tensorflow_speakerid.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lR2Rvs6NVAuS"
      },
      "source": [
        "# Speaker Identification"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import math\n",
        "import random\n",
        "import numpy as np\n",
        "import librosa\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "fivevgg = [0.94447660,0.95706898,0.95934319,0.94880545,0.32394856,0.96585649,0.95991272,0.95634270,0.96104479,0.90679061,0.92205840,0.95092440,0.92154872,0.95683944,0.95309526,0.94056165,0.94644040,0.90275872,0.96498823,0.93177980,0.93074518,0.94423437,0.96645796,0.92889559,0.95119870]\n",
        "tenvgg = [0.90047634,0.85916662,0.88760287,0.91713089,0.88994265,0.91689712,0.92651540,0.91207546,0.91933054,0.90255326,0.93229455,0.88642770,0.91912723,0.92506468,0.94099718,0.90749002,0.90065914,0.91762137,0.87230527,0.87131679,0.89855266,0.87554008,0.92138267,0.89781547,0.91194147]\n",
        "twentyvgg = [0.54047787,0.83028978,0.86905277,0.84565496,0.50137204,0.84981763,0.85792303]\n",
        "fivecnn = [76.5886287625418, 74.32216905901116, 86.10271903323263]\n",
        "tencnn = [63.9344262295082, 61.40295690468701, 51.57248157248157, 60.14833127317676]\n",
        "twentycnn = [35.53467808786958]\n",
        "import statistics as s\n",
        "avgvgg = [s.mean(fivevgg) * 100, s.mean(tenvgg) * 100, s.mean(twentyvgg) * 100]\n",
        "avgcnn = [s.mean(fivecnn), s.mean(tencnn), s.mean(twentycnn)]\n",
        "fig, axs = plt.subplots()\n",
        "plt.plot([5,10,20], avgvgg)\n",
        "plt.plot([5,10,20], avgcnn)\n",
        "plt.plot([5,10,20], avgvgg, '*', color='steelblue', markersize=10)\n",
        "plt.plot([5,10,20], avgcnn, 'o', color='orange')\n",
        "plt.ylim(0,100)\n",
        "# axs.set_title('Accuracy of VGGish Embeddings with SVM classifier')\n",
        "axs.set_title('Accuracy by number of speakers')\n",
        "axs.legend(['VGGish Embeddings', 'CNN'])\n",
        "axs.set_ylabel('Accuracy')\n",
        "axs.set_xlabel('Number of Speakers')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 490
        },
        "id": "_0S5qmLd3tB2",
        "outputId": "de5843fb-5bc2-4620-a43e-90be379021ea"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Number of Speakers')"
            ]
          },
          "metadata": {},
          "execution_count": 2
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjsAAAHHCAYAAABZbpmkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAByHElEQVR4nO3deXhMZ/8G8Hsm+77JShJBSEQEQRBrhWgttS9Vsb0oYqfq7WutpbTVokj1p2jFUkqr2tprLbHGUpEQsSexZt9nzu+Pk5lkZCeTSSb357rmavOcM2e+J03l9iznkQiCIICIiIhIS0k1XQARERGROjHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBEAYPPmzZBIJLh48aKmS6lwCxYsgEQiwfPnzzVdSqlcuHABbdq0gYmJCSQSCcLDwzVd0hvp2LEjGjVqpOkyqBpg2CGts27dOkgkEvj5+Wm6FKJyl52djQEDBuDly5f4+uuv8dNPP8HV1VXTZRFVarqaLoCovIWGhqJ27do4f/487ty5g3r16mm6JKJyEx0djfv37+P777/Hf/7zH02XQ1QlsGeHtEpMTAz++ecfrFy5Era2tggNDdV0SUVKTU3VdAlUwdLS0t76Gk+fPgUAWFpavvW1tB3/HyMFhh3SKqGhobCyskL37t3Rv3//IsNOQkICpk2bhtq1a8PAwAC1atVCUFCQypyNjIwMLFiwAPXr14ehoSEcHR3Rt29fREdHAwCOHz8OiUSC48ePq1z73r17kEgk2Lx5s7JtxIgRMDU1RXR0NN577z2YmZlh6NChAIBTp05hwIABcHFxgYGBAZydnTFt2jSkp6cXqPvWrVsYOHAgbG1tYWRkhAYNGuDTTz8FAPz999+QSCTYu3dvgfdt27YNEokEZ8+eLfF7mJaWhnHjxsHGxgbm5uYICgrCq1evlMeHDx+OGjVqIDs7u8B7u3btigYNGhR7fcU8jZs3b6JTp04wNjZGzZo1sWLFCpXzFHOI7t27p9Je2Pddcc1r166hQ4cOMDY2Rr169bB7924AwIkTJ+Dn56f8nh05cqTQ2p4/f46BAwfC3NwcNjY2mDJlCjIyMgqct3XrVvj6+sLIyAjW1tYYPHgwHj58WOh9Xrp0Ce3bt4exsTH++9//Fvu9OXbsGNq1awcTExNYWlri/fffR0REhPL4iBEj0KFDBwDAgAEDIJFI0LFjxyKvl52djYULF8Ld3R2GhoawsbFB27ZtcfjwYZVrmpqa4u7duwgMDISJiQmcnJywaNEiCIKgcj25XI5vvvkGXl5eMDQ0hL29PcaNG6fy8wEAv/32G7p37w4nJycYGBigbt26+OyzzyCTyYq9fwA4dOgQjI2NMWTIEOTk5AAQf+779+8Pa2trGBoaonnz5ti3b5/K+xQ/LydOnMCECRNgZ2eHWrVqAQCSk5MxdepU5f/vdnZ26NKlCy5fvlxiPaQdGHZIq4SGhqJv377Q19fHkCFDcPv2bVy4cEHlnJSUFLRr1w5r1qxB165dsWrVKnz00Ue4desWHj16BACQyWTo0aMHFi5cCF9fX3z11VeYMmUKEhMTcePGjTeqLScnB4GBgbCzs8OXX36Jfv36AQB27dqFtLQ0jB8/HmvWrEFgYCDWrFmDoKAglfdfu3YNfn5+OHbsGMaMGYNVq1ahd+/e+P333wGIv1ydnZ0LDXihoaGoW7cuWrduXWKdwcHBiIiIwIIFCxAUFITQ0FD07t1b+Ytv2LBhePHiBQ4ePKjyvri4OBw7dgwffvhhiZ/x6tUrdOvWDT4+Pvjqq6/g4eGB2bNn46+//irxvcVds0ePHvDz88OKFStgYGCAwYMHY+fOnRg8eDDee+89fP7550hNTUX//v2RnJxc4BoDBw5ERkYGli1bhvfeew+rV6/G2LFjVc5ZsmQJgoKC4O7ujpUrV2Lq1Kk4evQo2rdvj4SEBJVzX7x4gXfffRdNmjTBN998g06dOhVZ/5EjRxAYGIinT59iwYIFmD59Ov755x/4+/srA9+4ceOUgWny5Mn46aeflGG3MAsWLMDChQvRqVMnfPvtt/j000/h4uJS4Je8TCZDt27dYG9vjxUrVsDX1xfz58/H/PnzVc4bN24cZs2aBX9/f6xatQojR45EaGgoAgMDVcLv5s2bYWpqiunTp2PVqlXw9fXFvHnz8MknnxRZKwDs378fvXr1woABA7B161bo6uri33//RatWrRAREYFPPvkEX331FUxMTNC7d+9Cg/2ECRNw8+ZNlc/76KOPsH79evTr1w/r1q3DzJkzYWRkpBIkScsJRFri4sWLAgDh8OHDgiAIglwuF2rVqiVMmTJF5bx58+YJAIQ9e/YUuIZcLhcEQRB++OEHAYCwcuXKIs/5+++/BQDC33//rXI8JiZGACBs2rRJ2TZ8+HABgPDJJ58UuF5aWlqBtmXLlgkSiUS4f/++sq19+/aCmZmZSlv+egRBEObMmSMYGBgICQkJyranT58Kurq6wvz58wt8Tn6bNm0SAAi+vr5CVlaWsn3FihUCAOG3334TBEEQZDKZUKtWLWHQoEEq71+5cqUgkUiEu3fvFvs5HTp0EAAIP/74o7ItMzNTcHBwEPr161egnpiYGJX3F/Z9V1xz27ZtyrZbt24JAASpVCqcO3dO2X7w4MEC/33mz58vABB69eql8lkTJkwQAAhXr14VBEEQ7t27J+jo6AhLlixROe/69euCrq6uSruippCQkGK/HwpNmjQR7OzshBcvXijbrl69KkilUiEoKKjA/e/atavEa/r4+Ajdu3cv9hzFz+akSZOUbXK5XOjevbugr68vPHv2TBAEQTh16pQAQAgNDVV5/4EDBwq0F/YzPW7cOMHY2FjIyMhQtnXo0EHw8vISBEEQfvnlF0FPT08YM2aMIJPJlOd07txZ8Pb2VnmfXC4X2rRpI7i7uyvbFD8vbdu2FXJyclQ+28LCQpg4cWKx3wfSbuzZIa0RGhoKe3t75d+eJRIJBg0ahB07dqh0n//yyy/w8fFBnz59ClxDIpEoz6lRowYmTZpU5DlvYvz48QXajIyMlP+empqK58+fo02bNhAEAVeuXAEAPHv2DCdPnsSoUaPg4uJSZD1BQUHIzMxUDt8AwM6dO5GTk1OqHhcAGDt2LPT09FRq1tXVxZ9//gkAkEqlGDp0KPbt26fSOxIaGoo2bdrAzc2txM8wNTVVqUdfXx8tW7bE3bt3S1VjUdccPHiw8usGDRrA0tISnp6eKivzFP9e2GdNnDhR5WvFf3/Fve/ZswdyuRwDBw7E8+fPlS8HBwe4u7vj77//Vnm/gYEBRo4cWWLtsbGxCA8Px4gRI2Btba1sb9y4Mbp06aL8/LKytLTEv//+i9u3b5d4bnBwsPLfJRIJgoODkZWVpRzy27VrFywsLNClSxeVe/f19YWpqanKvef/mU5OTsbz58/Rrl07pKWl4datWwU+e/v27Rg0aBDGjRuH7777DlKp+Kvp5cuXOHbsGAYOHKi8zvPnz/HixQsEBgbi9u3bePz4scq1xowZAx0dnQLfh7CwMDx58qQU3zXSRgw7pBVkMhl27NiBTp06ISYmBnfu3MGdO3fg5+eH+Ph4HD16VHludHR0ic/2iI6ORoMGDaCrW34LFnV1dZVzCPJ78OCB8pecqakpbG1tlfMyEhMTAeT9Yi6pbg8PD7Ro0UJlKCs0NBStWrUq9ao0d3d3la9NTU3h6OioMncmKCgI6enpymGEyMhIXLp0CcOGDSvVZ9SqVatAaLSysiow96MsCrumhYUFnJ2dC7QBKPSzXr/3unXrQiqVKu/99u3bEAQB7u7usLW1VXlFREQoJw8r1KxZE/r6+iXWfv/+fQAodL6Tp6cnnj9//kaTbRctWoSEhATUr18f3t7emDVrFq5du1bgPKlUijp16qi01a9fHwBU7j0xMRF2dnYF7j0lJUXl3v/991/06dMHFhYWMDc3h62trTLcKn6mFWJiYvDhhx+iX79+WLNmjcp/wzt37kAQBMydO7fAZyqG2F7/nhcWtlesWIEbN27A2dkZLVu2xIIFC94qWFPVw6XnpBWOHTuG2NhY7NixAzt27ChwPDQ0FF27di3Xzyyqh6eoSZgGBgbKv7HmP7dLly54+fIlZs+eDQ8PD5iYmODx48cYMWIE5HJ5mesKCgrClClT8OjRI2RmZuLcuXP49ttvy3yd4jRs2BC+vr7YunUrgoKCsHXrVujr62PgwIGlev/rf/NWEPJNiC3r97eoa5bms4ryeg1yuRwSiQR//fVXodc1NTVV+Tp/D4cmtG/fHtHR0fjtt99w6NAh/N///R++/vprhISElHnZulwuh52dXZGT/m1tbQGIk/87dOgAc3NzLFq0CHXr1oWhoSEuX76M2bNnF/iZdnR0hKOjI/78809cvHgRzZs3V/lMAJg5cyYCAwML/dzXQ3xh3/OBAweiXbt22Lt3Lw4dOoQvvvgCy5cvx549e/Duu++W/ptAVRbDDmmF0NBQ2NnZYe3atQWO7dmzB3v37kVISAiMjIxQt27dEicZ161bF2FhYcjOzlYZ0snPysoKAApMSlX8Lb00rl+/jqioKGzZskVlQnL+1TIAlH/rLs3k6MGDB2P69OnYvn070tPToaenh0GDBpW6ptu3b6tMpE1JSUFsbCzee+89lfOCgoIwffp0xMbGYtu2bejevbvye1IeyuP7W1a3b99W6Rm4c+cO5HI5ateuDUD8uRAEAW5ubsqej/KgeChgZGRkgWO3bt1CjRo1YGJi8kbXtra2xsiRIzFy5EikpKSgffv2WLBggUrYkcvluHv3rso9RUVFAYDKvR85cgT+/v7Fhrjjx4/jxYsX2LNnD9q3b69sj4mJKfR8Q0ND7N+/H++88w66deuGEydOwMvLC0Dez72enh4CAgLe6P4VHB0dMWHCBEyYMAFPnz5Fs2bNsGTJEoadaoLDWFTlpaenY8+ePejRowf69+9f4BUcHIzk5GTlUtV+/frh6tWrha7kUPxtv1+/fnj+/HmhPSKKc1xdXaGjo4OTJ0+qHF+3bl2pa1f0DuTvZRAEAatWrVI5z9bWFu3bt8cPP/yABw8eFFqPQo0aNfDuu+9i69atCA0NRbdu3VCjRo1S17RhwwaVlTXr169HTk5OgV8KQ4YMgUQiwZQpU3D37t1Szwkqrbp16wKAyvdXJpNhw4YN5fo5+b0eltesWQMAynvv27cvdHR0sHDhwgLfd0EQ8OLFizf6XEdHRzRp0gRbtmxRCXc3btzAoUOHCgTN0nq9HlNTU9SrVw+ZmZkFzs3/sy4IAr799lvo6emhc+fOAMTeEZlMhs8++6zAe3NycpR1F/YznZWVVez/FxYWFjh48KBySbji8Q52dnbo2LEjvvvuO8TGxhZ437Nnz4q8poJMJiswdGZnZwcnJ6dCvw+kndizQ1WeYqJsr169Cj3eqlUr5QMGBw0ahFmzZmH37t0YMGAARo0aBV9fX7x8+RL79u1DSEgIfHx8EBQUhB9//BHTp0/H+fPn0a5dO6SmpuLIkSOYMGEC3n//fVhYWGDAgAHKeQZ169bF/v37C8whKI6Hhwfq1q2LmTNn4vHjxzA3N8cvv/xS6HyS1atXo23btmjWrBnGjh0LNzc33Lt3D3/88UeBvZGCgoLQv39/ACj0l1NxsrKy0LlzZwwcOBCRkZFYt24d2rZtW+D7a2tri27dumHXrl2wtLRE9+7dy/Q5JfHy8kKrVq0wZ84cvHz5EtbW1tixY4fy2SvqEBMTg169eqFbt244e/Ystm7dig8++AA+Pj4AxAC2ePFizJkzB/fu3UPv3r1hZmaGmJgY7N27F2PHjsXMmTPf6LO/+OILvPvuu2jdujVGjx6N9PR0rFmzBhYWFliwYMEbXbNhw4bo2LEjfH19YW1tjYsXL2L37t0qk5EBsXflwIEDGD58OPz8/PDXX3/hjz/+wH//+1/l8FSHDh0wbtw4LFu2DOHh4ejatSv09PRw+/Zt7Nq1C6tWrUL//v3Rpk0bWFlZYfjw4Zg8eTIkEgl++umnEocNa9SogcOHD6Nt27YICAjA6dOnUbNmTaxduxZt27aFt7c3xowZgzp16iA+Ph5nz57Fo0ePcPXq1WKvm5ycjFq1aqF///7w8fGBqakpjhw5ggsXLuCrr756o+8rVUEVvwCMqHz17NlTMDQ0FFJTU4s8Z8SIEYKenp7w/PlzQRAE4cWLF0JwcLBQs2ZNQV9fX6hVq5YwfPhw5XFBEJfPfvrpp4Kbm5ugp6cnODg4CP379xeio6OV5zx79kzo16+fYGxsLFhZWQnjxo0Tbty4UejScxMTk0Jru3nzphAQECCYmpoKNWrUEMaMGSNcvXq1wDUEQRBu3Lgh9OnTR7C0tBQMDQ2FBg0aCHPnzi1wzczMTMHKykqwsLAQ0tPTS/NtVC7dPXHihDB27FjByspKMDU1FYYOHaqyHDq/n3/+WQAgjB07tlSfIQiqy43zGz58uODq6qrSFh0dLQQEBAgGBgaCvb298N///lc4fPhwoUvPC7umq6troUuvAagsRVYsPb9586bQv39/wczMTLCyshKCg4ML/f798ssvQtu2bQUTExPBxMRE8PDwECZOnChERkaWWFNxjhw5Ivj7+wtGRkaCubm50LNnT+HmzZsq55Rl6fnixYuFli1bCpaWloKRkZHg4eEhLFmyROXRAoqfzejoaKFr166CsbGxYG9vL8yfP19lCbjChg0bBF9fX8HIyEgwMzMTvL29hY8//lh48uSJ8pwzZ84IrVq1EoyMjAQnJyfh448/Vi75L+m/2507dwRHR0fB09NTuew9OjpaCAoKEhwcHAQ9PT2hZs2aQo8ePYTdu3cr36f4+b1w4YLK9TIzM4VZs2YJPj4+gpmZmWBiYiL4+PgI69atK/H7R9pDIgilmKVHRFVKTk4OnJyc0LNnT2zcuFFtn/Pbb7+hd+/eOHnyJNq1a6e2zyH1GTFiBHbv3o2UlBRNl0KkNpyzQ6SFfv31Vzx79qzAU5jL2/fff486deqgbdu2av0cIqK3wTk7RFokLCwM165dw2effYamTZsqn9dT3nbs2IFr167hjz/+wKpVq97qQYtEROrGsEOkRdavX4+tW7eiSZMmKhuRlrchQ4bA1NQUo0ePxoQJE9T2OURE5UGjw1gnT55Ez5494eTkBIlEgl9//VXluCAImDdvHhwdHWFkZISAgIACjz1/+fIlhg4dCnNzc1haWmL06NEce6Zqa/PmzcjJycHFixdLfNry2xAEAcnJyfi///u/cn3KNFW8zZs3889M0noaDTupqanw8fEp9EFwgPiI79WrVyMkJARhYWEwMTFBYGAgMjIylOcMHToU//77Lw4fPoz9+/fj5MmTBXYpJiIiouqr0qzGkkgk2Lt3L3r37g1A/Jujk5MTZsyYoXxuRWJiIuzt7bF582YMHjwYERERaNiwIS5cuKB8xPiBAwfw3nvv4dGjR3ByctLU7RAREVElUWn7n2NiYhAXF6fyiHALCwv4+fnh7NmzGDx4MM6ePQtLS0uVvVQCAgIglUoRFhZW6K7WAJCZmany5Ey5XI6XL1/CxsaGEy2JiIiqCMWQupOTU4G9B/OrtGEnLi4OAGBvb6/Sbm9vrzwWFxcHOzs7leO6urqwtrZWnlOYZcuWYeHCheVcMREREWnCw4cPUatWrSKPV9qwo05z5szB9OnTlV8nJibCxcUFDx8+hLm5uQYrIyIiotJKSkqCs7MzzMzMij2v0oYdBwcHAEB8fDwcHR2V7fHx8WjSpInynNf3IcrJycHLly+V7y+MgYEBDAwMCrSbm5sz7BAREVUxJU1BqbRPUHZzc4ODgwOOHj2qbEtKSkJYWBhat24NAGjdujUSEhJw6dIl5TnHjh2DXC6Hn59fhddMRERElY9Ge3ZSUlJw584d5dcxMTEIDw+HtbU1XFxcMHXqVCxevBju7u5wc3PD3Llz4eTkpFyx5enpiW7dumHMmDEICQlBdnY2goODMXjwYK7EIiIiIgAaDjsXL15Ep06dlF8r5tEMHz4cmzdvxscff4zU1FSMHTsWCQkJaNu2LQ4cOABDQ0Ple0JDQxEcHIzOnTtDKpWiX79+WL16dYXfCxEREVVOleY5O5qUlJQECwsLJCYmcs4OEVVqcrkcWVlZmi6DqELo6elBR0enyOOl/f1daScoExGRqqysLMTExEAul2u6FKIKY2lpCQcHh7d6Dh7DDhFRFSAIAmJjY6GjowNnZ+diH6BGpA0EQUBaWppy1XX+ldllxbBDRFQF5OTkIC0tDU5OTjA2NtZ0OUQVwsjICADw9OlT2NnZFTukVRz+1YCIqAqQyWQAAH19fQ1XQlSxFOE+Ozv7ja/BsENEVIVw/z6qbsrjZ55hh4iIiLQaww4REWmljh07YurUqeV+rrpt3rwZlpaWarm2RCLBr7/+WuTxe/fuQSKRIDw8HABw/PhxSCQSJCQkqKWeisKwQ0REatGzZ09069at0GOnTp2CRCLBtWvXlG2//PIL3nnnHVhZWcHIyAgNGjTAqFGjcOXKFZX3ZmVl4YsvvkCzZs1gYmICCwsL+Pj44H//+x+ePHmiPG/Pnj347LPP1HNzABYsWACJRFLg5eHhobbPrGht2rRBbGwsLCwsNF3KW2HYISIitRg9ejQOHz6MR48eFTi2adMmNG/eHI0bNwYAzJ49G4MGDUKTJk2wb98+REZGYtu2bahTpw7mzJmjfF9mZia6dOmCpUuXYsSIETh58iSuX7+O1atX4/nz51izZo3yXGtr6xJ3w35bXl5eiI2NVXmdPn1arZ9ZkfT19d/6GTeVAcMOERGpRY8ePWBra4vNmzertKekpGDXrl0YPXo0AODcuXNYsWIFVq5ciZUrV6Jdu3ZwcXGBr68v/ve//+Gvv/5Svvfrr7/G6dOncezYMUyePBm+vr5wcXFBhw4dEBISgqVLlyrPfX1oat26dXB3d4ehoSHs7e3Rv39/lbrkcjk+/vhjWFtbw8HBAQsWLCjxHnV1deHg4KDyqlGjhvJ47dq1sXjxYgQFBcHU1BSurq7Yt28fnj17hvfffx+mpqZo3LgxLl68WODav/76q7LewMBAPHz4UOX4b7/9hmbNmsHQ0BB16tTBwoULkZOTozx++/ZttG/fHoaGhmjYsCEOHz5c4DPOnz+Ppk2bwtDQEM2bNy/Qi/b6MJZiiO3gwYPw9PSEqakpunXrhtjYWOV7cnJyMHnyZFhaWsLGxgazZ8/G8OHDlftaAsDu3bvh7e0NIyMj2NjYICAgAKmpqSV+v98Uww4RURUkCALSsnI08irtLkO6uroICgrC5s2bVd6za9cuyGQyDBkyBACwfft2mJqaYsKECYVeJ3+vwvbt29GlSxc0bdq0xHPzu3jxIiZPnoxFixYhMjISBw4cQPv27VXO2bJlC0xMTBAWFoYVK1Zg0aJFhQaEsvr666/h7++PK1euoHv37hg2bBiCgoLw4Ycf4vLly6hbty6CgoJUvkdpaWlYsmQJfvzxR5w5cwYJCQkYPHiw8vipU6cQFBSEKVOm4ObNm/juu++wefNmLFmyBIAY3Pr27Qt9fX2EhYUhJCQEs2fPVqkrJSUFPXr0QMOGDXHp0iUsWLAAM2fOLPF+0tLS8OWXX+Knn37CyZMn8eDBA5X3LV++HKGhodi0aRPOnDmDpKQklXlCsbGxGDJkCEaNGoWIiAgcP34cffv2LfXP1ZvgQwWJiKqg9GwZGs47qJHPvrkoEMb6pfv1MWrUKHzxxRc4ceIEOnbsCEAcwurXr59yHkhUVBTq1KkDXd28a65cuRLz5s1Tfv348WNYWFggKipKeR2FPn36KENJ48aN8c8//xSo48GDBzAxMUGPHj1gZmYGV1fXAoGpcePGmD9/PgDA3d0d3377LY4ePYouXboUeX/Xr1+HqampStuHH36IkJAQ5dfvvfcexo0bBwCYN28e1q9fjxYtWmDAgAEAxCG81q1bIz4+Hg4ODgDEZ8p8++238PPzAyAGMU9PT5w/fx4tW7bEwoUL8cknn2D48OEAgDp16uCzzz7Dxx9/jPnz5+PIkSO4desWDh48CCcnJwDA0qVL8e677yrr2rZtG+RyOTZu3AhDQ0N4eXnh0aNHGD9+fJH3q6gtJCQEdevWBQAEBwdj0aJFyuNr1qzBnDlz0KdPHwDAt99+iz///FN5PDY2Fjk5Oejbty9cXV0BAN7e3sV+5ttizw4REamNh4cH2rRpgx9++AEAcOfOHZw6dUo5hFWUUaNGITw8HN999x1SU1OL/Vv/unXrEB4ejlGjRiEtLa3Qc7p06QJXV1fUqVMHw4YNQ2hoaIFzFfOHFBwdHZVbFRSlQYMGCA8PV3nl/8X/+nXt7e0BqP5yV7Tl/yxdXV20aNFC+bWHhwcsLS0REREBALh69SoWLVoEU1NT5WvMmDGIjY1FWloaIiIi4OzsrAw6ANC6dWuVuiIiItC4cWMYGhoWeU5hjI2NlUEHUP0+JSYmIj4+Hi1btlQe19HRga+vr/JrHx8fdO7cGd7e3hgwYAC+//57vHr1qsTPfRvs2SEiqoKM9HRwc1Ggxj67LEaPHo1JkyZh7dq12LRpE+rWrYsOHTooj7u7u+P06dPIzs6Gnp4eAHHzR0tLywKTm93d3REZGanSptgzydrausgazMzMcPnyZRw/fhyHDh3CvHnzsGDBAly4cEG5zFvx2QoSiaTETVf19fVRr169Ys/Jf13FMFthbWXZ4DUlJQULFy5E3759CxzLH17UobDvU1mGoHR0dHD48GH8888/OHToENasWYNPP/0UYWFhcHNzK+9yAbBnh4ioSpJIJDDW19XIq6wrcwYOHAipVIpt27bhxx9/xKhRo1SuMWTIEKSkpGDdunUlXmvIkCE4fPhwgYm0paGrq4uAgACsWLEC165dw71793Ds2LEyX6ci5OTkqExajoyMREJCAjw9PQEAzZo1Q2RkJOrVq1fgJZVK4enpiYcPH6pMHD537pzKZ3h6euLatWvIyMgo8pyysrCwgL29PS5cuKBsk8lkuHz5ssp5EokE/v7+WLhwIa5cuQJ9fX3s3bv3rT67OOzZISIitTI1NcWgQYMwZ84cJCUlYcSIESrHW7dujRkzZmDGjBm4f/8++vbtC2dnZ8TGxmLjxo2QSCTKXd6nTZuGP/74A507d8b8+fPRrl07WFlZISoqCn/99VeRG0Xu378fd+/eRfv27WFlZYU///wTcrkcDRo0eKt7y8nJQVxcnEqbRCJRDk29KT09PUyaNAmrV6+Grq4ugoOD0apVK+Xw0Lx589CjRw+4uLigf//+kEqluHr1Km7cuIHFixcjICAA9evXx/Dhw/HFF18gKSkJn376qcpnfPDBB/j0008xZswYzJkzB/fu3cOXX375VnUDwKRJk7Bs2TLUq1cPHh4eWLNmDV69eqUMuGFhYTh69Ci6du0KOzs7hIWF4dmzZ8ogpw7s2VGjfx++xJQfzuDfhy81XQoRkUaNHj0ar169QmBgoMo8EoUvv/wS27Ztw5UrV9CjRw+4u7tjwIABkMvlOHv2LMzNzQGIQzRHjx7F7NmzsWnTJrRt2xaenp6YOnUq/P39i3w6sKWlJfbs2YN33nkHnp6eCAkJwfbt2+Hl5fVW9/Xvv//C0dFR5aWYdPs2jI2NMXv2bHzwwQfw9/eHqakpdu7cqTweGBiI/fv349ChQ2jRogVatWqFr7/+WvnZUqkUe/fuRXp6Olq2bIn//Oc/ypVaCqampvj9999x/fp1NG3aFJ9++imWL1/+1rXPnj0bQ4YMQVBQEFq3bg1TU1MEBgYqh9fMzc1x8uRJvPfee6hfvz7+97//4auvvlKZPF3eJII613pVEUlJSbCwsEBiYqLyf6jysHTPZZz4NxYdvZwwp2/hyySJiEojIyMDMTExcHNzU/ucDKLyJJfL4enpiYEDB77RE62L+9kv7e9v9uyoSWJaFk5HiF2bpyJikZiWpeGKiIiI1O/+/fv4/vvvERUVhevXr2P8+PGIiYnBBx98oLGaGHbU5PDVR5DndprJBQFHrhV8XDoREZG2kUql2Lx5M1q0aAF/f39cv34dR44cUeucnJJwgnI5eJ6UgVepmSptv1+8D8UAoSAA+y7cR2NXG5VzrEwMUMOc3dFERKQ9nJ2dcebMGU2XoYJhpxws23MZNx4W/0CkuIQ0BP+f6uZwXs5WWDmijTpLIy3178OX2HA4AmO7eMLLuehnixAREcNOuejW1AWRTxKRI5OjNLO9BQEQABy58wr1/vsnDPV0YKinAyN9KYxy/91QTyf338U2I33VdsUxw9xjhb4vX7uBrrTK71pLeX67cA+3Hidg34X7DDtERCVg2CkHXXxqob6TBRb+fAmxr1IhLybxCAKQLQCPM4AsAQAEpGTmICUzp+g3lQOJBDDUzQ1NulIY6ucPTTrK0GSoKy0kPEkLhi19ndzrSV8LWTrQkTJUqdPrk98npHnBwlhfw1UREVVeDDvlxNXWDOvGtMXK36/hxM3YIs/r1MgRU3s0hlQqQUaWHOnZMqRny5Ch+GeWDBk5MqTnO5aZLUN6luI8ed75ynPFrxXH8l8nWyYmL0GA8pi66etKiwhNih4qqUo4Uu25khYawvL3chnmBi09HUm17K0qbPJ7v1Z1NFwVEVHlxbBTjgz1deHtaoOTN2MLHc6SAPB2tYFR7m7BBro6sIBeIWeWn2yZXBmkMhVhKCtfuMrOF6Ky8oUrxStLjowcMTzlnS8vELYyc/L2dMnKkSMrR46kDPX2VulIJarDefl6nMSeq9dCVb7eKCM9HRjkGxIsLoRpcgiQk9+JiN4ew045uxObCKlUAlkhY1lSqQS3YxMrtB49HSn0dKQwM1RvqJLLBWTmyPOFJEVv02s9UlmqIUvZC5Wt2suVoXIducr5il/0MrliCFCttwYAJfY4FQhVhcy5MtDNe8/rc64U578+BPimk9+9Xazx5fCSdy8mIqoOGHbKWcTjV5DJBehIJZBKJOjZ3BW/X7wPmVyATC4g4pF6t7HXFKlUIv4i1y/bbshlJQgCsmRy5RCgSmhSDu/JX/tadWiwQJjKkStDWP7jiiFAALmBS45XyFbr/enrSFVDEEq/CzIg9h7q6kgR2MRZPQUSEVVBDDvlKCtHhofPUwEAjlbGmDfAF662ZujW1BkLf76Exy9T8fB5KrJyZNDXVW8o0FYSiQQGujoVMgSYI5MjI0decNgvK38oyjc0mG+47/XeLMVwn2IIMH/YUhkClMmRJVMdAtSXAE6G4j9LGk0zMtRDn9b1UMfJUhm6iSqDuLg4LFmyBH/88QceP34MOzs7NGnSBFOnTkXnzp1Ru3Zt3L9/H2fPnkWrVq2U75s6dSrCw8Nx/PhxAMCCBQuwcOFCjBs3DiEhIcrzwsPD0bRpU8TExKB27doVfHdU2THslKPMbDlq25minoMFJr7bCIZ6YqBRTF7+9sC/uBuXhKwcOcNOFaCrI4WpjhSmBur93+T1IUBlb1S+wJSUnoXDl+8jJq7oYdCkHCDqRTau7I8A9kfASE8H9R3M4OlgBk9Hc3g4mMHD0RwWRuoNiUSvu3fvHvz9/WFpaYkvvvgC3t7eyM7OxsGDBzFx4kTcunULgLjJ5+zZs3HixIlir2doaIiNGzdixowZcHd3r4hboCqOYaccmRnpYe2YdpAW8tdvQ31dzOzlA7kgFHqcqq/SDgHqC3Ks/SuxyGc5dfZyRI6OLiLikhEZl4T0bBmuPkzA1YcJKufVtDSCp6MiAJnD09EMrjYm7AUitZkwYQIkEgnOnz8PExMTZbuXlxdGjRql/Hrs2LEICQnBn3/+iffee6/I6zVo0AB2dnb49NNP8fPPP6u1dtIODDvlrKQgw6BDb6q4ye86UglsjHUxrUdjAOLk7XsvUnErNhkRsUm4FZeEiNhkPE5IV76ORDxVvl/RC9TQ0Sw3AJmjgYMZe4EqM0EAstM089l6xiWPqeZ6+fIlDhw4gCVLlqgEHQVLS0vlv7u5ueGjjz7CnDlz0K1bN0ilRW/f+Pnnn6NFixa4ePEimjdvXuZboOqFYYeoiijL5HcdqQR1bU1R19YU3Rs7KtsT07NxKzYpNwCJQSgyPrmEXiDzfD1B7AWqNLLTgKVOmvns/z4B9AsGl8LcuXMHgiDAw8OjVOf/73//w6ZNmxAaGophw4YVeV6zZs0wcOBAzJ49G0ePHi3Vtan6YtghqgLKa/K7hZEe/OrYwK9O3nN5FL1AEbFJ+XqCXu8Fileeb6SngwYOZipDYR6OZjBX8+MNqGoShNJsopPH1tYWM2fOxLx58zBo0KBiz128eDE8PT1x6NAh2NnZvU2ZpOUYdoiqAHVOfs/fC5Q7CgYASEzLRkRcUm5PUDJuxYkhKD1bhvCHCQh/rReolpURPBzMxaEwR3EozNXaGFL2AqmHnrHYw6Kpzy4ld3d3SCQS5STk0pg+fTrWrVuHdevWFXte3bp1MWbMGHzyySfYuHFjqa9P1Q/DDlEVoInJ7xbGemhVxwatXusFinmemjsHKK8n6EliBh69SsejV0X1AuUNhTVwYC9QuZBISj2UpEnW1tYIDAzE2rVrMXny5ALzdhISElTm7QCAqakp5s6diwULFqBXr17FXn/evHmoW7cuduzYUd6lkxZh2CGqIirD5HcdqQT17ExRz84UPRrnzRdJSMtSzgG6FZuMiLgkRJbQC+TpaJ63LJ69QFpt7dq18Pf3R8uWLbFo0SI0btwYOTk5OHz4MNavX4+IiIgC7xk7diy+/vprbNu2DX5+fkVe297eHtOnT8cXX3yhzlugKo5hh4jemqWxfoFeoByZHPdepKmsBrv1Wi/Q4Zt5vUDG+mIvUP6hMA8HM7VvdULqV6dOHVy+fBlLlizBjBkzEBsbC1tbW/j6+mL9+vWFvkdPTw+fffYZPvjggxKvP3PmTKxfvx4ZGRnlXTppCYlQ1tljWigpKQkWFhZITEyEubm5pssh0moJaVnKOUARufOBouKTVZ4knZ+ztZFyObyiJ8ilGvYCZWRkICYmBm5ubjA05CavVH0U97Nf2t/f7NkhogplaayP1nVt0Lru671AqYjItxosIjYJsYkZePgyHQ9fFt4LlD8ANWAvEBEVgWGHiDROV0eKenZmqGdnhp4+eXOBXqXmmwuUOxQWGZ+MtCwZrjxIwJUHCSrXcbY2gqeDOAdI8YDE6tgLRESqGHaIqNKyMim6F+hm7hwgxVBYXFJeL9ChfL1AJoq5QI55Q2EejuZq3/OMiCoP/t9ORFVK/l6gXq/1AonPBRJ7giLikhAVn4LULBkuP0jA5dd6gVysjeHx2rJ4Zyv2AhFpI4YdItIKVib6aFO3BtrUraFsy5HJEfM8FRG5Q2GKpfFxSRl48DIND16mFdoLpFgO39DRDA0cKlcvENeUUHVTHj/zlef/YCKicqarI4W7vRnc7VV7gV6mZinnACnmA5XUC+SZb5PUho7mqGVlVKG9QDo64pOxs7KyYGRkVGGfS6RpaWnihrd6em++AIFLz8Gl50SU1wt0M9/2GBGxSYhPyiz0fBN9HeWzgBRDYersBRIEAQ8ePEB2djacnJyK3RGcSBsIgoC0tDQ8ffoUlpaWcHR0LHBOaX9/M+yAYYeIivYyNUucCJ1vKOx2fAqyZIU/F8jVJm8ukPiAxPLrBcrKykJMTAzk8sI/m0gbWVpawsHBAZJCnhLPsFMGDDtEVBbZirlAsapDYUX1Apka6Cp3ilcMhXk4mMHkDXqB5HI5srKy3vYWiKoEPT095RBuYRh2yoBhh4jKg6IXKP9QWEm9QOJzgcxU5gIV9jdYIiqIYacMGHaISF3y9wLdzLdT/NPkonuBPBzMlAHI09EcDezfrBeISNsx7JQBww4RVbQXKZnKp0MrhsLuPC28F0giAVytjfOGwBzN2AtEBIadMmHYIaLKIFsmx91nqbgVl28orBS9QIoApJgLZKzPXiCqHhh2yoBhh4gqs/y9QIqhsJJ6gRSrwRRPh2YvEGkjhp0yYNghoqpG0QukWA6vWBr/rIheIDMDXXjkWw0mPheIvUBUtTHslAHDDhFpi+cpmSr7g0XEJuPO02Rkywr+US+RALVtTPI9GFEcBmMvEFUVDDtlwLBDRNosWyZH9LMU5d5gN2OTcCsuucReoPwBiL1AVBkx7JQBww4RVUfPUzKVAUgxFFZSL5Dna0NhNS3ZC0Saw7BTBgw7RESirByxFyj/RqkRscl4nlJEL5ChrsqDET0dzVHf3pS9QFQhGHbKgGGHiKh4z5IzlZujKobCop+lFNkL5GZjIgagfM8GYi8QlTeGnTJg2CEiKjtFL1BE7hwgxcqw5ymF792l6AVSLIf3yH06tJF+0XsfERWHYacMGHaIiMrPs+RM5eao+Z8OnSMvuhdIMRHa09Ecnk7mcLIwZC8QlYhhpwwYdoiI1CsrR447T1PyhsJye4KK6gUyN9SFh6M5PB3yzwViLxCpYtgpA4YdIiLNeJqcoVwNpghARfUCSSVA7RomBYbC2AtUfWlF2JHJZFiwYAG2bt2KuLg4ODk5YcSIEfjf//6n/MEWBAHz58/H999/j4SEBPj7+2P9+vVwd3cv9ecw7BARVR6ZOTJEP00tMBT2IrX4XqCGucvhPRzM0cDBDIZ67AXSdqX9/V2p1wYuX74c69evx5YtW+Dl5YWLFy9i5MiRsLCwwOTJkwEAK1aswOrVq7Flyxa4ublh7ty5CAwMxM2bN2FoaKjhOyAiorIy0NVBQydzNHTK++UlCAKepWQqN0dV9ATdeZqCpIwcnI95ifMxL5XnK3uBckOQYj6QI3uBqqVK3bPTo0cP2NvbY+PGjcq2fv36wcjICFu3boUgCHBycsKMGTMwc+ZMAEBiYiLs7e2xefNmDB48uFSfw54dIqKqKTNHJs4Fem2LjJdF9AJZGOnl2x7DTDkXiL1AVZNW9Oy0adMGGzZsQFRUFOrXr4+rV6/i9OnTWLlyJQAgJiYGcXFxCAgIUL7HwsICfn5+OHv2bJFhJzMzE5mZeQ/ISkpKKv/i5TLg2SkgPRYwcgRs2wFS/s9ERFSeDHR14OVkAS8nC2WbIAjiirDcOUC3ch+MGP0sBYnp2QiLeYmw13qB3GqYFBgKYy+Q9qjUYeeTTz5BUlISPDw8oKOjA5lMhiVLlmDo0KEAgLi4OACAvb29yvvs7e2VxwqzbNkyLFy4UH2FP9wDXJoCpD3KazOuBfiuApz7qu9ziYgIEokEduaGsDM3RIf6tsp2RS+QcigsXy9Q9LNURD9LxR/XYpXn5+8Fapj7YET2AlVNlTrs/PzzzwgNDcW2bdvg5eWF8PBwTJ06FU5OThg+fPgbX3fOnDmYPn268uukpCQ4OzuXR8li0DnVH8Bro4Npj8X2drsZeIiINKC4XqCbrz0YMfpZarG9QIrl8IqhMAdz9gJVZpU67MyaNQuffPKJcjjK29sb9+/fx7JlyzB8+HA4ODgAAOLj4+Ho6Kh8X3x8PJo0aVLkdQ0MDGBgYFD+BctlYo/O60EHyG2TAJemAjXf55AWEVElkL8XqGMDO2V7Zo4Mt+NTVAJQRGwSXqVlK3uB9ufrBbI0zjcXKHeLDHd7U/YCVRKVOuykpaVBKpWqtOno6EAulwMA3Nzc4ODggKNHjyrDTVJSEsLCwjB+/PiKLleco5N/6KoAAUh7KJ5n37GiqiIiojIy0NVBo5oWaFRTtRfoae7ToSOUzwYSe4ES0rJx7u5LnLur2gtUx9ZU+XRoxVAYe4EqXqUOOz179sSSJUvg4uICLy8vXLlyBStXrsSoUaMAiIl86tSpWLx4Mdzd3ZVLz52cnNC7d++KLzg9tuRzACC1uEBERESVkUQigb25Iexf6wXKyFbMBRJDkOIp0a/SsnHnaQruPE3B71fzrqPSC5TbE8ReIPWq1EvPk5OTMXfuXOzduxdPnz6Fk5MThgwZgnnz5kFfXx9A3kMFN2zYgISEBLRt2xbr1q1D/fr1S/055bb0PP44cLRTyecl1gVaLQAaDwJ09N7884iIqFJS9ALdjM3bKT4iNgl3n6dCVsjToXWkEtTJXRHmmW+3eHtzA/YCFUMrnqBcUcot7MhlwL7a4mTkQuftAMjRAe4aA5AAFi5A2ylAkw8BPT4AkYhI273eC6R4NlBCWnah51sZ68EjN/h4OIpDYfXs2AukwLBTBuX6UEHlaixANfDkJvNWPwGxL4B/1gCpT8U2M0egzWTAdwSgb/x2n09ERFWKIAiIT8rMXQqfpFwaX1IvkCIAKYbCqmMvEMNOGZT7E5QLfc6OM+D7Td6y8+x04PJPwJlvgKTHuefUAFpPBFr8BzDkk5yJiKozRS+QylBYCb1A4mTovCXx2t4LxLBTBmrZLqK0T1DOyQSubgdOrQQS7otthpZAq/GA3zjAyKp86iEioipP2QsUm6TybKC7z1JQSCeQSi9Q/qEwO7OK6wX69+FLbDgcgbFdPOHlbF2u12bYKYNKsTeWLAe4sRs4+SXw4rbYpm8GtPwP0DoYMKmhmbqIiKjSy8gWnwuUtz+YOByWmF54L5C1ib5yRZjin+rqBVq65zJO/BuLjl5OmNO3ablem2GnDCpF2FGQy4Cbv4mh5+m/YpuuEdB8FNBmEmDuWPz7iYiIIPYCxSVlvPZcoORie4Hq2poUGAp7m16gxLQsDPn6CGRyATpSCbZPC4CFsf5b3lkehp0yqFRhR0EuB6IOACdXAE+uiG06BkCzYYD/FMDSRbP1ERFRlZS/F+hm7oMRS+oFUmyOqugJcrc3hYFuyb1Au8/exf8djYAgABIJMCbAE/1a1Sm3e2HYKYNKGXYUBAGIPgqc+AJ4eE5sk+oCPoOBttMBm7qarY+IiKq8wnqBImKTEPM8tdBeIF2pBHVtTZWrwTwczOBgpg8poNILtHj3ZcQlpCm/drA0xv/6N1O5lpWJAWqYv9njVxh2yqBShx0FQQDunQZOfgHEnBDbJFKgUX+g3QzAzkOz9RERkdbJyJYhKj65QAhKysgpcK6zIWD8BlN+vF2s8eXw1m9UH8NOGVSJsJPfw/PinJ7bB/PaPHsB7WcCjj6aq4uIiLSeIAiITcxQzgG6GZuEW7FJeP4qFXb64lPlSjPFRwJAV0eKKd290cWn1hvVwrBTBlUu7Cg8CQdOfQlE/J7XVr8b0H4WUKu5xsoiIqLqJz1LhlOR8fj+0L9ISssq9lypBHCyNsG8Ab5wtTV7489k2CmDKht2FOJvAqdXAjd+AQRxR3jU6SiGntptNVoaERFVLxlZOVj5+zWcuFn05tgdvRwxrafPWy91L+3vb+lbfQpVDvYNgX7/B0y8IO6zJdUF7h4HNncHfngXuHNUnPNDRESkZob6uvB2tUFRI1kSAN6uNhX6ZGeGHW1Sox7Qey0w6TLQfDSgow88+AfY2hf4/h3g1p8MPUREpHZ3YhMhlRYed6RSCW7HJlZoPQw72sjKFeixEphyFWg1QXwo4ZPLwI4hQEhb4N+94sMLiYiI1CDi8SvlgwT1dKTo6+cGPR0ppBIJZHIBEY9eVWg9DDvazNwJ6LYMmHod8J8K6JsC8TeAXSOAda2AqzvFbSqIiIjKSVaODA+fpwIAHK2MsXZMW4zr2hBrx7SFo5UxAODh81Rk5VTcX7oZdqoDU1ugy0Ix9HT4BDC0AJ5HAXvHAt/6Ape2ADnFz5wnIiIqjcxsOWrbmaKrTy2sHdNOudrK1dYM68a0RRefWnCzM0NWjrzCauJqLGjBaqyyykgCLnwPnF0LpL0Q28xridtQNBsG6Blptj4iIqrS5IIAaTEP2ynpeGlx6XkZVLuwo5CVClzaDJxZDaTEiW2m9uKGo74jAQNTjZZHRERUHIadMqi2YUchOwO48hNwZhWQ+FBsM7IGWk8AWo4Vh72IiIgqGYadMqj2YUchJwu4thM49RXwKkZsM7AA/MYBrcYDxtaarY+IiCgfhp0yYNh5jSwH+HePuP/W80ixTd8UaDEaaB0MmNpptj4iIiIw7JQJw04R5HLg1u/iTutx18U2XUPAdwTQZjJgUVOj5RERUfXGsFMGDDslEAQg6qAYeh5fFNt09IEmQ4G2UwGr2pqsjoiIqimGnTJg2CklQRD33Dr5BXD/jNgm0QEaDwLaTQdquGu0PCIiql4YdsqAYecN3DsDnPoSiD6W2yABGvUF2s0A7L00WhoREVUP3PWc1Ku2PzBsL/CfY0D9dwEIwI1fgPVtgB1DgSdXNF0hERERAIYdelu1fIEPdgDjTgENewOQALf2Axs6Alv7Aw/CNFwgERFVdxzGAoexytWzSODUSuD6LkDI3eStdjug/SzArT1QDo8HJyIiAjhnp0wYdtTg5V3g9NdA+HZAni22OfuJoadeAEMPERG9NYadMmDYUaOEh+I2FJd/BGSZYptjEzH0NHgPkHIklYiI3gzDThkw7FSA5DjgnzXAxR+A7DSxza6huHrLqw8g1dFsfUREVOUw7JQBw04FSn0BnFsLhG0AspLFNpt6YujxHgDo6Gm2PiIiqjIYdsqAYUcD0l8B578Hzq4FMhLENksXoO008cnMugYaLY+IiCo/hp0yYNjRoMxk4MJG4Oy3QOozsc3MCfCfAjQLAvSNNVsfERFVWgw7ZcCwUwlkpYmTmM+sApKfiG0mtuIu6y1GAwZmmq2PiIgqHYadMmDYqURyMoHwUHHZesIDsc3ICmg1AWg5FjCy1Gh5RERUeTDslAHDTiUkyxYfTHjqK+DFHbHNwBxoOQZoNREwsdFsfUREpHEMO2XAsFOJyWXAv3uBk18CzyLENj1joPkooM0kwMxBs/UREZHGMOyUAcNOFSCXA5F/ACe/AGKvim06BuIkZv8pgKWzZusjIqIKx7BTBgw7VYggAHeOACdWAI/Oi21SPaDJEHHZunUdzdZHREQVhmGnDBh2qiBBAGJOij09906JbRId8cGE7aYDtg00Wx8REakdw04ZMOxUcQ/OiXN67hzObZAADd8H2s8EHLw1WhoREalPaX9/cxdGqvpcWgEf7gbG/A149AAgADd/BULaAtsGA48uabpCIiLSIIYd0h41mwGDQ4Hx/wCN+gGQAFF/Af/3DvBTH+D+P5qukIiINIDDWOAwltZ6fhs4tRK4thMQZGKbqz/QfhZQpyMgkWi0PCIiejucs1MGDDta7tU94PQ3wJWtgDxbbKvZXAw99QMZeoiIqiiGnTJg2KkmEh8D/6wGLm0GcjLENgdvMfR49ASkHNUlIqpKGHbKgGGnmkl5CvyzRtxtPTtVbLP1ANrNALz6Ajq6mq2PiIhKhWGnDBh2qqm0l8C59UDYd0BmothmXQdoOx1oPAjQ1ddsfUREVCyGnTJg2KnmMhKB8xuAs+uA9Jdim4WzuA1F02GAnqFm6yMiokIx7JQBww4BADJTgEubgDOrgdSnYpupA+A/GfAdAeibaLQ8IiJSxbBTBgw7pCI7Hbj8E3DmGyDpsdhmbAO0ngi0GAMY8meEiKgyYNgpA4YdKlROFnB1O3B6pbh8HQAMLQC/8YDfOMDYWqPlERFVdww7ZcCwQ8WS5QA3dgOnvgKeR4lt+qZAi/8ArYMBU1vN1kdEVE0x7JQBww6VilwGROwTNx2NvyG26RoBzUcCbSYB5k6arY+IqJph2CkDhh0qE7kciDoAnPwCeHJZbNPRF1du+U8BrFw1Wx8RUTXBsFMGDDv0RgQBiD4mhp4HZ8U2qS7QeDDQbjpgU1ez9RERaTmGnTJg2KG3du+0GHruHhe/lkjFndfbzQDsPDVaGhGRtirt729uBkRUHmq3BYJ+A0YfAep3AwQ5cH0XsK4VsPND4Em4piskIqq2GHaIypNzC+CDncC4k4BnL7Et4ndgQwcgdCDw8IJm6yMiqoY4jAUOY5EaPY0Ql6zf+EXs7QEAtw7iTuu12wISiWbrIyKqwjhnpwwYdkjtXkSLDye8ugOQ54htLq2B9jOBup0ZeoiI3gDDThkw7FCFSXgAnP4GuPITIMsS25yaij099d8FpBxZJiIqLa2ZoPz48WN8+OGHsLGxgZGREby9vXHx4kXlcUEQMG/ePDg6OsLIyAgBAQG4ffu2BismKoalC9BjJTDlGtBqovhQwidXgB0fAN+1A27sER9eSERE5aZSh51Xr17B398fenp6+Ouvv3Dz5k189dVXsLKyUp6zYsUKrF69GiEhIQgLC4OJiQkCAwORkZGhwcqJSmDuCHRbCky9DrSdJm4/EX8D2D1SXMF1dYe4TQUREb21Sj2M9cknn+DMmTM4depUoccFQYCTkxNmzJiBmTNnAgASExNhb2+PzZs3Y/DgwaX6HA5jkcalvQTObwDOrQMyEsU2q9piEPIZAuga5J0rlwHPTgHpsYCRI2DbDpDqaKRsIiJN0ophrH379qF58+YYMGAA7Ozs0LRpU3z//ffK4zExMYiLi0NAQICyzcLCAn5+fjh79myR183MzERSUpLKi0ijjK2Bjp8AU28AnecDxjbiTuu/TwFWNwXCvgOy04GHe4B9tYGjnYB/PhD/ua+22E5ERIWq1GHn7t27WL9+Pdzd3XHw4EGMHz8ekydPxpYtWwAAcXFxAAB7e3uV99nb2yuPFWbZsmWwsLBQvpydndV3E0RlYWgubjUx9ToQuAwwdQCSHgN/fQyE1AFO9QPSHqm+J+0xcKo/Aw8RUREqddiRy+Vo1qwZli5diqZNm2Ls2LEYM2YMQkJC3uq6c+bMQWJiovL18OHDcqqYqJzomwCtJwBTrgLdvwIsagHm8UChg865jZemcnIzEVEhKnXYcXR0RMOGDVXaPD098eDBAwCAg4MDACA+Pl7lnPj4eOWxwhgYGMDc3FzlRVQp6RkCLf4DDF4F6AlAkY/jEYC0h+JcHiIiUlGpw46/vz8iIyNV2qKiouDq6goAcHNzg4ODA44ePao8npSUhLCwMLRu3bpCayVSq8xnpTvv+DwgfDuQ+ly99RARVSG6mi6gONOmTUObNm2wdOlSDBw4EOfPn8eGDRuwYcMGAIBEIsHUqVOxePFiuLu7w83NDXPnzoWTkxN69+6t2eKJypORY+nOu3cJiPgIgASo2QxwDwTqdwUcfPjAQiKqtir10nMA2L9/P+bMmYPbt2/Dzc0N06dPx5gxY5THBUHA/PnzsWHDBiQkJKBt27ZYt24d6tevX+rP4NJzqvTkMnHVVdpjFD5xRwIY2AEW44E7h4G466qHTe2Bel3E4FOnkzgRmoioiuN2EWXAsENVwsM94qorAKqBJ3ciT7vdgHNf8d+TngC3DwO3DwF3jwNZKXmnS3XFfbnqB4o9PzXcuTcXEVVJDDtlwLBDVcbDPcClKarLz42dAd9v8oLO63Iygfv/iMHn9iHgxR3V45auecGndltxUjQRURWgtrBTu3ZtjBo1CiNGjICLi8tbF1oZMOxQlfK2T1B+EZ0XfO6dztuQFBD36qrTAXDvKr4s+QwqIqq81BZ2vvnmG2zevBk3btxAp06dMHr0aPTp0wcGBgYlv7mSYtihaiszBYg5IQafqENA8hPV43YN84KPsx+gU6nXNBBRNaP2YazLly9j8+bN2L59O2QyGT744AOMGjUKzZo1e+OiNYVhhwiAIIibkUYdFOf7PDoPCPK844YWQN3O4pBXvQDApIbmaiUiQgXO2cnOzsa6deswe/ZsZGdnw9vbG5MnT8bIkSMhqSKTHhl2iAqR9hK4c1Ts9blzGEh/le+gBKjpmzvXpwuXthORRqg97GRnZ2Pv3r3YtGkTDh8+jFatWmH06NF49OgR1q5di3feeQfbtm174xuoSAw7RCWQy4BHF3Pn+hwsfGm7exdxuItL24mogqgt7Fy+fBmbNm3C9u3bIZVKERQUhP/85z/w8PBQnnPjxg20aNEC6enpb34HFYhhh6iM8i9tj/4byE7NOybVA1xb58714dJ2IlIftYUdHR0ddOnSBaNHj0bv3r2hp6dX4JzU1FQEBwdj06ZNZa9cAxh2iN5C/qXtUQeBl9Gqx61q5wUfLm0nonKktrBz//595d5U2oJhh6gcKZa2Rx0E7p9RXdquZwy4dRCHvOoHiru5ExG9IbWFnQsXLkAul8PPz0+lPSwsDDo6OmjevPmbVaxBDDtEaqJY2q5Y4VVgabtXXvCp1ZJL24moTNQWdlq2bImPP/4Y/fv3V2nfs2cPli9fjrCwsDerWIMYdogqgMrS9kPAowtc2k5Eb0VtYcfU1BTXrl1DnTp1VNpjYmLQuHFjJCcnv1nFGsSwQ6QByqXtB4E7Rwouba/VPO+Bho4+nORMRAWU9vd3mfuMDQwMEB8fXyDsxMbGQleXXdBEVErG1kDjAeJLubT9oPgk5/jrYs/PowvA30sAUwfAPUCc5Fy3E2BgpunqiagKKXPPzpAhQxAbG4vffvsNFhYWAICEhAT07t0bdnZ2+Pnnn9VSqDqxZ4eokkl6kreFxd3jRSxtDxSHvGzqsdeHqJpS2zDW48eP0b59e7x48QJNmzYFAISHh8Pe3h6HDx+Gs3PV2ziQYYeoEsvJFFd1ReU+0PDlXdXjVm7iUFf9roArl7YTVSdqfYJyamoqQkNDcfXqVRgZGaFx48YYMmRIoc/cqQoYdoiqkBfRuZOcDwL3zgDy7LxjiqXt9XPn+nBpO5FWq7C9sbQBww5RFZWZDNw9IQaf24eB5FjV43ZeucEnEKjVgkvbibSM2sPOzZs38eDBA2RlZam09+rV600up1EMO0RaQBDEPbsUk5wfXQCQ7483Q0ugXmcx+NQLAExsNFUpEZUTtYWdu3fvok+fPrh+/TokEgkUb1fscC6Tyd6ibM1g2CHSQqkvgOij4pDXnSNARkK+g4ql7YFiz49DY05yJqqC1BZ2evbsCR0dHfzf//0f3NzccP78ebx48QIzZszAl19+iXbt2r118RWNYYdIy8lygMcX857kHP/6ru0OeU9yrtORS9uJqgi1hZ0aNWrg2LFjaNy4MSwsLHD+/Hk0aNAAx44dw4wZM3DlypW3Lr6iMewQVTOJj8Wl7bcPF7G0vY0YfNy7cmk7USWmtocKymQymJmJf+upUaMGnjx5ggYNGsDV1RWRkZFvXjERUUWxqAk0Hym+cjKBe6fzNi99FSPu5xVzAjj4X3FpuyL4uPpzaTtRFVTmsNOoUSNcvXoVbm5u8PPzw4oVK6Cvr48NGzYUeKoyEVGlp2sgTlyu1xl4dznw/E7u6q5D4tL2VzFAWIj40jMWh7kU21hY1NR09URUCmUexjp48CBSU1PRt29f3LlzBz169EBUVBRsbGywc+dOvPPOO+qqVW04jEVEhcpMFoe5FENery9tt2+UF3y4tJ2owlXoc3ZevnwJKysr5YqsqoZhh4hKJAhA3LW8bSwKXdoeIA551e3Mpe1EFUAtYSc7OxtGRkYIDw9Ho0aNyqXQyoBhh4jKLPWFuKT99qGCS9slUqBm87wnOXNpO5FaqGWCsp6eHlxcXKrks3SIiMqViQ3gM0h8yXLEnp7bh8RX/A3g0XnxdWwxYOYoLm1378ql7UQaUOZhrI0bN2LPnj346aefYG1tra66KhR7doioXCU+Euf43Fbs2p6Wd0yqB9T2Fx9o6N4VqFFPY2USVXVqm7PTtGlT3LlzB9nZ2XB1dYWJiYnK8cuXL79ZxRrEsENEapOdAdw/LYYfxdL2/Kzr5AafLkDttuLqMCIqFbU9Z6d3795vUxcRUfWiZyhOXK4XAHT7HHhxJ++ZPvf/AV7eBcLWiy89k9yl7V24tJ2oHHHXc7Bnh4g0RLG0XbGNRUqc6nF777xJzrVaAFIdjZRJVFlV6NLzqo5hh4g0TrG0PeqQ+FDDRxehsrTdyErsHXLvKv7TWDvmTBK9DbWFHalUWuzzdKriSi2GHSKqdFKfA3eOisHnzhEgIzHvmEQq9vS4dxHn+zh4c2k7VUtqCzu//fabytfZ2dm4cuUKtmzZgoULF2L06NFvVrEGMewQUaWmXNp+UOz5efqv6nEzp9eWtptqpEyiilbhw1jbtm3Dzp07C4ShqoBhh4iqlMRHeU9yjjmhurRdR1/ctd09UHyas01dzdVJpGYVHnbu3r2Lxo0bIyUlpTwuV6EYdoioylIsbVfM9Xl1T/W4dV2xx6d+7q7tXNpOWqRCw056ejrmzJmDv/76C5GRkW97uQrHsENEWkEQxKXtUQfF4HP/H0Cek3dcsbRdscLL3EljpRKVB7U9Z+f1DT8FQUBycjKMjY2xdevWN6uWiIjenkQC1HAXX22CgYyk3F3bFUvb44HIP8QXkG9peyBQqzmXtpPWKnPPzubNm1XCjlQqha2tLfz8/GBlZVXuBVYE9uwQkdaTy/Pt2n4QeHwJhS9tDwTqdebSdqoS+JydMmDYIaJqJ/X5a7u2F7a0vas4ydm+EZe2U6WktrCzadMmmJqaYsCAASrtu3btQlpaGoYPH/5mFWsQww4RVWuyHHGHdsWTnIta2l4/EHDrwKXtVGmoLezUr18f3333HTp16qTSfuLECYwdO5YTlImIqrqEh2KPz+1DwN0TQE563jEdfXFVV/3cXdu5tJ00SG1hx9DQELdu3ULt2rVV2u/duwdPT0+kp6cX/sZKjGGHiKgI2RnAvdO54aeIpe2K4OPahkvbqUKpbTWWnZ0drl27ViDsXL16FTY2NmUulIiIKjE9Q8A9QHwJy4Hnt3NXdx3K3bU9Gji3Tnzpm+bu2t5VHPbi0naqJMocdoYMGYLJkyfDzMwM7du3ByAOYU2ZMgWDBw8u9wKJiKiSkEgA2/riq82k3KXtf+f2+uQubb+1X3wB4p5diic51/Tl0nbSmDIPY2VlZWHYsGHYtWsXdHXFrCSXyxEUFISQkBDo6+urpVB14jAWEdFbksuBuKti6Cl0abu1uLS9fiBQ9x0ubadyofal57dv30Z4eDiMjIzg7e0NV1fXNy5W0xh2iIjKWcqzvKXt0UcLWdreMu9JzlzaTm+Iz9kpA4YdIiI1kuUAD8PyVng9val63Lxm3q7tXNpOZaC2sNOvXz+0bNkSs2fPVmlfsWIFLly4gF27dr1ZxRrEsENEVIESHojDXUUtba/dVpzr496FS9upWGoLO7a2tjh27Bi8vb1V2q9fv46AgADEx8e/WcUaxLBDRKQh2enAvTPiCq+og0DCfdXjNvXygo+rP6Bb9eaFkvqobel5SkpKoZOQ9fT0kJSUVNbLERFRdaZnlLe0/d0VeUvbow4CD86Ku7i/uAOcW/va0vaugLmjpqunKqLMYcfb2xs7d+7EvHnzVNp37NiBhg0bllthRERUzRRY2p4o7toelTvXJ/Xpa0vbG+c90JBL26kYZQ47c+fORd++fREdHY133nkHAHD06FFs27YNu3fvLvcCiYiomjK0ABq+L74US9ujcp/k/PiyuIt73DXg5BeAsU3uru1dubSdCnij1Vh//PEHli5dqlx67uPjg/nz58Pa2hqNGjVSR51qxTk7RERVjHJp+0HgzjEg87Wl7c5+uSu8AgF7Ly5t11IVtvQ8KSkJ27dvx8aNG3Hp0iXIZLK3uZxGMOwQEVVhsmzg4fncuT6HgGcRqsfNa+Utba/TAdA30UydVO7UHnZOnjyJjRs34pdffoGTkxP69u2Lfv36oUWLFm9ctKYw7BARaZGEB+Icn6hDQMzJ15a2G+Qube8qPtTQuo7m6qS3ppawExcXh82bN2Pjxo1ISkrCwIEDERISgqtXr1bpyckMO0REWio7Xdy1Peqg2POT8ED1uI17XvBxacOl7VVMuYednj174uTJk+jevTuGDh2Kbt26QUdHB3p6egw7RERU+QkC8DwqN/gcEpe2y3PyjiuWtitWeJk5aKxUKp1yDzu6urqYPHkyxo8fD3d3d2U7ww4REVVJGYlAdL5d21Ofqh539Ml9pk8gULMZl7ZXQuX+UMHTp09j48aN8PX1haenJ4YNG4bBgweXS7FEREQVztAC8OotvuRyIDY8b/+ux5eB2Kvi6/Wl7fU6A0ZWGi6eyqLME5RTU1Oxc+dO/PDDDzh//jxkMhlWrlyJUaNGwczMTF11qhV7doiISEXKM+BO7v5dRS5t7yoOedk15NJ2DamQpeeRkZHYuHEjfvrpJyQkJKBLly7Yt2/fm15OYxh2iIioSLJscdf2qIPicFdRS9vrBwJu7bm0vQJV2HN2AEAmk+H333/HDz/8wLBDRETa7dX9vHk+MSeAnIy8Y4ql7YpJztZumquzGqjQsFNRPv/8c8yZMwdTpkzBN998AwDIyMjAjBkzsGPHDmRmZiIwMBDr1q2Dvb19qa/LsENERG8kOx2IOZUbfopY2q4IPi6tubS9nGld2Llw4QIGDhwIc3NzdOrUSRl2xo8fjz/++AObN2+GhYUFgoODIZVKcebMmVJfm2GHiIjemiAAzyLzJjkXWNpuBtTtKK7ucu/Cpe3lQKvCTkpKCpo1a4Z169Zh8eLFaNKkCb755hskJibC1tYW27ZtQ//+/QEAt27dgqenJ86ePYtWrVqV6voMO0REVO4yEoHoY+Jw1+1DQOoz1eOOPmLwqR8IODXl0vY3UO5LzzVp4sSJ6N69OwICArB48WJl+6VLl5CdnY2AgABlm4eHB1xcXIoNO5mZmcjMzFR+nZSUpL7iiYioejK0ALz6iC+5HIi9IgafqIPAk/xL21fkLm3vIj7Jue47XNpezip92NmxYwcuX76MCxcuFDgWFxcHfX19WFpaqrTb29sjLi6uyGsuW7YMCxcuLO9SiYiICieVAjV9xVfHT4CUp+Ku7VEHxd6ftBfAtR3iS6IjLm2v31Wc68Ol7W+tUoedhw8fYsqUKTh8+DAMDQ3L7bpz5szB9OnTlV8nJSXB2dm53K5PRERULFM7oMkH4ktlafsh4Nkt4ME/4uvIAsDCOXfXdsXSdmNNV1/lVOqwc+nSJTx9+hTNmjVTtslkMpw8eRLffvstDh48iKysLCQkJKj07sTHx8PBoeiJXwYGBjAwMFBn6URERKWjoycuV6/dFuj6Wb6l7bm7tic+BC7+IL50DAC3dnmTnLm0vVQq9QTl5ORk3L9/X6Vt5MiR8PDwwOzZs+Hs7AxbW1ts374d/fr1AyA+6NDDw4MTlImIqOrLShN3bb99EIg6BCS+trS9Rv3c/buq59J2rZigbGZmhkaNGqm0mZiYwMbGRtk+evRoTJ8+HdbW1jA3N8ekSZPQunXrUgcdIiKiSkvfWJy7U78r8J5iaXtu8HlwVtzF/XkUcPbb3KXtncTVXfW6AGalf96ctqvUYac0vv76a0ilUvTr10/loYJERERaRSIB7DzEl/8UID0BuPu3GHzuHBaXtkfsE18A4Ngk74GGTs3ESdLVVKUexqooHMYiIqIqTbG0PSr3Sc5PrqgeN66RO8m5C1C3M2BkqZEyy5tWPVRQ3Rh2iIhIq6Q8zX2Y4UEg+m8gM9/z5CQ6gEurvBVedp5Vdmk7w04ZMOwQEZHWkmUDD87lzfV5Hql63MI5b5JzFVvazrBTBgw7RERUbby6l/ck53unVHdt1zUEarcTg0/9roBVbU1VWSoMO2XAsENERNVSVpoYeBQPNEx8qHq8RgNxuKt+oLi0XUdPM3UWgWGnDBh2iIio2hME8enNiuDz4BwgyPKOG5gDdTqWbWm7XAY8OwWkxwJGjoBtu3Ld8JRhpwwYdoiIiF6TnqC6a3vac9XjTk1z5/oodm1/bWn7wz3ApSlA2qO8NuNagO8qwLlvuZTIsFMGDDtERETFkMvF5ey3c3t9ilzanrtr+/NjwKn+AF6PGLmrvtrtLpfAw7BTBgw7REREZZAcLz7IMCp3aXtWct4xiRSolwFIMop4s0Ts4ekV89ZDWlqxXQQRERFVQmb2QNMPxVdOFvDwnNjjE3UISP23mKADAAKQ9lCcy2PfsULKrb7PjiYiIqK3p6svPp+n62Ig+DzQY1np3pceq9668mHYISIiovJTo3HpzjNyVG8d+TDsEBERUfmxbSfOyUFRW1BIAGNn8bwKwrBDRERE5UeqIy4vB1Aw8OR+7ftNuT5vp8SSKuyTiIiIqHpw7isuLzeuqdpuXKvclp2XBVdjERERUflz7gvUfF+tT1AuLYYdIiIiUg+pToUtLy+2DE0XQERERKRODDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0GsMOERERaTWGHSIiItJqDDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0GsMOERERaTWGHSIiItJqDDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0GsMOERERaTWGHSIiItJqDDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0GsMOERERaTWGHSIiItJqDDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0GsMOERERaTWGHSIiItJqDDtERESk1Rh2iIiISKsx7BAREZFWY9ghIiIircawQ0RERFqNYYeIiIi0WqUOO8uWLUOLFi1gZmYGOzs79O7dG5GRkSrnZGRkYOLEibCxsYGpqSn69euH+Ph4DVVMRERElU2lDjsnTpzAxIkTce7cORw+fBjZ2dno2rUrUlNTledMmzYNv//+O3bt2oUTJ07gyZMn6Nu3rwarJiIiospEIgiCoOkiSuvZs2ews7PDiRMn0L59eyQmJsLW1hbbtm1D//79AQC3bt2Cp6cnzp49i1atWpXquklJSbCwsEBiYiLMzc3VeQtERERUTkr7+7tS9+y8LjExEQBgbW0NALh06RKys7MREBCgPMfDwwMuLi44e/ZskdfJzMxEUlKSyouIiIi0U5UJO3K5HFOnToW/vz8aNWoEAIiLi4O+vj4sLS1VzrW3t0dcXFyR11q2bBksLCyUL2dnZ3WWTkRERBpUZcLOxIkTcePGDezYseOtrzVnzhwkJiYqXw8fPiyHComIiKgy0tV0AaURHByM/fv34+TJk6hVq5ay3cHBAVlZWUhISFDp3YmPj4eDg0OR1zMwMICBgYE6SyYiIqJKolL37AiCgODgYOzduxfHjh2Dm5ubynFfX1/o6enh6NGjyrbIyEg8ePAArVu3ruhyiYiIqBKq1D07EydOxLZt2/Dbb7/BzMxMOQ/HwsICRkZGsLCwwOjRozF9+nRYW1vD3NwckyZNQuvWrUu9EouIiIi0W6Veei6RSApt37RpE0aMGAFAfKjgjBkzsH37dmRmZiIwMBDr1q0rdhjrdVx6TkREVPWU9vd3pQ47FYVhh4iIqOrRyufsEBEREZUVww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVbTmrCzdu1a1K5dG4aGhvDz88P58+c1XRIRERFVAloRdnbu3Inp06dj/vz5uHz5Mnx8fBAYGIinT59qujQiIiLSMK0IOytXrsSYMWMwcuRINGzYECEhITA2NsYPP/yg6dKIiIhIw6p82MnKysKlS5cQEBCgbJNKpQgICMDZs2c1WBkRERFVBrqaLuBtPX/+HDKZDPb29irt9vb2uHXrVqHvyczMRGZmpvLrxMREAEBSUpL6CiUiIqJypfi9LQhCsedV+bDzJpYtW4aFCxcWaHd2dtZANURERPQ2kpOTYWFhUeTxKh92atSoAR0dHcTHx6u0x8fHw8HBodD3zJkzB9OnT1d+LZfL8fLlS9jY2EAikai13oqQlJQEZ2dnPHz4EObm5pouR+2q2/0C1e+eeb/ajfer3dR5v4IgIDk5GU5OTsWeV+XDjr6+Pnx9fXH06FH07t0bgBhejh49iuDg4ELfY2BgAAMDA5U2S0tLNVda8czNzavF/0gK1e1+gep3z7xf7cb71W7qut/ienQUqnzYAYDp06dj+PDhaN68OVq2bIlvvvkGqampGDlypKZLIyIiIg3TirAzaNAgPHv2DPPmzUNcXByaNGmCAwcOFJi0TERERNWPVoQdAAgODi5y2Kq6MTAwwPz58wsM1Wmr6na/QPW7Z96vduP9arfKcL8SoaT1WkRERERVWJV/qCARERFRcRh2iIiISKsx7BAREZFWY9ghIiIircawo2UeP36MDz/8EDY2NjAyMoK3tzcuXryo6bLUQiaTYe7cuXBzc4ORkRHq1q2Lzz77rMQ9UqqKkydPomfPnnBycoJEIsGvv/6qclwQBMybNw+Ojo4wMjJCQEAAbt++rZliy0Fx95udnY3Zs2fD29sbJiYmcHJyQlBQEJ48eaK5gt9SSf998/voo48gkUjwzTffVFh96lCae46IiECvXr1gYWEBExMTtGjRAg8ePKj4YstBSfebkpKC4OBg1KpVC0ZGRmjYsCFCQkI0U+xbWrZsGVq0aAEzMzPY2dmhd+/eiIyMVDknIyMDEydOhI2NDUxNTdGvX78Cux+oC8OOFnn16hX8/f2hp6eHv/76Czdv3sRXX30FKysrTZemFsuXL8f69evx7bffIiIiAsuXL8eKFSuwZs0aTZdWLlJTU+Hj44O1a9cWenzFihVYvXo1QkJCEBYWBhMTEwQGBiIjI6OCKy0fxd1vWloaLl++jLlz5+Ly5cvYs2cPIiMj0atXLw1UWj5K+u+rsHfvXpw7d67Ex+FXBSXdc3R0NNq2bQsPDw8cP34c165dw9y5c2FoaFjBlZaPku53+vTpOHDgALZu3YqIiAhMnToVwcHB2LdvXwVX+vZOnDiBiRMn4ty5czh8+DCys7PRtWtXpKamKs+ZNm0afv/9d+zatQsnTpzAkydP0Ldv34opUCCtMXv2bKFt27aaLqPCdO/eXRg1apRKW9++fYWhQ4dqqCL1ASDs3btX+bVcLhccHByEL774QtmWkJAgGBgYCNu3b9dAheXr9fstzPnz5wUAwv379yumKDUq6n4fPXok1KxZU7hx44bg6uoqfP311xVem7oUds+DBg0SPvzwQ80UpGaF3a+Xl5ewaNEilbZmzZoJn376aQVWph5Pnz4VAAgnTpwQBEH880lPT0/YtWuX8pyIiAgBgHD27Fm118OeHS2yb98+NG/eHAMGDICdnR2aNm2K77//XtNlqU2bNm1w9OhRREVFAQCuXr2K06dP491339VwZeoXExODuLg4BAQEKNssLCzg5+eHs2fParCyipOYmAiJRKKV+9oB4h5/w4YNw6xZs+Dl5aXpctROLpfjjz/+QP369REYGAg7Ozv4+fkVO7xX1bVp0wb79u3D48ePIQgC/v77b0RFRaFr166aLu2tJSYmAgCsra0BAJcuXUJ2drbKn1keHh5wcXGpkD+zGHa0yN27d7F+/Xq4u7vj4MGDGD9+PCZPnowtW7ZoujS1+OSTTzB48GB4eHhAT08PTZs2xdSpUzF06FBNl6Z2cXFxAFBgSxR7e3vlMW2WkZGB2bNnY8iQIVq7keLy5cuhq6uLyZMna7qUCvH06VOkpKTg888/R7du3XDo0CH06dMHffv2xYkTJzRdnlqsWbMGDRs2RK1ataCvr49u3bph7dq1aN++vaZLeytyuRxTp06Fv78/GjVqBED8M0tfX7/AX04q6s8srdkugsQfsObNm2Pp0qUAgKZNm+LGjRsICQnB8OHDNVxd+fv5558RGhqKbdu2wcvLC+Hh4Zg6dSqcnJy08n5JlJ2djYEDB0IQBKxfv17T5ajFpUuXsGrVKly+fBkSiUTT5VQIuVwOAHj//fcxbdo0AECTJk3wzz//ICQkBB06dNBkeWqxZs0anDt3Dvv27YOrqytOnjyJiRMnwsnJSaUHpKqZOHEibty4gdOnT2u6FCX27GgRR0dHNGzYUKXN09Ozyq5kKMmsWbOUvTve3t4YNmwYpk2bhmXLlmm6NLVzcHAAgAIrGeLj45XHtJEi6Ny/fx+HDx/W2l6dU6dO4enTp3BxcYGuri50dXVx//59zJgxA7Vr19Z0eWpRo0YN6OrqVps/w9LT0/Hf//4XK1euRM+ePdG4cWMEBwdj0KBB+PLLLzVd3hsLDg7G/v378ffff6NWrVrKdgcHB2RlZSEhIUHl/Ir6M4thR4v4+/sXWOoXFRUFV1dXDVWkXmlpaZBKVX+EdXR0lH9D1GZubm5wcHDA0aNHlW1JSUkICwtD69atNViZ+iiCzu3bt3HkyBHY2NhouiS1GTZsGK5du4bw8HDly8nJCbNmzcLBgwc1XZ5a6Ovro0WLFtXmz7Ds7GxkZ2drzZ9hgiAgODgYe/fuxbFjx+Dm5qZy3NfXF3p6eip/ZkVGRuLBgwcV8mcWh7G0yLRp09CmTRssXboUAwcOxPnz57FhwwZs2LBB06WpRc+ePbFkyRK4uLjAy8sLV65cwcqVKzFq1ChNl1YuUlJScOfOHeXXMTExCA8Ph7W1NVxcXDB16lQsXrwY7u7ucHNzw9y5c+Hk5ITevXtrrui3UNz9Ojo6on///rh8+TL2798PmUymHOe3traGvr6+psp+YyX99309zOnp6cHBwQENGjSo6FLLTUn3PGvWLAwaNAjt27dHp06dcODAAfz+++84fvy45op+CyXdb4cOHTBr1iwYGRnB1dUVJ06cwI8//oiVK1dqsOo3M3HiRGzbtg2//fYbzMzMlP9/WlhYwMjICBYWFhg9ejSmT58Oa2trmJubY9KkSWjdujVatWql/gLVvt6LKtTvv/8uNGrUSDAwMBA8PDyEDRs2aLoktUlKShKmTJkiuLi4CIaGhkKdOnWETz/9VMjMzNR0aeXi77//FgAUeA0fPlwQBHH5+dy5cwV7e3vBwMBA6Ny5sxAZGanZot9CcfcbExNT6DEAwt9//63p0t9ISf99X6cNS89Lc88bN24U6tWrJxgaGgo+Pj7Cr7/+qrmC31JJ9xsbGyuMGDFCcHJyEgwNDYUGDRoIX331lSCXyzVb+Bso6v/PTZs2Kc9JT08XJkyYIFhZWQnGxsZCnz59hNjY2AqpT5JbJBEREZFW4pwdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKtxrBDREREWo1hh4iIiLQaww4RlZt79+5BIpEgPDxc06Uo3bp1C61atYKhoSGaNGmi6XJKJJFI8Ouvv2q6DCKtwrBDpEVGjBgBiUSCzz//XKX9119/rTa7Z79u/vz5MDExQWRkpMq+PPk9e/YM48ePh4uLCwwMDODg4IDAwECcOXOmgqslInVg2CHSMoaGhli+fDlevXql6VLKTVZW1hu/Nzo6Gm3btoWrq2uRm4f269cPV65cwZYtWxAVFYV9+/ahY8eOePHixRt/bmXxNt87Im3BsEOkZQICAuDg4IBly5YVec6CBQsKDOl88803qF27tvLrESNGoHfv3li6dCns7e1haWmJRYsWIScnB7NmzYK1tTVq1aqFTZs2Fbj+rVu30KZNGxgaGqJRo0Y4ceKEyvEbN27g3XffhampKezt7TFs2DA8f/5cebxjx44IDg7G1KlTUaNGDQQGBhZ6H3K5HIsWLUKtWrVgYGCAJk2a4MCBA8rjEokEly5dwqJFiyCRSLBgwYIC10hISMCpU6ewfPlydOrUCa6urmjZsiXmzJmDXr16qVxr/fr1ePfdd2FkZIQ6depg9+7dKtd6+PAhBg4cCEtLS1hbW+P999/HvXv3lMcvXLiALl26oEaNGrCwsECHDh1w+fLlQu9NYf78+XB0dMS1a9cAAKdPn0a7du1gZGQEZ2dnTJ48Gampqcrza9eujc8++wxBQUEwNzfH2LFjkZWVheDgYDg6OsLQ0BCurq7F/nwQaRuGHSIto6Ojg6VLl2LNmjV49OjRW13r2LFjePLkCU6ePImVK1di/vz56NGjB6ysrBAWFoaPPvoI48aNK/A5s2bNwowZM3DlyhW0bt0aPXv2VPaSJCQk4J133kHTpk1x8eJFHDhwAPHx8Rg4cKDKNbZs2QJ9fX2cOXMGISEhhda3atUqfPXVV/jyyy9x7do1BAYGolevXrh9+zYAIDY2Fl5eXpgxYwZiY2Mxc+bMAtcwNTWFqakpfv31V2RmZhb7/Zg7dy769euHq1evYujQoRg8eDAiIiIAANnZ2QgMDISZmRlOnTqFM2fOwNTUFN26dVP2riQnJ2P48OE4ffo0zp07B3d3d7z33ntITk4u8FmCIGDSpEn48ccfcerUKTRu3BjR0dHo1q0b+vXrh2vXrmHnzp04ffo0goODVd775ZdfwsfHB1euXMHcuXOxevVq7Nu3Dz///DMiIyMRGhqqEmyJtF6FbDdKRBVi+PDhwvvvvy8IgiC0atVKGDVqlCAIgrB3714h///u8+fPF3x8fFTe+/XXXwuurq4q13J1dRVkMpmyrUGDBkK7du2UX+fk5AgmJibC9u3bBUEQlLuTf/7558pzsrOzhVq1agnLly8XBEEQPvvsM6Fr164qn/3w4UMBgHLX9g4dOghNmzYt8X6dnJyEJUuWqLS1aNFCmDBhgvJrHx8fYf78+cVeZ/fu3YKVlZVgaGgotGnTRpgzZ45w9epVlXMACB999JFKm5+fnzB+/HhBEAThp59+Eho0aKCyY3VmZqZgZGQkHDx4sNDPlclkgpmZmfD777+rfM6uXbuEDz74QPD09BQePXqkPDZ69Ghh7NixKtc4deqUIJVKhfT0dEEQxN3Re/furXLOpEmThHfeeadK7qZNVB7Ys0OkpZYvX44tW7Yoex7ehJeXF6TSvD8m7O3t4e3trfxaR0cHNjY2ePr0qcr7Wrdurfx3XV1dNG/eXFnH1atX8ffffyt7VExNTeHh4QFAnF+j4OvrW2xtSUlJePLkCfz9/VXa/f39y3zP/fr1w5MnT7Bv3z5069YNx48fR7NmzbB58+Yi70vxdf77unPnDszMzJT3ZW1tjYyMDOV9xcfHY8yYMXB3d4eFhQXMzc2RkpKCBw8eqFx32rRpCAsLw8mTJ1GzZk1l+9WrV7F582aV711gYCDkcjliYmKU5zVv3lzleiNGjEB4eDgaNGiAyZMn49ChQ2X6/hBVdbqaLoCI1KN9+/YIDAzEnDlzMGLECJVjUqkUgiCotGVnZxe4hp6ensrXEomk0Da5XF7qulJSUtCzZ08sX768wDFHR0flv5uYmJT6muXB0NAQXbp0QZcuXTB37lz85z//wfz58wt874qSkpICX19fhIaGFjhma2sLABg+fDhevHiBVatWwdXVFQYGBmjdunWBScRdunTB9u3bcfDgQQwdOlTlM8aNG4fJkycX+AwXFxflv7/+vWvWrBliYmLw119/4ciRIxg4cCACAgIKzDki0lYMO0Ra7PPPP0eTJk3QoEEDlXZbW1vExcVBEATlkvTyfDbOuXPn0L59ewBATk4OLl26pJxX0qxZM/zyyy+oXbs2dHXf/I8gc3NzODk54cyZM+jQoYOy/cyZM2jZsuXb3QCAhg0bFnjezblz5xAUFKTyddOmTQGI97Vz507Y2dnB3Ny80GueOXMG69atw3vvvQdAnNCcf2K2Qq9evdCzZ0988MEH0NHRweDBg5WfcfPmTdSrV6/M92Nubo5BgwZh0KBB6N+/P7p164aXL1/C2tq6zNciqmo4jEWkxby9vTF06FCsXr1apb1jx4549uwZVqxYgejoaKxduxZ//fVXuX3u2rVrsXfvXty6dQsTJ07Eq1evMGrUKADAxIkT8fLlSwwZMgQXLlxAdHQ0Dh48iJEjR0Imk5Xpc2bNmoXly5dj586diIyMxCeffILw8HBMmTKl1Nd48eIF3nnnHWzduhXXrl1DTEwMdu3ahRUrVuD9999XOXfXrl344YcfEBUVhfnz5+P8+fPKEDd06FDUqFED77//Pk6dOoWYmBgcP34ckydPVk7gdnd3x08//YSIiAiEhYVh6NChMDIyKrSuPn364KeffsLIkSOVPTCzZ8/GP//8g+DgYISHh+P27dv47bffCkxQft3KlSuxfft23Lp1C1FRUdi1axccHBxgaWlZ6u8TUVXGsEOk5RYtWlRgmMnT0xPr1q3D2rVr4ePjg/Pnzxe6UulNff755/j888/h4+OD06dPY9++fahRowYAKHtjZDIZunbtCm9vb0ydOhWWlpYq84NKY/LkyZg+fTpmzJgBb29vHDhwAPv27YO7u3upr2Fqago/Pz98/fXXaN++PRo1aoS5c+dizJgx+Pbbb1XOXbhwIXbs2IHGjRvjxx9/xPbt29GwYUMAgLGxMU6ePAkXFxf07dsXnp6eGD16NDIyMpQ9PRs3bsSrV6/QrFkzDBs2DJMnT4adnV2RtfXv3x9btmzBsGHDsGfPHjRu3BgnTpxAVFQU2rVrh6ZNm2LevHlwcnIq9h7NzMywYsUKNG/eHC1atMC9e/fw559/lvn7TVRVSYTXB+6JiKgAiUSCvXv3onfv3pouhYjKiLGeiIiItBrDDhEREWk1rsYiIioFjvgTVV3s2SEiIiKtxrBDREREWo1hh4iIiLQaww4RERFpNYYdIiIi0moMO0RERKTVGHaIiIhIqzHsEBERkVZj2CEiIiKt9v8ZqcejCxB84gAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fivefft_train = [0.9731, 0.9439, 0.9915]\n",
        "fivefft_val = [0.8586, 0.6735, 0.8205]\n",
        "fivemfcc_train = [0.9359, 0.9712, 0.9979]\n",
        "fivemfcc_val = [0.5455, 0.5692, 0.6102]"
      ],
      "metadata": {
        "id": "FIjuvgYn5HBB"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "# Data\n",
        "fivefft_train = [0.9731, 0.9439, 0.9915]\n",
        "fivefft_val = [0.8586, 0.6735, 0.8205]\n",
        "fivemfcc_train = [0.9359, 0.9712, 0.9979]\n",
        "fivemfcc_val = [0.5455, 0.5692, 0.6102]\n",
        "\n",
        "# Calculate averages\n",
        "avg_fivefft_train = np.mean(fivefft_train)\n",
        "avg_fivemfcc_train = np.mean(fivemfcc_train)\n",
        "avg_fivefft_val = np.mean(fivefft_val)\n",
        "avg_fivemfcc_val = np.mean(fivemfcc_val)\n",
        "\n",
        "# Normalize to percentage\n",
        "avg_fivefft_train *= 100\n",
        "avg_fivemfcc_train *= 100\n",
        "avg_fivefft_val *= 100\n",
        "avg_fivemfcc_val *= 100\n",
        "\n",
        "# Plotting\n",
        "bar_width = 0.35\n",
        "index = np.arange(2)  # Two feature extraction methods\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "\n",
        "# Plotting training bars\n",
        "ax.bar(index - bar_width/2, [avg_fivefft_train, avg_fivemfcc_train], bar_width, label='Train')\n",
        "\n",
        "# Plotting validation bars\n",
        "ax.bar(index + bar_width/2, [avg_fivefft_val, avg_fivemfcc_val], bar_width, label='Validation')\n",
        "\n",
        "# Customize the plot\n",
        "ax.set_xlabel('Feature Extraction Method')\n",
        "ax.set_ylabel('Average Accuracy (%)')\n",
        "ax.set_title('Average Training and Validation Accuracy for Different Feature Extraction Methods')\n",
        "ax.set_xticks(index)\n",
        "ax.set_xticklabels(['FFT', 'MFCC'])\n",
        "\n",
        "# Move legend outside the graph\n",
        "ax.legend(bbox_to_anchor=(1.05, 1), loc='upper left')\n",
        "\n",
        "# Show the plot\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "fSl5Ac_I7D2q",
        "outputId": "ab463c88-db1c-4233-cd38-ece4cbcaf0e0"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GUsmqYZIgdc8"
      },
      "source": [
        "## Installation & Imports"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "U3U7NbTCXS6T",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9566b746-0196-47fc-ed8d-63223d7e6c80"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: tensorflow==2.8.0 in /usr/local/lib/python3.10/dist-packages (2.8.0)\n",
            "Requirement already satisfied: absl-py>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.4.0)\n",
            "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.6.3)\n",
            "Requirement already satisfied: flatbuffers>=1.12 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (23.5.26)\n",
            "Requirement already satisfied: gast>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (0.5.4)\n",
            "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (0.2.0)\n",
            "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (3.9.0)\n",
            "Requirement already satisfied: keras-preprocessing>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.1.2)\n",
            "Requirement already satisfied: libclang>=9.0.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (16.0.6)\n",
            "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.23.5)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (3.3.0)\n",
            "Requirement already satisfied: protobuf>=3.9.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (3.20.3)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (67.7.2)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.16.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (2.3.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (4.5.0)\n",
            "Requirement already satisfied: wrapt>=1.11.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.14.1)\n",
            "Requirement already satisfied: tensorboard<2.9,>=2.8 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (2.8.0)\n",
            "Requirement already satisfied: tf-estimator-nightly==2.8.0.dev2021122109 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (2.8.0.dev2021122109)\n",
            "Requirement already satisfied: keras<2.9,>=2.8.0rc0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (2.8.0)\n",
            "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (0.25.0)\n",
            "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.8.0) (1.59.3)\n",
            "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from astunparse>=1.6.0->tensorflow==2.8.0) (0.42.0)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (2.17.3)\n",
            "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (0.4.6)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (3.5.1)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (2.31.0)\n",
            "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (0.6.1)\n",
            "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (1.8.1)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.9,>=2.8->tensorflow==2.8.0) (3.0.1)\n",
            "Requirement already satisfied: cachetools<6.0,>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (5.3.2)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (0.3.0)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (4.9)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (1.3.1)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (2023.11.17)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from werkzeug>=0.11.15->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (2.1.3)\n",
            "Requirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (0.5.1)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.10/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.9,>=2.8->tensorflow==2.8.0) (3.2.2)\n",
            "Requirement already satisfied: tensorflow-io==0.25.0 in /usr/local/lib/python3.10/dist-packages (0.25.0)\n",
            "Requirement already satisfied: tensorflow-io-gcs-filesystem==0.25.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow-io==0.25.0) (0.25.0)\n"
          ]
        }
      ],
      "source": [
        "!pip3 install tensorflow==2.8.0\n",
        "!pip3 install tensorflow-io==0.25.0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uL3EdaUOhGUv"
      },
      "source": [
        "### Project imports"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SwG9JJB6U3OF",
        "outputId": "62780642-abc5-4211-df1a-d8324c076e81"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "TensorFlow version: 2.8.0\n",
            "TensorFlow IO version: 0.25.0\n",
            "True\n"
          ]
        }
      ],
      "source": [
        "import sys\n",
        "import os\n",
        "import random\n",
        "from datetime import datetime\n",
        "from pathlib import Path\n",
        "from typing import Tuple\n",
        "\n",
        "import librosa\n",
        "import numpy as np\n",
        "import keras\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib import cm\n",
        "\n",
        "import tensorflow as tf\n",
        "import tensorflow_io as tfio\n",
        "print(\"TensorFlow version:\", tf.__version__)\n",
        "print(\"TensorFlow IO version:\", tfio.__version__)\n",
        "tf.config.run_functions_eagerly(True)\n",
        "print(tf.executing_eagerly())\n",
        "%load_ext tensorboard"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-x6YLgMwXYmD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "18396b8d-c81a-49a4-8c32-81e4f92e873c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "ROOT_DIR='/content/drive/MyDrive/College/Research/Linh_2023_Research'\n",
        "\n",
        "DATASET_PATH=ROOT_DIR+'/test_data/vox'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OsF8O-L0VRZA"
      },
      "source": [
        "## Data Processing\n",
        "[Keras Speaker Recognition](https://keras.io/examples/audio/speaker_recognition_using_cnn/)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CXc8U7QY8Jwn"
      },
      "outputs": [],
      "source": [
        "SAMPLING_RATE = 16000\n",
        "\n",
        "def paths_and_labels_to_dataset(audio_paths, labels):\n",
        "    \"\"\"Constructs a dataset of audios and labels.\"\"\"\n",
        "    path_ds = tf.data.Dataset.from_tensor_slices(audio_paths)\n",
        "    audio_ds = path_ds.map(lambda x: path_to_audio(x))\n",
        "    label_ds = tf.data.Dataset.from_tensor_slices(labels)\n",
        "    return tf.data.Dataset.zip((audio_ds, label_ds))\n",
        "\n",
        "\n",
        "def path_to_audio(path):\n",
        "    \"\"\"Reads and decodes an audio file.\"\"\"\n",
        "    audio = tf.io.read_file(path)\n",
        "    audio, _ = tf.audio.decode_wav(audio, 1, SAMPLING_RATE)\n",
        "    return audio"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### FFT & MFCC Pre-processing"
      ],
      "metadata": {
        "id": "Ogn_zBISnt9W"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "N_MFCC = 13\n",
        "\n",
        "def audio_to_fft(audio):\n",
        "    # Since tf.signal.fft applies FFT on the innermost dimension,\n",
        "    # we need to squeeze the dimensions and then expand them again\n",
        "    # after FFT\n",
        "    audio = tf.squeeze(audio, axis=-1)\n",
        "    fft = tf.signal.fft(\n",
        "        tf.cast(tf.complex(real=audio, imag=tf.zeros_like(audio)), tf.complex64)\n",
        "    )\n",
        "    fft = tf.expand_dims(fft, axis=-1)\n",
        "\n",
        "    # Return the absolute value of the first half of the FFT\n",
        "    # which represents the positive frequencies\n",
        "    return tf.math.abs(fft[:, : (audio.shape[1] // 2), :])\n",
        "\n",
        "def audio_to_mfcc(audio):\n",
        "    audio = tf.squeeze(audio, axis=-1)\n",
        "    # Convert the audio to MFCC\n",
        "    stfts = tf.signal.stft(audio, frame_length=1024, frame_step=256, fft_length=1024)\n",
        "    spectrograms = tf.abs(stfts)\n",
        "\n",
        "    # Warp the linear scale spectrograms into the mel-scale\n",
        "    num_spectrogram_bins = stfts.shape[-1]\n",
        "    lower_edge_hertz, upper_edge_hertz = 80.0, 7600.0\n",
        "    linear_to_mel_weight_matrix = tf.signal.linear_to_mel_weight_matrix(\n",
        "        N_MFCC, num_spectrogram_bins, SAMPLING_RATE, lower_edge_hertz, upper_edge_hertz)\n",
        "    mel_spectrograms = tf.tensordot(spectrograms, linear_to_mel_weight_matrix, 1)\n",
        "    mel_spectrograms.set_shape(spectrograms.shape[:-1].concatenate(linear_to_mel_weight_matrix.shape[-1:]))\n",
        "\n",
        "    # Compute a stabilized log to get log-magnitude mel-scale spectrograms\n",
        "    log_mel_spectrograms = tf.math.log(mel_spectrograms + 1e-6)\n",
        "\n",
        "    # Compute MFCCs from log_mel_spectrograms and take the first N_MFCC\n",
        "    mfccs = tf.signal.mfccs_from_log_mel_spectrograms(log_mel_spectrograms)[..., :N_MFCC]\n",
        "    print(mfccs.shape)\n",
        "\n",
        "    return mfccs"
      ],
      "metadata": {
        "id": "4B5UlsIgnqTH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t8WW3Ke_gny_"
      },
      "source": [
        "### [VGGish Embedding Colab](https://colab.research.google.com/drive/1E3CaPAqCai9P9QhJ3WYPNCVmrJU4lAhF)\n",
        "\n",
        "Tensorflow implementation of VGGish model takes `numpy` array as input. While this is possible in eager mode (default mode of Tensorflow 2.x), we cannot run VGGish model in a dataset mapping function as mapping functions do not run in eager model."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install numpy scipy\n",
        "!pip install resampy tensorflow\n",
        "!pip install tf_slim\n",
        "!rm -rf models\n",
        "!git clone https://github.com/tensorflow/models.git"
      ],
      "metadata": {
        "id": "I4h29FcBHvwY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BEWOcr68gjyZ",
        "outputId": "cee1f527-a163-4f57-9858-839483c16f16"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content\n"
          ]
        }
      ],
      "source": [
        "# Check to see where are in the kernel's file system.\n",
        "!pwd"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YiVWsKdzgmQs",
        "outputId": "b84bbf5b-ae63-4073-c5e7-725b1db49396"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100  277M  100  277M    0     0  98.8M      0  0:00:02  0:00:02 --:--:-- 98.8M\n",
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100 73020  100 73020    0     0   156k      0 --:--:-- --:--:-- --:--:--  156k\n"
          ]
        }
      ],
      "source": [
        "# Grab the VGGish model\n",
        "!curl -O https://storage.googleapis.com/audioset/vggish_model.ckpt\n",
        "!curl -O https://storage.googleapis.com/audioset/vggish_pca_params.npz"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IwfNMNu6g7RI",
        "outputId": "274e4ccd-bf70-486b-f074-0311e153ee69"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "drive\t\t requirements.txt\t   vggish_model.ckpt\t  vggish_smoke_test.py\n",
            "mel_features.py  sample_data\t\t   vggish_params.py\t  vggish_train_demo.py\n",
            "models\t\t vggish_export_tfhub.py    vggish_pca_params.npz\n",
            "__pycache__\t vggish_inference_demo.py  vggish_postprocess.py\n",
            "README.md\t vggish_input.py\t   vggish_slim.py\n"
          ]
        }
      ],
      "source": [
        "# Make sure we got the model data.\n",
        "!ls"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qnYEND0vgSkK",
        "outputId": "6d9dd1d4-0fa0-462a-ada9-bf580f4bcb64"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "mel_features.py   vggish_export_tfhub.py    vggish_params.py\t   vggish_smoke_test.py\n",
            "README.md\t  vggish_inference_demo.py  vggish_postprocess.py  vggish_train_demo.py\n",
            "requirements.txt  vggish_input.py\t    vggish_slim.py\n"
          ]
        }
      ],
      "source": [
        "# Verify the location of the AudioSet source files\n",
        "!ls models/research/audioset/vggish"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a6BgFFjCg_7u"
      },
      "outputs": [],
      "source": [
        "# Copy the source files to the current directory.\n",
        "!cp models/research/audioset/vggish/* ."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mR6DF8CJhJri",
        "outputId": "35afd681-95da-4b5c-f3ef-555968edf4da"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "drive\t\t requirements.txt\t   vggish_model.ckpt\t  vggish_smoke_test.py\n",
            "mel_features.py  sample_data\t\t   vggish_params.py\t  vggish_train_demo.py\n",
            "models\t\t vggish_export_tfhub.py    vggish_pca_params.npz\n",
            "__pycache__\t vggish_inference_demo.py  vggish_postprocess.py\n",
            "README.md\t vggish_input.py\t   vggish_slim.py\n"
          ]
        }
      ],
      "source": [
        "# Make sure the source files got copied correctly.\n",
        "!ls"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZSrTZluHhMQ2"
      },
      "outputs": [],
      "source": [
        "# Run the test, which also loads all the necessary functions.\n",
        "from vggish_smoke_test import *"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import vggish_slim\n",
        "import vggish_params\n",
        "import vggish_input\n",
        "\n",
        "def create_vggish_net(sess, hop_size=0.96):   # Hop size is in seconds.\n",
        "    \"\"\"Define VGGish model, load the checkpoint, and return a dictionary that points\n",
        "    to the different tensors defined by the model.\n",
        "    \"\"\"\n",
        "    vggish_slim.define_vggish_slim()\n",
        "    checkpoint_path = 'vggish_model.ckpt'\n",
        "    vggish_params.EXAMPLE_HOP_SECONDS = hop_size\n",
        "\n",
        "    vggish_slim.load_vggish_slim_checkpoint(sess, checkpoint_path)\n",
        "\n",
        "    features_tensor = sess.graph.get_tensor_by_name(vggish_params.INPUT_TENSOR_NAME)\n",
        "    embedding_tensor = sess.graph.get_tensor_by_name(vggish_params.OUTPUT_TENSOR_NAME)\n",
        "\n",
        "    layers = {'conv1': 'vggish/conv1/Relu',\n",
        "            'pool1': 'vggish/pool1/MaxPool',\n",
        "            'conv2': 'vggish/conv2/Relu',\n",
        "            'pool2': 'vggish/pool2/MaxPool',\n",
        "            'conv3': 'vggish/conv3/conv3_2/Relu',\n",
        "            'pool3': 'vggish/pool3/MaxPool',\n",
        "            'conv4': 'vggish/conv4/conv4_2/Relu',\n",
        "            'pool4': 'vggish/pool4/MaxPool',\n",
        "            'fc1': 'vggish/fc1/fc1_2/Relu',\n",
        "            # 'fc2': 'vggish/fc2/Relu',\n",
        "            'embedding': 'vggish/embedding',\n",
        "            'features': 'vggish/input_features',\n",
        "            }\n",
        "    g = tf.compat.v1.get_default_graph()\n",
        "    for k in layers:\n",
        "        layers[k] = g.get_tensor_by_name( layers[k] + ':0')\n",
        "\n",
        "    return {'features': features_tensor,\n",
        "            'embedding': embedding_tensor,\n",
        "            'layers': layers,\n",
        "            }\n",
        "\n",
        "@tf.function\n",
        "def embeddings_from_vggish(sess, vgg, x, sr=SAMPLING_RATE):\n",
        "    \"\"\"Run the VGGish model, starting with a sound (x) at sample rate\n",
        "    (sr). Return a dictionary of embeddings from the different layers\n",
        "    of the model.\"\"\"\n",
        "    # Produce a batch of log mel spectrogram examples.\n",
        "    def waveform_to_examples_py(x_np, sr):\n",
        "        x_np = np.array(x_np)\n",
        "        return vggish_input.waveform_to_examples(x_np, sr)\n",
        "\n",
        "    # Use tf.py_function to execute the Python function\n",
        "    input_batch = tf.py_function(waveform_to_examples_py, [x, sr], tf.float32)\n",
        "\n",
        "    layer_names = vgg['layers'].keys()\n",
        "    tensors = [vgg['layers'][k] for k in layer_names]\n",
        "\n",
        "    def concrete_func(x, sr):\n",
        "        return sess.run(tensors, feed_dict={vgg['features']: x})\n",
        "\n",
        "    # Execute the concrete function with the input tensor evaluated\n",
        "    results = concrete_func(input_batch, sr)\n",
        "\n",
        "    resdict = {}\n",
        "    for i, k in enumerate(layer_names):\n",
        "        resdict[k] = results[i]\n",
        "\n",
        "    embedding = tf.convert_to_tensor(resdict['embedding'])\n",
        "    return embedding\n",
        "\n",
        "# Define the graph and load the model\n",
        "with tf.Graph().as_default():\n",
        "    sess = tf.compat.v1.Session()\n",
        "    vgg = create_vggish_net(sess, 0.01)\n",
        "\n",
        "# Iterate through dataset\n",
        "# ...\n",
        "\n",
        "# Close the session when done\n",
        "sess.close()"
      ],
      "metadata": {
        "id": "I0zh1PH3n88M"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### SpeechIdentity"
      ],
      "metadata": {
        "id": "FL_BvU0dIY3Q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/douglas125/SpeechIdentity.git"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VG-m3hHeIbSY",
        "outputId": "f66f67a7-3e49-40dd-bfea-1136f2cb43f7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'SpeechIdentity'...\n",
            "remote: Enumerating objects: 55, done.\u001b[K\n",
            "remote: Counting objects: 100% (55/55), done.\u001b[K\n",
            "remote: Compressing objects: 100% (55/55), done.\u001b[K\n",
            "remote: Total 55 (delta 11), reused 23 (delta 0), pack-reused 0\u001b[K\n",
            "Receiving objects: 100% (55/55), 8.66 MiB | 1.60 MiB/s, done.\n",
            "Resolving deltas: 100% (11/11), done.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!cp -r /content/SpeechIdentity/speech-id-model-110/ .\n",
        "!ls"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Iq1HerVIItyL",
        "outputId": "aef6c2ed-ddbe-4ff3-8d2e-937590715113"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "drive  sample_data  SpeechIdentity  speech-id-model-110\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import train_speech_id_model\n",
        "\n",
        "model = tf.keras.models.load_model('speech-id-model-110')\n",
        "cur_data = tfio.audio.AudioIOTensor(file)\n",
        "audio_data = cur_data.to_tensor()[:, 0]\n",
        "# set batch size to 1, extract first element\n",
        "audio_embedding = model.predict(\n",
        "\ttf.expand_dims(audio_data, axis=0)\n",
        ")[0]"
      ],
      "metadata": {
        "id": "uUap9O53HSb-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### SpeechT5"
      ],
      "metadata": {
        "id": "VRPjNtbEf-H8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install git+https://github.com/huggingface/transformers.git\n",
        "!pip install datasets\n",
        "!pip install sentencepiece"
      ],
      "metadata": {
        "id": "LUkuQE0-gZI6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install transformers"
      ],
      "metadata": {
        "id": "cULOPqmmlTvD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import SpeechT5Processor, SpeechT5ForSpeechToText\n",
        "\n",
        "processor = SpeechT5Processor.from_pretrained(\"microsoft/speecht5_asr\")\n",
        "model = SpeechT5ForSpeechToText.from_pretrained(\"microsoft/speecht5_asr\")"
      ],
      "metadata": {
        "id": "IKabrHj-gBNU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Dataset preparation"
      ],
      "metadata": {
        "id": "AUfHLUxcn7NS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Create dataset"
      ],
      "metadata": {
        "id": "ddNperMNvrWk"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rSzqjnpz4uhM",
        "outputId": "ce32aecc-3327-487f-a308-72b8f755ec50"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing speaker id10004\n",
            "Processing speaker id10836\n",
            "Processing speaker id10964\n",
            "Processing speaker id10365\n",
            "Processing speaker id10022\n",
            "Found 699 files belonging to 395 classes.\n",
            "Using 560 files for training.\n",
            "Using 139 files for validation.\n"
          ]
        }
      ],
      "source": [
        "BATCH_SIZE = 128\n",
        "SHUFFLE_SEED = 43\n",
        "TRAIN_VALID_SPLIT = 0.2\n",
        "EPOCHS = 100\n",
        "N_CLASS = 5\n",
        "\n",
        "class_names = os.listdir(DATASET_PATH)\n",
        "random.shuffle(class_names)\n",
        "audio_paths = []\n",
        "labels = []\n",
        "for label, name in enumerate(class_names):\n",
        "    if label > N_CLASS - 1: break\n",
        "    print(\"Processing speaker {}\".format(name,))\n",
        "    dir_path = Path(DATASET_PATH) / name\n",
        "    speaker_sample_paths = [\n",
        "        os.path.join(dir_path, filepath)\n",
        "        for filepath in os.listdir(dir_path)\n",
        "        if filepath.endswith(\".wav\")\n",
        "    ]\n",
        "    audio_paths += speaker_sample_paths\n",
        "    labels += [label] * len(speaker_sample_paths)\n",
        "\n",
        "print(\n",
        "    \"Found {} files belonging to {} classes.\".format(len(audio_paths), len(class_names))\n",
        ")\n",
        "\n",
        "# Shuffle\n",
        "rng = np.random.RandomState(SHUFFLE_SEED)\n",
        "rng.shuffle(audio_paths)\n",
        "rng = np.random.RandomState(SHUFFLE_SEED)\n",
        "rng.shuffle(labels)\n",
        "\n",
        "# Split into training and validation\n",
        "num_val_samples = int(TRAIN_VALID_SPLIT * len(audio_paths))\n",
        "print(\"Using {} files for training.\".format(len(audio_paths) - num_val_samples))\n",
        "train_audio_paths = audio_paths[:-num_val_samples]\n",
        "train_labels = labels[:-num_val_samples]\n",
        "\n",
        "print(\"Using {} files for validation.\".format(num_val_samples))\n",
        "valid_audio_paths = audio_paths[-num_val_samples:]\n",
        "valid_labels = labels[-num_val_samples:]\n",
        "\n",
        "# Create 2 datasets, one for training and the other for validation\n",
        "train_ds = paths_and_labels_to_dataset(train_audio_paths, train_labels)\n",
        "train_ds = train_ds.shuffle(buffer_size=BATCH_SIZE * 8, seed=SHUFFLE_SEED).batch(\n",
        "    BATCH_SIZE\n",
        ")\n",
        "\n",
        "valid_ds = paths_and_labels_to_dataset(valid_audio_paths, valid_labels)\n",
        "valid_ds = valid_ds.shuffle(buffer_size=32 * 8, seed=SHUFFLE_SEED).batch(32)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Data transformation"
      ],
      "metadata": {
        "id": "ZkJ3BqwLvt34"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_ds = train_ds.map(\n",
        "    lambda x, y: (audio_to_fft(x), y), num_parallel_calls=tf.data.AUTOTUNE\n",
        ")\n",
        "train_ds = train_ds.prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "valid_ds = valid_ds.map(\n",
        "    lambda x, y: (audio_to_fft(x), y), num_parallel_calls=tf.data.AUTOTUNE\n",
        ")\n",
        "valid_ds = valid_ds.prefetch(tf.data.AUTOTUNE)"
      ],
      "metadata": {
        "id": "VBfBOqArvl6E"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Data visualization (MFCC)"
      ],
      "metadata": {
        "id": "AB82g1OeGw2r"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "frTFswK4CGRr"
      },
      "outputs": [],
      "source": [
        "# for i, data in enumerate(train_ds):\n",
        "#     if i > 0: break\n",
        "#     x, y = data\n",
        "#     # Extract the chosen sample\n",
        "#     selected_sample_np = x[i].numpy()\n",
        "\n",
        "#     # Display the MFCC for the selected sample\n",
        "#     plt.imshow(selected_sample_np, cmap='viridis', origin='lower', aspect='auto')\n",
        "#     plt.title(f'MFCC for Sample {i}')\n",
        "#     plt.xlabel('MFCC Coefficient')\n",
        "#     plt.ylabel('Time Step')\n",
        "#     plt.colorbar(label='Magnitude')\n",
        "#     plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9jlin8sbJRYj"
      },
      "source": [
        "## Model\n",
        "\n",
        "MFCC\n",
        "\n",
        "FFT (focus on low freq) ---> CNN (max pool in one direction)\n",
        "\n",
        "Is speaker unique in consonant or vowel?\n",
        "\n",
        "To try: Cifar, Transfer learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iyhOcSlQeYdw"
      },
      "source": [
        "### Simple ResNet"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "imt11oYweVzh"
      },
      "outputs": [],
      "source": [
        "def residual_block(x, filters, conv_num=3, activation=\"relu\"):\n",
        "    # Shortcut\n",
        "    s = keras.layers.Conv1D(filters, 1, padding=\"same\")(x)\n",
        "    for i in range(conv_num - 1):\n",
        "        x = keras.layers.Conv1D(filters, 3, padding=\"same\")(x)\n",
        "        x = keras.layers.Activation(activation)(x)\n",
        "    x = keras.layers.Conv1D(filters, 3, padding=\"same\")(x)\n",
        "    x = keras.layers.Add()([x, s])\n",
        "    x = keras.layers.Activation(activation)(x)\n",
        "    return keras.layers.MaxPool1D(pool_size=2, strides=2)(x)\n",
        "\n",
        "\n",
        "def simple_resnet(input_shape, num_classes):\n",
        "    inputs = keras.layers.Input(shape=input_shape, name=\"input\")\n",
        "\n",
        "    x = residual_block(inputs, 16, 2)\n",
        "    x = residual_block(x, 32, 2)\n",
        "    x = residual_block(x, 64, 3)\n",
        "    x = residual_block(x, 128, 3)\n",
        "    x = residual_block(x, 128, 3)\n",
        "\n",
        "    x = keras.layers.Flatten()(x)\n",
        "    # x = keras.layers.Dense(256, activation=\"relu\")(x)\n",
        "    x = keras.layers.Dense(128, activation=\"relu\")(x)\n",
        "\n",
        "    outputs = keras.layers.Dense(num_classes, activation=\"softmax\", name=\"output\")(x)\n",
        "\n",
        "    return keras.models.Model(inputs=inputs, outputs=outputs)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XnFLlceDEwmQ"
      },
      "outputs": [],
      "source": [
        "# model = simple_resnet((SAMPLING_RATE//2, 1), N_CLASS)\n",
        "model = simple_resnet((59, N_MFCC), N_CLASS)\n",
        "\n",
        "model.summary()\n",
        "\n",
        "# Compile the model using Adam's default learning rate\n",
        "model.compile(\n",
        "    optimizer=\"Adam\", loss=\"sparse_categorical_crossentropy\", metrics=[\"accuracy\"]\n",
        ")\n",
        "\n",
        "# Add callbacks:\n",
        "# 'EarlyStopping' to stop training when the model is not enhancing anymore\n",
        "# 'ModelCheckPoint' to always keep the model that has the best val_accuracy\n",
        "model_save_filename = \"model.h5\"\n",
        "\n",
        "earlystopping_cb = keras.callbacks.EarlyStopping(patience=10, restore_best_weights=True)\n",
        "mdlcheckpoint_cb = keras.callbacks.ModelCheckpoint(\n",
        "    model_save_filename, monitor=\"val_accuracy\", save_best_only=True\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xpMTXaQmJlRo"
      },
      "source": [
        "## Training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VRecLUo_ELVz"
      },
      "source": [
        "`fit()` is for training the model with the given inputs (and corresponding training labels).\n",
        "\n",
        "`evaluate()` is for evaluating the already trained model using the validation (or test) data and the corresponding labels. Returns the loss value and metrics values for the model.\n",
        "\n",
        "`predict()` is for the actual prediction. It generates output predictions for the input samples."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "33HkHl5aJnO2"
      },
      "outputs": [],
      "source": [
        "# Define the Keras TensorBoard callback\n",
        "logdir=\"logs/fit/\" + datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
        "tensorboard_callback = keras.callbacks.TensorBoard(log_dir=logdir)\n",
        "\n",
        "# Train the model\n",
        "history = model.fit(\n",
        "    train_ds,\n",
        "    epochs=EPOCHS,\n",
        "    validation_data=valid_ds,\n",
        "    callbacks=[earlystopping_cb, mdlcheckpoint_cb, tensorboard_callback],\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CUHcoTI4j1en"
      },
      "outputs": [],
      "source": [
        "%tensorboard --logdir logs"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LF-Lf3hsLCZB"
      },
      "source": [
        "## Evaluate"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yYufHcUMLD_C"
      },
      "outputs": [],
      "source": [
        "print(model.evaluate(valid_ds))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-fsVZtGsEZ8z"
      },
      "outputs": [],
      "source": [
        "print(model.predict(valid_ds))"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyN5dC6y50arRwHlAtYGt8Q5"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}