmMbfEahFHJRiBCae6W6z2aO9qtltfcx3j4GaElyx0E/2K57oaR1axa7Ky5iljhkzCahAi4gc+AS4ByoHVwA2quiHmmC8B81T1dhFZBlytqteLyCnAflXdIyJzgFdUdcJAn5eMGgRAXUs7F3//DcZmh/jVlxYT8NtF+hKmZocbclu2CsrXQOUnQMzPbzDdLYU+/WI3mzx3ki1aaMxhJKsGsQjYoqrbvEI8DywFNsQcsxS4z3u8HPiRiIiqro05Zj0QEZGQqrYmsLxHJScS5L6/nc2Xf/YBT/55B184Z2qyizR65ZXCglJYcLN7Hq2Dys2ub6NhL8eo2psAABXcSURBVFR8DFv/6Fa47dY9GzzbmySYMxEKZkDBiZA/zfo2jBlAIgNiAlAW87wcOL2/Y1S1Q0TqgHwg9sLPnwY+iBcOInIbcBvApEmThq7kR+jyucVceFIR3/+fT7hs7jgm5NqksWERznFNTX1Vb3PX9q7f7cKjfo97vO8v0FTRe5z4ekdSZY51CxiGsiCUDTklbnRV/jQXKtZ0ZVLQiB7FJCKzgQeAS+PtV9VHgUfBNTENY9EOIiLcv3Q2l3z/Tf7hxXU89XeLCAXsF0rSjPHmXcTT1gxVm6HiE3ffsBca9kPjfqja4s0gr4eu9t7X+IKuuWrMFBcW3X0eRTPd3A6b02FGqUQGxG4gdrxjibct3jHlIhIAcnCd1YhICfAS8DlV3ZrAcg6Jkrx0vnX1HL764of8/fPr+M8bTrH+iJEoLR3Gnexu/VF1wVG9Daq2uvua7W6E1Z510BxTwU3Lclf9m3Smm8cxdrYLDxtlZUaBRAbEamCGiEzBBcEy4DN9jlkB3Ay8A1wDvKqqKiK5wMvAPar6pwSWcUj9X6eWUNvczv2/3cDXX/qIBz49D7FfFMcfEcge726lZx+6v6MNGve5VXG3v+E6zj/5Xe/+jCKYuAgmneEWNxw3z/o6zHEpYQHh9SncAbyCG+b6E1VdLyL3A2tUdQXwOPC0iGwBqnEhAnAHMB24V0Tu9bZdqqoHElXeofJ3Z0+htqWdB/+4mVDAzzevnEVawGoSo0ogrXcW+eyr3LaWGti/wc3l2P0BlL3bOyzXH4Lx892Q3Eie6+OI5Ln1qrJL3H0o22odZsSxiXIJoKr8+//exKNvbmP2+Gx+uGw+04uykl0sM9wa9rkhuWXvQflqdx3x1oZDl1MH8Ke5+R2ZhZA/HQpPchMC80rdsiThXAsQkxADDXO1gEig36/fxz2//Iim1g7uXnISnztzsvVLpDpVaGuE5mrXz1FX7u6bKqCpyjVdVX4CtbsOfl0ox62WO7m7v2O2rZRrhoQFRBIdaIhy9/K/8NrHFcwoyuQbV8zk/OP5GhJmeLQ1uVFVNTtdWFRvdTWR/evpmRyYluVqGGOm9A7JzZ/hah52USczSBYQSaaq/M+G/Xx75UZ2VDVz5tR8Pr2ghEtnjyU7bEMkzRFornbNVVVb3czy7tFVNTsOHpqbnu+uOz5miguRrGLXzxHOcc1VGfmuSSuUZbWQFGcBMUK0dXTx03d28MSfdrC7toU0v4+zZxSwsDSPUybmMa8kh4zQiJ6aYkaqzg63Qm7VVnfd8YqPveG5O1wzFv38Pw/nuhFXExe52eUdrdDe5PpJMsf2zkLPLLIgGaUsIEYYVWVtWS2//XAvr318gO2VTQD4BE4Ym8Upk/I4ZWIus8Znc8LYLBsFZY5NRys0V7kJgK0NbsRVcyU0VbrJgmXvuQ70gQTTey9Pm1EAkTHuPq/UNW/lTbF1r45TFhAjXE1TG+vKa1m7q5Z1ZbWs3VVDQ7QDgKBfmFaYybSiTKYVZDC1MJMZYzOZXpRps7XN0GmpcX0dwXR3Aze7vGEv1O3ubc6q2enCprnq4CYtcDWN7n6Q/OnebZpr7gpl2XIlI5QFxHGmq0vZXtXEhj31bNhbz8a99WyvbKKsupku7+vy+4QpBRnMKMpkWmEmUwszGJ8bYWx2mKKskDVVmcRShWitC47qbVC1zXWqV22Gyi3QWnfoa4LpvTWPjMLemkh67LZCL1C8/pLAKLre+whlATFKtHZ0srOqmY/3NfDJ/gY27Wtga0UjO6ua6ew6+HvMTQ8yaUw6k8akc1Kxa7aaV5JDlnWKm0RTdTWMqq1u9FVLjWvaita7x00V0HTAdbg3V7s+j/5ExrhlUcbPd30koSxIy4Bwdu8ii7YW1jGxgBjl2jq62FXdzL66KAcaouyvb2V3bTM7q5rZVe3uwfUxTh6TTmlBBqX5GZTmpzPZe1ySFyFoczRMMrRHXaA0Vbhbc7W3aGKta9La+yEc2ABdHfFfn1HkJhPmTnar8AbTXc3DH3IXo0rLdKESyXVzRyJjrNM9RtKuKGeGR1rAx/Qi1y8RT11LOx+W1fLBrho2729ke2UTq7dX09TW2XOM3ydMyI0wOT+dyfnpjM+NMCE3QklehMn5GeRnpNm6UiYxgmHvWh0DXBOso9WNxmprcrdorZup3rjfba/d5Yb/bvhV/0ESKzLGrcQ7/hTXyZ5V7EZsRXJdmAQzwG+/Hq0GkaJUlcrGNnZWNbG9somdVc3srG5mZ1UTu6qbqW0+uAMyKxRgUn46E3IjPeFRmBWiIDNEUXaI0vwMG21lRoauTuhsg46oW969vdlr4qrzmrgq3bVB9qx110LXzvjvExnj1tvKm+zWzMosck1aGYWQnufVRvLcLHff8fuzbzUIcwgRoTArRGFWiIWlYw7Z39TawZ7aFspqmtlR6YJjZ3UzO6qa+NOWyoNqHwABnzC9KJMTi7OYnJ/BxLwIk8akMyEvQnF22JYYMcPH5wdfxDUvRfIGPrY96l0TZB807HH9JO3N0Nrolj2p2elmr2/+wwB9JeI61dPHuJFc2eN7r16YO8k9D+e62kkw/bhq2rIahDliqkp9tIPKxlYqGlrZVxft6TT/eF8De+paiP2x8gkUZ4cpyUtnotdxPj43zPjcCONywhRlh8lI81sTlhnZWhtdk1ZTpauJtFS7+2iduzVVQP1ed/XC+j2HDgMGCITdOlrjTu5dT6u7j6R7iHFaumviSkt3xyf4/4V1Upth1dbRxZ7aFnZVN7OntoU9tS2U17ZQXu227auPHvKaUMBHQWaIcTlhJuWnM3lMBpPyI0z0QqUwM4TPZwFijhNdnS5Masu8mkkdtNS6bfs+ch3vrfWHfx/xeUODvWHBuZN6r5iYO9nVTjKLjmmOiQWEGVGi7Z3sr4+ypzbK3roWKhtbqWpso6KxlT21LeysamZv3cEhkub3MT43zIQ81/9RnONqH8U5YUq8fhGb+2GOG11drmmrrdHdWhtd01Z7c2+/SXeHfOwIr5qdrukrli8A85bBVQ8dVVGsD8KMKOGgn8n5GUzOz+j3mGh7J+U1rg+kvLqZ8toWdte0sLu2hdc/rqCisZW+f9vkRIIUZ4cpyg4xNjvM2J77MJPzXdNWepr9yJsRwOcbeNTWQNqa3AKNdeVQX+6as/KnD235PPa/xYxI4aB/wKG77Z1dVDS0sreuhXIvOPbUtnCgvpX9Da1s3l9JRWPrIRMIi7JCjM+NMD43THF2hDEZQXIiQXLS05iQG2bSmAwKMm1IrxnB0jKgeI67JZgFhDkuBf0+7xd9hAWT4x/T2aVUN7Wxt66lZ9Lgjsom9tZF2bSvgdc/rqC57dAhjulpfopzwozNcrWQ4hwXKONyIuRnppEddqGSlx600VlmVLOAMKOW39c7lHdeSW7cY1o7Oqlraae2uZ3dNS09w3n310c5UN/Kmp017K/fS3vnoX11fp8wLidMSZ4LquJs1ydSlBUmPzONMRlpFGSGyA4HrEZijksWECalhQJ+irL8FGWFOWFs/OuGd3UplU2t7KmNUtPURn20nbqWdg7Ut1Je00x5TQurtlWzvz5KR9ehQRIJejUSr0+kKMu7zw5T7PWVFGWFiaTZaqdmZLGAMOYwfD6hKMvVDAbSHSQH6lupamqjuqmVyoY29tVH2VcfZX9dlLW7atlfH6W1o+uQ12eGAq7GkxnyZqmnkRUOkhkOkBkK9PSfFOeEbekTMywsIIwZIoMNElWlvqWD/Q1R9tW58KhocJMOKxpbqWxoZeO+eiobWmls7SBOpYS0gI9xOWHG5YQpyAyRl55GXnqQwqxQT82kKDtEfkbIlkAxR80CwphhJiLkpAfJSQ/226zVTVVpae+kvqWDAw1R9tZF2Vvbwt66KHvqouypbWH9nnpqmtuoa2k/ZOgvQHY4QH5miDEZ3f0iaRRmuaaugsw0MkOulpIVDpAbcR3w1vluIMEBISJLgB8CfuAxVf1On/0h4KfAAqAKuF5Vd4hIPrAcOA14UlXvSGQ5jRmpRIT0tADpaQGKc8LMK+n/2M4upaqplf11rT21kqrGVjcRsamN6qY2yqqbWburlqqmQ+eRxMoOByjwQiU/M83VTLxQKcrufZyTHrQrG45iCQsIEfEDDwGXAOXAahFZoaobYg77PFCjqtNFZBnwAHA9EAW+CczxbsaYw/DHNHHNJWfAYzs6u6hsbKOqqZWm1k6aWjuoj7ZT09RGTXM7Nc1tLlQa29hW0cSq7dWHrPDbLT3N7+aSRIJke/cFmWmuLyU7TE4kSFY4QHY4SG56kPwMN1TYlk4Z+RJZg1gEbFHVbQAi8jywFIgNiKXAfd7j5cCPRERUtQl4W0QSMz3QmBQX8Pso9pYqGazWjk4qGlo50OA64isaW6lrbqO2uZ3alnbqW9qpj7YPqpbi9wk5ERcYuZGg60PJcP0ouelpZEeCZHvNXhlpATJC7rGN9hpeiQyICUBZzPNy4PT+jlHVDhGpA/KBysF8gIjcBtwGMGnSpGMtrzFmAKGAn5K8dEry0gd1fEdnF9VNrm+kPtpBQ9TNN6n2mrtqmtuobWmnttmN9Nq4t56qpra4I7xiZYfdaK/c9LSePpPsmNpLdjjg1VqChII+0vw+QgEfuemuD8ZvNZdBO647qVX1UeBRcIv1Jbk4xpgYAb+PIm++x5GItndSH22nvsWFSktbJ01tndQ2t3Gge7RXQyu1LW3sb3Cz4uuj7TRED38lOZ9AXnoameEAkaCfSJqf/IwQxTkhirPD5HnNX1nhQE8tJjscJD0UIBTwEfBJSg0vTmRA7AYmxjwv8bbFO6ZcRAJADq6z2hiTosJBP+Ggn6KBB3gdorNLafCCpT7qmrtaO7po7+gi2tFFTVMblY2tVDa20dzWQUtbJ81tnZTXNLNmZ/99LLF84srnBg74yY4EGJvVO6Q4OxIgywuY7qax9DQ/maHeZrJQwHfchEwiA2I1MENEpuCCYBnwmT7HrABuBt4BrgFe1dGy/rgxZlj5feKandLTjur10fZOapvbXch4TWL10Q7qWtqJtnUSbe8k2tFJtL2L5rZOWto6qG1pZ29dlA/L66huao07Z6WvgE/ICvcGSffjno7+cJCMkKvdpKf5yQp1N6EFSPP78IngEyE74prSEhk2CQsIr0/hDuAV3DDXn6jqehG5H1ijqiuAx4GnRWQLUI0LEQBEZAeQDaSJyFXApX1GQBljzJAJB/0U5/iPqOM+lqrS1NZJg9fc1dTaQXNbJ42t7nFTawcNrR00Rjto8AKowXtcVt3M+hbX2R9vAcn+pPl9FGaFuGxOMf985ayjKvdAEtoHoaorgZV9tt0b8zgKXNvPa0sTWTZjjBlKIkJmyC2LMm7gUcYDau/soqW9s6cJrLvZrK6lnY6uLrpU6ezCrQfW4Oa7jMuNDN2JxDiuO6mNMWa0Cfp9BP0+ssPBZBcFm09vjDEmLgsIY4wxcVlAGGOMicsCwhhjTFwWEMYYY+KygDDGGBOXBYQxxpi4LCCMMcbEJaNl6SMRqQB2HsNbFDDIZcZHkVQ8Z0jN87ZzTh1Het6TVbUw3o5RExDHSkTWqOrCZJdjOKXiOUNqnredc+oYyvO2JiZjjDFxWUAYY4yJywKi16PJLkASpOI5Q2qet51z6hiy87Y+CGOMMXFZDcIYY0xcFhDGGGPiSvmAEJElIvKxiGwRkXuSXZ5EEJGJIvKaiGwQkfUi8hVv+xgR+R8R2ezd5yW7rIkgIn4RWSsiv/WeTxGRVd53/oKIHN1FjEcoEckVkeUisklENorImanwXYvI/+v9fP9VRJ4TkfBo/K5F5CcickBE/hqzLe73K86D3vn/RUROPZLPSumAEBE/8BBwGTALuEFEhv7CrsnXAfyDqs4CzgC+7J3nPcAfVXUG8Efv+Wj0FWBjzPMHgB+o6nSgBvh8UkqVOD8EfqeqJwEn4859VH/XIjIBuBNYqKpzAD/uGvej8bt+EljSZ1t/3+9lwAzvdhvw4yP5oJQOCGARsEVVt6lqG/A8sDTJZRpyqrpXVT/wHjfgfmFMwJ3rU95hTwFXJaeEiSMiJcAVwGPecwEuBJZ7h4yq8xaRHOBc4HEAVW1T1VpS4LvGXUI5IiIBIB3Yyyj8rlX1TaC6z+b+vt+lwE/VeRfIFZFxg/2sVA+ICUBZzPNyb9uoJSKlwCnAKmCsqu71du0DxiapWIn0H8A/AV3e83ygVlU7vOej7TufAlQAT3jNao+JSAaj/LtW1d3A94BduGCoA95ndH/Xsfr7fo/pd1yqB0RKEZFM4BfA36tqfew+deOdR9WYZxG5Ejigqu8nuyzDKACcCvxYVU8BmujTnDRKv+s83F/LU4DxQAaHNsOkhKH8flM9IHYDE2Oel3jbRh0RCeLC4VlV/aW3eX93ddO7P5Cs8iXIYuBTIrID13x4Ia59PtdrhoDR952XA+Wqusp7vhwXGKP9u74Y2K6qFaraDvwS9/2P5u86Vn/f7zH9jkv1gFgNzPBGOqThOrVWJLlMQ85rd38c2Kiq34/ZtQK42Xt8M/Dr4S5bIqnq11S1RFVLcd/tq6p6I/AacI132Kg6b1XdB5SJyInepouADYzy7xrXtHSGiKR7P+/d5z1qv+s++vt+VwCf80YznQHUxTRFHVbKz6QWkctx7dR+4Ceq+q0kF2nIicjZwFvAR/S2xX8d1w/xIjAJt1T6darat/NrVBCR84F/VNUrRWQqrkYxBlgL3KSqrcks31ASkfm4Tvk0YBtwK+6PwVH9XYvIvwLX40btrQW+gGtvH1XftYg8B5yPW9Z7P/AvwK+I8/16YfkjXHNbM3Crqq4Z9GelekAYY4yJL9WbmIwxxvTDAsIYY0xcFhDGGGPisoAwxhgTlwWEMcaYuCwgjBkBROT87tVmjRkpLCCMMcbEZQFhzBEQkZtE5D0RWScij3jXmmgUkR941yL4o4gUesfOF5F3vXX4X4pZo3+6iPxBRD4UkQ9EZJr39pkx13F41pvkZEzSWEAYM0giMhM3U3exqs4HOoEbcQvDrVHV2cAbuJmtAD8F7lbVebhZ7N3bnwUeUtWTgbNwq4+CW2X373HXJpmKW0vImKQJHP4QY4znImABsNr74z6CWxStC3jBO+YZ4JfedRlyVfUNb/tTwM9FJAuYoKovAahqFMB7v/dUtdx7vg4oBd5O/GkZE58FhDGDJ8BTqvq1gzaKfLPPcUe7fk3sGkGd2P9Pk2TWxGTM4P0RuEZEiqDnOsCTcf+PulcM/QzwtqrWATUico63/bPAG94V/cpF5CrvPUIikj6sZ2HMINlfKMYMkqpuEJF/Bn4vIj6gHfgy7qI8i7x9B3D9FOCWXX7YC4DuVVXBhcUjInK/9x7XDuNpGDNotpqrMcdIRBpVNTPZ5TBmqFkTkzHGmLisBmGMMSYuq0EYY4yJywLCGGNMXBYQxhhj4rKAMMYYE5cFhDHGmLj+D0T7bSK2/FQNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Many-to-one LSTM \n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, GRU\n",
    "from keras.layers import Dense\n",
    "from numpy import array  \n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1)\n",
    "\n",
    "######################################################################################################\n",
    "model = Sequential()\n",
    "model.add(GRU(10, input_shape = (2, 1)))           ## (3 GRU's cells , 1 feature) \n",
    "model.add(Dense(1))                                ##  output y \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "print(model.summary())\n",
    "########################################################################################################\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=100)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train loss, validation loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "POgcpPDKWadV",
    "outputId": "e627edc6-70ec-4128-97df-5dee72d485d6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input :  [[[498.06949807]\n",
      "  [581.08108108]]]\n",
      "Predict output :  [489.7979]\n",
      "Actual Output:  604.2471042471043\n"
     ]
    }
   ],
   "source": [
    "x_test = X_test[1]\n",
    "x_test = x_test.reshape(1, 2, 1)\n",
    "\n",
    "testPredict = model.predict(x_test)\n",
    "print('input : ', x_test*1000)\n",
    "print('Predict output : ', testPredict.reshape(-1)*1000)\n",
    "print('Actual Output: ', Y_test[1]*1000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tGpc3GC57RyU"
   },
   "source": [
    "2. From stock price prediction, modify the above code to predict the PPT's next-day closed price using the past 30 days of \"open\", \"closed\" and \"volume\".\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "dHd_zhFgnm3j"
   },
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "\n",
    "def generate_dataset(timesteps, train, test):\n",
    "    X_train = []\n",
    "    Y_train = []\n",
    "    X_val = []\n",
    "    Y_val = []\n",
    "    for i in range(timesteps,train.shape[0]):\n",
    "        X_train.append(train[i-timesteps:i])\n",
    "        Y_train.append(train[i][3])\n",
    "    X_train,Y_train = np.array(X_train),np.array(Y_train)\n",
    "  \n",
    "    for i in range(timesteps,test.shape[0]):\n",
    "        X_val.append(test[i-timesteps:i])\n",
    "        Y_val.append(test[i][0])\n",
    "    X_val,Y_val = np.array(X_val),np.array(Y_val)\n",
    "    return X_train, Y_train, X_val, Y_val\n",
    "\n",
    "sc = MinMaxScaler()\n",
    "dataset = sc.fit_transform(stock_data)\n",
    "train = dataset[:2000]\n",
    "test = dataset[2000:]\n",
    "\n",
    "timesteps = 50\n",
    "\n",
    "X_train, Y_train, X_test, Y_test = generate_dataset(50, train, test)\n",
    "print(X_train.shape, Y_train.shape)\n",
    "print(X_test.shape, Y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "kSf_BG7-n0gx"
   },
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import GRU\n",
    "from keras.layers import Dense\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "model = Sequential()\n",
    "model.add(GRU(50, input_shape = (50, 5)))          \n",
    "model.add(Dense(1))                               \n",
    "\n",
    "model.compile(loss='mean_squared_error', optimizer='sgd')\n",
    "print(model.summary())\n",
    "\n",
    "history = model.fit(X_train, Y_train, validation_split = 0.1, epochs=50, verbose=1)\n",
    "\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train_loss, val_loss'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "_8Oyz2FrldEF",
    "outputId": "2bfdceed-2898-4f85-9e6f-03e223ddfa72"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1.44000006, 1.44000006, 1.45000005, 1.47000003, 1.46000004,\n",
       "       1.42999995, 1.44000006, 1.45000005, 1.47000003, 1.44000006,\n",
       "       1.45000005, 1.44000006, 1.42999995, 1.46000004, 1.44000006,\n",
       "       1.41999996, 1.42999995, 1.40999997, 1.38999999, 1.35000002,\n",
       "       1.38      , 1.37      , 1.33000004, 1.35000002, 1.38999999,\n",
       "       1.40999997, 1.41999996, 1.38999999, 1.38      , 1.37      ,\n",
       "       1.13999999, 1.01999998, 0.91000003, 0.85000002, 0.81      ,\n",
       "       0.77999997, 0.81      , 0.76999998, 0.76999998, 0.81999999,\n",
       "       0.80000001, 0.80000001, 0.77999997, 0.76999998, 0.76999998,\n",
       "       0.88      , 0.99000001, 0.95999998, 0.93000001, 0.95999998,\n",
       "       0.94999999, 0.95999998, 0.93000001, 0.92000002, 0.94      ,\n",
       "       1.00999999, 1.        , 0.99000001, 1.00999999, 1.        ,\n",
       "       0.97000003, 0.98000002, 0.98000002, 0.95999998, 0.91000003,\n",
       "       0.92000002, 0.92000002, 0.88999999, 0.88      , 0.91000003,\n",
       "       0.93000001, 0.92000002, 0.91000003, 0.89999998, 0.88      ,\n",
       "       0.85000002, 0.82999998, 0.85000002, 0.85000002, 0.82999998,\n",
       "       0.81999999, 0.80000001, 0.79000002, 0.77999997, 0.77999997,\n",
       "       0.79000002, 0.76999998, 0.76999998, 0.79000002, 0.79000002,\n",
       "       0.81999999, 0.94999999, 1.01999998, 0.93000001, 0.92000002,\n",
       "       0.86000001, 0.88      , 0.91000003, 0.88999999, 0.89999998,\n",
       "       1.00999999, 0.95999998, 0.98000002, 1.05999994, 1.10000002,\n",
       "       1.04999995, 1.01999998, 1.08000004, 1.09000003, 1.09000003,\n",
       "       1.08000004, 1.05999994, 1.03999996, 0.97000003, 0.94      ,\n",
       "       0.93000001, 0.95999998, 0.97000003, 0.98000002, 0.94999999,\n",
       "       0.94      , 0.95999998, 0.94      , 0.88999999, 0.89999998,\n",
       "       0.88999999, 0.91000003, 0.92000002, 0.93000001, 0.93000001,\n",
       "       0.89999998, 0.88      , 0.88      , 0.92000002, 0.88      ,\n",
       "       0.98000002, 0.94999999, 0.91000003, 0.91000003, 0.93000001,\n",
       "       0.93000001, 0.91000003, 0.91000003, 0.91000003, 0.92000002,\n",
       "       0.88999999, 0.85000002, 0.86000001, 0.88999999, 0.89999998,\n",
       "       0.88      , 0.98000002, 0.95999998, 1.03999996, 1.01999998,\n",
       "       1.05999994, 1.01999998, 1.01999998, 1.10000002, 1.42999995,\n",
       "       1.49000001, 1.52999997, 1.72000003, 1.72000003, 1.94000006,\n",
       "       2.0999999 , 2.0999999 , 1.87      , 1.80999994, 1.88999999,\n",
       "       1.64999998, 1.72000003, 1.78999996, 1.83000004, 1.78999996,\n",
       "       1.82000005, 1.72000003, 1.75999999, 1.75999999, 1.69000006,\n",
       "       1.74000001, 1.74000001, 1.74000001, 1.75      , 1.73000002,\n",
       "       1.74000001, 1.75999999, 1.72000003, 1.95000005, 1.82000005])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "Y_hat = model.predict(X_test)\n",
    "\n",
    "print(' MSE : ', mean_squared_error(Y_hat, Y_test))\n",
    "plt.plot(Y_test,c = 'r')\n",
    "plt.plot(Y_hat,c = 'y')\n",
    "plt.xlabel('Day')\n",
    "plt.ylabel('Price')\n",
    "plt.title(\"Stock Price Prediction\")\n",
    "plt.legend(['Actual','Predicted'],loc = 'lower right')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "RNN_LAB.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}