diff --git a/1-neural_networks.ipynb b/1-neural_networks.ipynb index a3e834a..ba229e4 100644 --- a/1-neural_networks.ipynb +++ b/1-neural_networks.ipynb @@ -181,7 +181,7 @@ }, { "cell_type": "code", - "execution_count": 72, + "execution_count": null, "id": "e538eebc", "metadata": { "slideshow": { @@ -223,7 +223,7 @@ }, { "cell_type": "code", - "execution_count": 73, + "execution_count": null, "id": "626bcc7b", "metadata": { "slideshow": { @@ -238,20 +238,33 @@ "np.random.seed(SEED)" ] }, + { + "cell_type": "markdown", + "id": "bff3655e", + "metadata": { + "slideshow": { + "slide_type": "slide" + } + }, + "source": [ + "## Accelerated computing\n", + "\n", + "- **Accelerated computing** = when we add extra hardware to accelerate computation, like GPUs (needed in deep machine learning).\n", + "- **GPU**: many \"not-so-intelligent\" cores that are parallelizable. They can carry out specific operations in a very efficient way, e.g. tensor cores perform very efficient sparse tensor multiplication.\n", + "\n", + "We will be working with torch tensors in this notebook! instead of the usual numpy arrays. This means you could execute this code on a GPU if you have access to one with a simple command `torch.device(\"cuda\")`." + ] + }, { "cell_type": "code", - "execution_count": 74, + "execution_count": null, "id": "acb09c91", - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Using cpu device\n" - ] + "metadata": { + "slideshow": { + "slide_type": "-" } - ], + }, + "outputs": [], "source": [ "# Use GPU if available\n", "device = ('cuda' if torch.cuda.is_available()\n", @@ -359,37 +372,16 @@ }, { "cell_type": "code", - "execution_count": 75, + "execution_count": null, "id": "b27267d5", "metadata": { "slideshow": { "slide_type": "-" } }, - "outputs": [ - { - "data": { - "text/plain": [ - "Text(0.5, 1.0, 'Function to be fitted')" - ] - }, - "execution_count": 75, - "metadata": {}, - "output_type": "execute_result" - }, - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ - "sample_points = 1e2\n", + "sample_points = 1e3\n", "x_lim = 100\n", "x = np.linspace(0, x_lim, int(sample_points))\n", "y = np.sin(x * x_lim * 1e-4) * np.cos(x * x_lim * 1e-3) * 3\n", @@ -417,7 +409,7 @@ }, { "cell_type": "code", - "execution_count": 76, + "execution_count": null, "id": "d964b9bd", "metadata": { "slideshow": { @@ -432,19 +424,10 @@ }, { "cell_type": "code", - "execution_count": 77, + "execution_count": null, "id": "cda2d1f8", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "100\n", - "100\n" - ] - } - ], + "outputs": [], "source": [ "print(len(x))\n", "print(len(y))" @@ -464,7 +447,7 @@ }, { "cell_type": "code", - "execution_count": 78, + "execution_count": null, "id": "436f80a6", "metadata": { "slideshow": { @@ -479,24 +462,14 @@ }, { "cell_type": "code", - "execution_count": 79, + "execution_count": null, "id": "b669f0a9", "metadata": { "slideshow": { "slide_type": "-" } }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "(100,) (100,)\n", - "(100, 1) (100, 1)\n", - "10.101010101010102 [10.1010101]\n" - ] - } - ], + "outputs": [], "source": [ "# Uncomment to check the shape change\n", "print(x.shape, y.shape)\n", @@ -506,7 +479,7 @@ }, { "cell_type": "code", - "execution_count": 80, + "execution_count": null, "id": "2844eee1", "metadata": {}, "outputs": [], @@ -554,7 +527,7 @@ }, { "cell_type": "code", - "execution_count": 81, + "execution_count": null, "id": "340bc8b8", "metadata": {}, "outputs": [], @@ -565,19 +538,10 @@ }, { "cell_type": "code", - "execution_count": 82, + "execution_count": null, "id": "95a39f64", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "float64 float64\n", - "torch.float64 torch.float64\n" - ] - } - ], + "outputs": [], "source": [ "# Type checking:\n", "print(x.dtype, y.dtype)\n", @@ -598,7 +562,7 @@ }, { "cell_type": "code", - "execution_count": 83, + "execution_count": null, "id": "a2a41119", "metadata": {}, "outputs": [], @@ -609,19 +573,10 @@ }, { "cell_type": "code", - "execution_count": 84, + "execution_count": null, "id": "74760bf6", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "float64 float64\n", - "torch.float32 torch.float32\n" - ] - } - ], + "outputs": [], "source": [ "# Type checking:\n", "print(x.dtype, y.dtype)\n", @@ -630,31 +585,10 @@ }, { "cell_type": "code", - "execution_count": 85, + "execution_count": null, "id": "f74ae34d", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "[]" - ] - }, - "execution_count": 85, - "metadata": {}, - "output_type": "execute_result" - }, - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "plt.plot(x_torch.numpy(), y_torch.numpy())" ] @@ -688,7 +622,7 @@ }, { "cell_type": "code", - "execution_count": 86, + "execution_count": null, "id": "da256a5c", "metadata": {}, "outputs": [], @@ -699,31 +633,10 @@ }, { "cell_type": "code", - "execution_count": 87, + "execution_count": null, "id": "fa4d6625", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "Text(0.5, 1.0, 'Normalized function')" - ] - }, - "execution_count": 87, - "metadata": {}, - "output_type": "execute_result" - }, - { - "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGJCAYAAABYRTOkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAAA9hAAAPYQGoP6dpAABW30lEQVR4nO3dd1hUV/4G8Hf60AZEOlIElCLYFbHEhj0mmmZsUaOmqL/djcnuapoaEzVZN2WN0aixxJg11RQ1BizELjasgApiAWkiXYZh5v7+QFgJooAz3Bnm/TyPzxMu9975zsmUl3PPPUciCIIAIiIiIpFIxS6AiIiIrBvDCBEREYmKYYSIiIhExTBCREREomIYISIiIlExjBAREZGoGEaIiIhIVAwjREREJCqGESIiIhIVwwgRmUS/fv3Qr1+/6p/T0tIgkUiwfv36Jq1j8uTJ8Pf3r9e+//rXvxAQEACZTIaOHTuatK7GiouLg0QiQVxcnNilEBkNwwiRSNavXw+JRAK1Wo309PRav+/Xrx/Cw8NFqMw6xcTE4B//+Ad69eqFdevWYdGiRaLW89lnnzV5cCMSi1zsAoisnVarxZIlS7Bs2TKxSzEpPz8/3L59GwqFQuxS7mn37t2QSqX44osvoFQqxS4Hn332GVxcXDB58uQa2x955BHcvn3bLGokMhb2jBCJrGPHjli9ejUyMjJM9hiCIOD27dsmO399VPUCyWQyUeuoS3Z2NmxsbMz+S14qlUKtVkMq5cc3NR98NROJ7PXXX4der8eSJUseuG9FRQUWLlyIwMBAqFQq+Pv74/XXX4dWq62xn7+/Px599FH8/vvv6Nq1K2xsbPD5559Xjzf49ttvsWDBAnh7e8PBwQFPPfUUCgoKoNVq8be//Q1ubm6wt7fHlClTap173bp1GDBgANzc3KBSqRAWFoYVK1Y8sPY/jxmpquVe//48xuO3335Dnz59YGdnBwcHB4wYMQLnzp2r9Rg//fQTwsPDoVarER4eji1btjywLqAyKK1btw4lJSXVNaxfv/6+41wkEgnmz59f/fP8+fMhkUhw6dIlTJ48GU5OTnB0dMSUKVNQWlpa6/ivvvoK3bt3h62tLVq0aIFHHnkEMTExACr//507dw5//PFHdT1V42/qGjPy3XffoUuXLrCxsYGLiwsmTJhQ6/Lf5MmTYW9vj/T0dIwaNQr29vZwdXXFa6+9Br1eX6+2IjIFXqYhElnr1q3x3HPPYfXq1ZgzZw68vLzq3HfatGnYsGEDnnrqKbz66qs4cuQIFi9ejMTExFpfvMnJyRg7dixefPFFTJ8+HcHBwdW/W7x4MWxsbDBnzhxcunQJy5Ytg0KhgFQqxa1btzB//nwcPnwY69evR+vWrfH2229XH7tixQq0a9cOjz32GORyOX799VfMmDEDBoMBM2fOrPfzDg0NxcaNG2tsy8/Px+zZs+Hm5la9bePGjZg0aRKGDBmC999/H6WlpVixYgV69+6NkydPVgeXmJgYPPnkkwgLC8PixYtx8+ZNTJkyBa1atXpgLRs3bsSqVasQHx+PNWvWAAB69uxZ7+dyt2eeeQatW7fG4sWLceLECaxZswZubm54//33q/dZsGAB5s+fj549e+Kdd96BUqnEkSNHsHv3bgwePBgff/wx/u///g/29vZ44403AADu7u51Pub69esxZcoUdOvWDYsXL0ZWVhY++eQTHDhwACdPnoSTk1P1vnq9HkOGDEFkZCSWLl2KnTt34t///jcCAwPx8ssvN+o5Ez00gYhEsW7dOgGAcPToUSElJUWQy+XCX/7yl+rf9+3bV2jXrl31zwkJCQIAYdq0aTXO89prrwkAhN27d1dv8/PzEwAIO3bsqLHvnj17BABCeHi4UF5eXr197NixgkQiEYYNG1Zj/6ioKMHPz6/GttLS0lrPZciQIUJAQECNbX379hX69u1b/fPly5cFAMK6devu2R4Gg0F49NFHBXt7e+HcuXOCIAhCUVGR4OTkJEyfPr3GvpmZmYKjo2ON7R07dhQ8PT2F/Pz86m0xMTECgFrP4V4mTZok2NnZ1dh2v5oBCPPmzav+ed68eQIA4fnnn6+x3+jRo4WWLVtW/3zx4kVBKpUKo0ePFvR6fa02qNKuXbsa7Vel6v/hnj17BEEQhPLycsHNzU0IDw8Xbt++Xb3f1q1bBQDC22+/XeM5AhDeeeedGufs1KmT0KVLl1qPRdRUeJmGyAwEBARg4sSJWLVqFW7cuHHPfbZv3w4AmD17do3tr776KgBg27ZtNba3bt0aQ4YMuee5nnvuuRoDSSMjIyEIAp5//vka+0VGRuLatWuoqKio3mZjY1P93wUFBcjNzUXfvn2RmpqKgoKCBz3VOi1cuBBbt27F+vXrERYWBgCIjY1Ffn4+xo4di9zc3Op/MpkMkZGR2LNnDwDgxo0bSEhIwKRJk+Do6Fh9zkGDBlWfq6m89NJLNX7u06cPbt68icLCQgCVl5IMBgPefvvtWuM+JBJJgx/v2LFjyM7OxowZM6BWq6u3jxgxAiEhIbVeF3XVmJqa2uDHJjIWhhEiM/Hmm2+ioqKizrEjV65cgVQqRVBQUI3tHh4ecHJywpUrV2psb926dZ2P5evrW+Pnqi9wHx+fWtsNBkONkHHgwAFER0fDzs4OTk5OcHV1xeuvvw4AjQ4jO3bswIIFCzB37lw8+eST1dsvXrwIABgwYABcXV1r/IuJiUF2djYAVD/3Nm3a1Dr33ZenmsKf27ZFixYAgFu3bgEAUlJSIJVKjRaSqp77vZ5nSEhIrdeFWq2Gq6trrRqr6iMSA8eMEJmJgIAATJgwAatWrcKcOXPq3K++fz3f3YPxZ3Xd0VLXdkEQAFR+kQ4cOBAhISH48MMP4ePjA6VSie3bt+Ojjz6CwWCoV213u3z5MsaPH49Bgwbh3XffrfG7qvNt3LgRHh4etY6Vy037EVZXW99vsOeD2lBs5no3E1k3hhEiM/Lmm2/iq6++qjHYsYqfnx8MBgMuXryI0NDQ6u1ZWVnIz8+Hn5+fyev79ddfodVq8csvv9ToAai6XNJQt2/fxhNPPAEnJyf897//rXXZIjAwEADg5uaG6OjoOs9T9dyrelLulpyc3KjagP/1auTn59fY/ufehoYIDAyEwWDA+fPn7zvLa31DZ9VzT05OxoABA2r8Ljk5uUleF0QPi5dpiMxIYGAgJkyYgM8//xyZmZk1fjd8+HAAwMcff1xj+4cffgigcoyAqVX9VX33X/kFBQVYt25do8730ksv4cKFC9iyZUv1F//dhgwZAo1Gg0WLFkGn09X6fU5ODgDA09MTHTt2xIYNG2pcKoqNjcX58+cbVRsAaDQauLi4YO/evTW2f/bZZ40+56hRoyCVSvHOO+/U6km6u13t7OxqhaB76dq1K9zc3LBy5coat2H/9ttvSExMbJLXBdHDYs8IkZl54403sHHjRiQnJ6Ndu3bV2zt06IBJkyZh1apVyM/PR9++fREfH48NGzZg1KhR6N+/v8lrGzx4MJRKJUaOHIkXX3wRxcXFWL16Ndzc3OoceFuXbdu24csvv8STTz6J06dP4/Tp09W/s7e3x6hRo6DRaLBixQpMnDgRnTt3xrPPPgtXV1dcvXoV27ZtQ69evfDpp58CqLxdecSIEejduzeef/555OXlYdmyZWjXrh2Ki4sb/ZynTZuGJUuWYNq0aejatSv27t2LCxcuNPp8QUFBeOONN7Bw4UL06dMHTzzxBFQqFY4ePQovLy8sXrwYANClSxesWLEC7777LoKCguDm5lar5wMAFAoF3n//fUyZMgV9+/bF2LFjq2/t9ff3xyuvvNLoWomaCsMIkZkJCgrChAkTsGHDhlq/W7NmDQICArB+/Xps2bIFHh4emDt3LubNm9cktQUHB+P777/Hm2++iddeew0eHh54+eWX4erqWutOnAep6tX44Ycf8MMPP9T4nZ+fH0aNGgUAGDduHLy8vLBkyRL861//glarhbe3N/r06YMpU6ZUHzN06FB89913ePPNNzF37lwEBgZi3bp1+Pnnnx9qUbm3334bOTk5+P777/Htt99i2LBh+O2332rMhdJQ77zzDlq3bo1ly5bhjTfegK2tLdq3b4+JEyfWeNwrV67ggw8+QFFREfr27XvPMAJUTmZma2uLJUuW4J///Cfs7OwwevRovP/++zXmGCEyVxLBXEZVERERkVXimBEiIiISFcMIERERiYphhIiIiETFMEJERESiYhghIiIiUTGMEBERkag4z8gDGAwGZGRkwMHBoVErahIREVkrQRBQVFQELy+vWss93I1h5AEyMjJqrWRKRERE9Xft2jW0atWqzt8zjDyAg4MDgMqG1Gg0RjmnTqdDTEwMBg8eDIVCYZRzWjO2p/GxTY2PbWpcbE/jM0WbFhYWwsfHp/q7tC4MIw9QdWlGo9EYNYzY2tpCo9HwTWQEbE/jY5saH9vUuNiexmfKNn3QMAcOYCUiIiJRMYwQERGRqBhGiIiISFQMI0RERCQqhhEiIiISFcMIERERiYphhIiIiETFMEJERESiYhghIiIiUXEGViIiM2YwCDibUYDsQi0EAAZBgCAI0Ngo0Nm3BdQKmdglEj00hhEiIjNTVKbDvou52J2UjbjkbOQWl99zPxuFDL2CXDAgxA0DQtzg4ahu4kqJjINhhIjITBRrK/DZnktYs/8yyisM1dsdVHIEuNpBIpFAKqlc5+NaXimyi7TYmZiFnYlZAIAnO7fCnGEhcHVQifUUiBqFYYSISGQGg4AfT6bj/R1JyCnSAgD8W9piYKg7Boa6oZu/MxSymkP8BEHAuYxCxCVnY3dSNk5czccPJ64j5lwmXhnUFs9F+UEu47BAsgwMI0REIkq8UYg5P5zGqesFAAC/lrZ4c0QYokPd7rvSqUQiQbi3I8K9HTFrQBucvHoLb/98DmfSC/DO1vP45ug1vP9Ue3T0cWqiZ0LUeIzNREQiiTmXiSc+O4hT1wtgr5Jj7rAQxLzyCAaFuT9wyfU/6+TbAj/N7IVFoyPgZKtAclYRxnx+CL+fyzRR9UTGwzBCRNTEBEHAqr0pePGr47it06N3kAt2v9YXL/YNhEre+LtjZFIJxkX6Ys+r/TAwxA3aCgNe/uo4vj5y1YjVExkfwwgRURMqrzBgzg9nsGh7EgQBGB/pi3VTusHNwXh3wrSwU+LziV0wpqsPDALw+pYz+HjnBQiCYLTHIDImjhkhImoiOgPwwlcncSDlJqQS4K1HwzC5p3+DL8nUh1wmxZInI+CuUeE/uy/h450XkVusxcLHw03yeEQPg2GEiKgJ6A0Cvrwoxem8m7BTyvDpuM7oH+Jm0seUSCSYPTgYrho13v75LL46fBWejjaY2T/IpI9L1FC8TENEZGKCIGDer4k4nSeFQibB6ue6mjyI3G1iDz8sfDwcAPCv35Ox4ywHtZJ5YRghIjKxD2Mv4Jtj1yGBgA+fbo+eQS5NXsOEHn6YFOUHAHjlmwScyyho8hqI6sIwQkRkQusOXMay3ZcAAE8HGDC0nbtotbz1aBj6tHHBbZ0e0zccQ3ZRmWi1EN2NYYSIyET2JGfjna3nAQB/GxiEXu7i3s0il0nx6bjOCHC1Q0ZBGV748jjKdHpRayICGEaIiEwiI/82Zn+TAEEAxnb3xYy+rcUuCQDgaKPAF5O6wdFGgYRr+fho5wWxSyJiGCEiMjad3oD/++9J3CrVIdxbg/mPhZnV7bStXeyw9OkOAIDVe1Nx4uotkSsia8cwQkRkZEt/T8bxK7fgoJJj+bjODzWrqqkMCnPH6E7eMAjA3787xcs1JCqGESIiI9p5Pguf700FAPzr6fbwa2knckV1mzcyDK4OKqTklOCjWF6uIfEwjBARGcn1W6V49btTAIApvfwxNNxT5Iruz8lWicWjIwAAq/fxcg2Jh2GEiMgIBEHAnB/OoOC2Dh18nDB3WKjYJdVLdJg7nrhzueY1Xq4hkTCMEBEZwbfHrmH/pVyo5FJ8PKYjlHLL+XidN7Id3BxUSM0pwbLdF8Uuh6yQ5bxbiIjMVFZhGd7dlggAeHVwW7R2Md9xIvfiaKvAwlGV08Wv3ncZ12+VilwRWRuGESKihyAIAt7YchZFZRXo0MoRz/cyj/lEGmpwmDuiAlqivMKAD3Yki10OWRmGESKih/Dr6RvYmZgFhUyCD57qALnMMj9WJRIJ3hgRCokE+OVUBk5yMCs1Ict81xARmYG8knLM/+UcAGBGvyAEeziIXNHDCfd2xFOdWwEA3t2WCEEQd/p6sh4MI0REjfTu1vPIKylHsLsDZvYPErsco3htSDBsFDIcv3IL289kil0OWQmLCyPLly+Hv78/1Go1IiMjER8fX6/jNm/eDIlEglGjRpm2QCKyCsev5OHHk+mQSIAlT0ZY1N0z9+OuUePFvgEAgCU7EnmrLzUJi3r3fPPNN5g9ezbmzZuHEydOoEOHDhgyZAiys7Pve1xaWhpee+019OnTp4kqJaLmzGAQsODXytV4n+7SCp18W4hckXG98EgA3DUqXMu7jfUH08Quh6yARYWRDz/8ENOnT8eUKVMQFhaGlStXwtbWFmvXrq3zGL1ej/Hjx2PBggUICAhowmqJqLn6/vh1nL5eAAeVHH8fEiJ2OUZnq/zf81oRl4KiMp3IFVFzJxe7gPoqLy/H8ePHMXfu3OptUqkU0dHROHToUJ3HvfPOO3Bzc8PUqVOxb9++Bz6OVquFVqut/rmwsBAAoNPpoNMZ5w1ZdR5jnc/asT2Nj21at6IyHd7fkQQAmNk/AE5qab3aydLa9NFwN3y2xxapuaXYcOAyXnzEvG5ZtrT2tASmaNP6nstiwkhubi70ej3c3d1rbHd3d0dSUtI9j9m/fz+++OILJCQk1PtxFi9ejAULFtTaHhMTA1tb2wbV/CCxsbFGPZ+1Y3saH9u0tp/SpLhZIoWbWoDrrfPYvv18g463pDbt6SRBaq4MK/ZcgFtBIlTmt/iwRbWnpTBmm5aW1m8CPYsJIw1VVFSEiRMnYvXq1XBxcan3cXPnzsXs2bOrfy4sLISPjw8GDx4MjUZjlNp0Oh1iY2MxaNAgKBQKo5zTmrE9jY9tem+pOSXYd+QgAAGLnu6Mvm1d632sJbbpYL0Be/9zAFfzbiO/ZTtM6ekndknVLLE9zZ0p2rTq6sKDWEwYcXFxgUwmQ1ZWVo3tWVlZ8PDwqLV/SkoK0tLSMHLkyOptBoMBACCXy5GcnIzAwMBax6lUKqhUqlrbFQqF0V/wpjinNWN7Gh/btKYlv19AhUHAgBA3RLfzatQ5LKlNFQpgZv8g/POHM1izPw3P9WwNtcK8ukcsqT0thTHbtL7nsZgBrEqlEl26dMGuXbuqtxkMBuzatQtRUVG19g8JCcGZM2eQkJBQ/e+xxx5D//79kZCQAB8fn6Ysn4gs3MFLudiTnAOFTII3R1jGirzGMLpTK3g72SC7SItvj10TuxxqpiymZwQAZs+ejUmTJqFr167o3r07Pv74Y5SUlGDKlCkAgOeeew7e3t5YvHgx1Go1wsPDaxzv5OQEALW2ExHdjyAIeP/3yvVaxnX3RYCrvcgVNR2lXIqX+gbgrZ/PYWVcCp7t5tts5lQh82FRYWTMmDHIycnB22+/jczMTHTs2BE7duyoHtR69epVSKV8kxCRccWcz8Kpa/mwVcowa0Absctpck939cGy3ZeQUVCGH09cx7PdfcUuiZoZiwojADBr1izMmjXrnr+Li4u777Hr1683fkFE1KzpDQL+dadXZGrv1nB1qD2mrLlTK2R44ZEAvLstEZ/FpeCpLq0sdkFAMk98NRER3cePJ67jUnYxnGwVmP6I9U6cOD7SD852SlzNK0XM+awHH0DUAAwjRER1KNPp8fHOiwCAGf0CoVFb710bNkoZxkdWXp5Zd+CyyNVQc8MwQkRUh01HriI9/zY8NGo8F+Uvdjmim9DDD3KpBEfTbuFseoHY5VAzwjBCRHQPRWU6LN9zCQDwt+g2Zje/hhjcNWqMaO8JAFjL3hEyIoYRIqJ72HAwDXkl5QhwscNTXVqJXY7ZmNKrco2araduIKdI+4C9ieqHYYSI6E+KtRVYs7/yL/+/DGzDO0fu0tHHCZ18nVCuN2DTkStil0PNBN9hRER/8uWhNOSX6hDgYoeRHRo37XtzVtU78tXhq9BW6EWuhpoDhhEioruUaCuwem8qAGDWgCDIpBKRKzI/w8I94KFRI7dYi22nb4hdDjUDDCNERHfZePgKbpXq4N/SFo+xV+SeFDIpJkZVruC77kAaBEEQuSKydAwjRER3lJbf3SvCsSL3M7a7L1RyKc6kF+D4lVtil0MWju80IqI7vjp8BTdLyuHX0hajOrJX5H6c7ZR4/E4bfX3kqsjVkKVjGCEiAnC7XI9Vd3pFZvYPYq9IPYy9s2DetjM3UFCqE7kasmR8txERAdh05Apyi8vh42yD0Z28xS7HInT0cUKwuwO0FQb8fCpd7HLIgjGMEJHV01bosWZf5bwiM/oFQcFekXqRSCR4trsPAOC/8dc4kJUaje84IrJ6P5/MQGZhGdw1KjzRmb0iDTG6kzeUcikSbxTi9HWuV0ONwzBCRFZNbxCwcm8KAGBa7wCo5FyDpiGcbJUYFu4BANh89JrI1ZClYhghIqsWcy4TqTklcLRRYGykr9jlWKRnu1W22y8J6SjRVohcDVkihhEislqCIGDFH5W9IpOi/GCvkotckWXqEeCM1i52KCnXY+vpDLHLIQvEMEJEVutgyk2cvl4AtUKKST39xS7HYkkkEozpVjmQlZdqqDEYRojIan0WdwlA5WWGlvYqkauxbE92bgW5VIKTV/ORnFkkdjlkYRhGiMgqnbqWjwOXbkIulWBan9Zil2PxXB1UiA51BwBsPsoZWalhGEaIyCqtvDNW5LGOXmjVwlbkapqHqks1vyRkQKc3iFwNWRKGESKyOmm5JdhxLhMA8FLfQJGraT76tHGBi70SN0vKse9ijtjlkAVhGCEiq/PF/ssQBGBAiBvaujuIXU6zIZdJMbJD5eJ5P57g9PBUfwwjRGRV8krK8d3xyjs+pvcJELma5ueJTq0AALHns1BYxsXzqH4YRojIqnx1+ArKdAZEeDuiR4Cz2OU0O+HeGgS52UNbYcCOM5lil0MWgmGEiKxGmU6PDQfTAADTHwmARCIRt6BmSCKRVK96/OPJ6yJXQ5aCYYSIrMaWk+m4WVIObycbDL+zngoZ36g7YeRwah7S82+LXA1ZAoYRIrIKBoOA1ftSAQBTevlDLuPHn6l4O9lUXwL76SQHstKD8d1IRFZhd1I2UnNK4KCW49nuXBDP1KoGsm45mQ5BEESuhswdwwgRWYVVd3pFxkX6ckG8JjAswgMquRSXsotxNr1Q7HLIzDGMEFGzd/p6PuIv50EulWBKT0793hQc1AoMCqucHp4DWelBGEaIqNn7Yv9lAMDIDl7wcFSLXI31qLqr5tdTGajg9PB0HwwjRNSsZRaUYdvpGwCAqb3ZK9KUHmnriha2CuQWl+PI5TyxyyEzxjBCRM3ahkNpqDAI6N7aGeHejmKXY1UUMimG3rmFeuvpDJGrIXPGMEJEzVZpeQW+PlK5nP009oqI4tH2lWvV/HY2kyv5Up0YRoio2frh+HUU3NbBr6UtBoa6i12OVYps7QwXeyXyS3U4cClX7HLITDGMEFGzZDAIWHsgDQAwpac/ZFJO/S4GuUyKYeGeAICtd8buEP0ZwwgRNUt7krNxObdykrOnu/qIXY5Ve7R9ZRj5/VwmtBV6kashc8QwQkTNUtXtvGO7+8KOk5yJqpu/M9w1KhSVVWDfBV6qodoYRoio2Um8UYiDKTchk0owqae/2OVYPalUguERlb0jv/KuGroHhhEianbW3ukVGRruAW8nG5GrIeB/d9XsPJ+FMh0v1VBNDCNE1KzkFmvxc0LlX9+c5Mx8dPZ1greTDUrK9diTlC12OWRmGEaIqFnZdPgqyvUGdPRxQmffFmKXQ3dIJBKMaM+7aujeGEaIqNnQVuix8fAVAMDz7BUxO1V31exKykKJtkLkasicWFwYWb58Ofz9/aFWqxEZGYn4+Pg69129ejX69OmDFi1aoEWLFoiOjr7v/kRk2baeuoHcYi08NGoMuzMNOZmPCG9H+DrbokxnQFxyjtjlkBmxqDDyzTffYPbs2Zg3bx5OnDiBDh06YMiQIcjOvvf1x7i4OIwdOxZ79uzBoUOH4OPjg8GDByM9Pb2JKyciUxMEAWsPVA5cfa6nHxQyi/p4swoSiaQ6JP52lpdq6H8s6t364YcfYvr06ZgyZQrCwsKwcuVK2NraYu3atffcf9OmTZgxYwY6duyIkJAQrFmzBgaDAbt27WriyonI1OIv5+FcRiHUCinGdvMVuxyqQ9XCeXuSsnlXDVWzmJmAysvLcfz4ccydO7d6m1QqRXR0NA4dOlSvc5SWlkKn08HZ2bnOfbRaLbRabfXPhYWFAACdTgedTtfI6muqOo+xzmft2J7GZ4ltumZfKgBgVEcv2CslZle7JbapKYS528Fdo0JWoRZxSZkYGOLWqPOwPY3PFG1a33NZTBjJzc2FXq+Hu3vNxa7c3d2RlJRUr3P885//hJeXF6Kjo+vcZ/HixViwYEGt7TExMbC1tW1Y0Q8QGxtr1PNZO7an8VlKm+aWATsTZQAkaF2ehu3b08QuqU6W0qamFGwrRVahFGtjTkCb+nAr+bI9jc+YbVpaWlqv/SwmjDysJUuWYPPmzYiLi4Nara5zv7lz52L27NnVPxcWFlaPNdFoNEapRafTITY2FoMGDYJCoTDKOa0Z29P4LK1NF/2WDAFX0CeoJZ5/qovY5dyTpbWpKbW8nIe9a48huViJQUP6NWp8D9vT+EzRplVXFx7EYsKIi4sLZDIZsrKyamzPysqCh8f9R80vXboUS5Yswc6dO9G+ffv77qtSqaBSqWptVygURn/Bm+Kc1oztaXyW0KZFZTp8d7xyUPrzfQLMvl5LaFNTiwpyQ0s7JW6WlOP4tUL0aePa6HOxPY3PmG1a3/NYzABWpVKJLl261Bh8WjUYNSoqqs7jPvjgAyxcuBA7duxA165dm6JUImpC3x+/jmJtBQJc7dD3Ib7UqOnIpBIMbld5yf23s5kiV0PmwGLCCADMnj0bq1evxoYNG5CYmIiXX34ZJSUlmDJlCgDgueeeqzHA9f3338dbb72FtWvXwt/fH5mZmcjMzERxcbFYT4GIjEhvELD+YBoAYEqv1pBKJeIWRPU2NLxyArSYc1nQGwSRqyGxWcxlGgAYM2YMcnJy8PbbbyMzMxMdO3bEjh07qge1Xr16FVLp//LVihUrUF5ejqeeeqrGeebNm4f58+c3ZelEZAK7k7Jx5WYpNGo5nuzsLXY51ABRAS3hoJYjt1iL41duoXvruu9ypObPosIIAMyaNQuzZs265+/i4uJq/JyWlmb6gohINFWr846N9IWt0uI+zqyaUi7FoFB3/HgyHb+dvcEwYuUs6jINEVGV8xmFOJR6EzKpBM9F+YtdDjVC1QRov5/NhCDwUo01YxghIou07s7U70PDPeDtZCNyNdQYj7R1ha1ShoyCMpy+XiB2OSQihhEisji5xVr8nJABAHi+F1fntVRqhQz9gytnYN1xjnfVWDOGESKyOJsOX0W53oAOPk7o7Oskdjn0EKpu8Y09n/WAPak5YxghIouirdBj4+ErAIDne/lDIuHtvJasX7Ab5FIJLmUXIzWH0y5YK4YRIrIovyRkILdYCw+NGsMjPMUuhx6So40CPQJaAmDviDVjGCEiiyEIAr64czvvcz39GrWmCZkfXqohvpOJyGIcSrmJpMwi2ChkGNfdV+xyyEiiQyvDyPGrt5BTpBW5GhIDwwgRWYw1d3pFnurSCk62SpGrIWPxcrJBhLcjBAHYncTeEWvEMEJEFiElpxi7k7IhkQBTevmLXQ4Z2aAwXqqxZgwjRGQRqiY5GxjihgBXe5GrIWOrGjey72IuSssrRK6GmhrDCBGZvVsl5fj++HUAwPO9OclZcxTs7gAfZxtoKwzYeyFX7HKoiTGMEJHZ+zr+Ksp0BoR5ahB15zZQal4kEgkGhVauVRNznrOxWhuGESIya+UVBnx5KA0AMLV3a05y1oxVXarZnZSNCr1B5GqoKXHNbSuk0xuQW6xFTpEWxdqa12blUilcHVRwc1DBTsWXB4lv6+kMZBVq4eqgwsgOXmKXQybU1a8FnGwVyC/V4diVW9WToVHzx2+bZuxWSTkSbxTi/I1CJN4oQnJWIW7klyGvtBz1Wa3bXiWHm0aFEA8HtG/lhPatHBHh7QgHtcL0xROhcpKzVXtTAQCTe/pDKWdnbnMml0kxMMQdP5y4jphzWQwjVoRhpBkpLNPhSGoeDlzKxcGUXFzIqnudB7lUAhd7FRzUctzd611eYUBOkRYl5XoUaytQnFOB1JwSbD9TeQ1XIgHaeWkwOMwDQ9p5oK27PbvNyWT2X8pFUmYRbJUyTIj0E7scagKDwirDyM7ELLz1aCg/X6wEw4iFyy4sw45zmdh+5gaOpt2C3lCzy8PX2Rahng4I9dQg1FMDv5a2cLVXoYWtElJp3W/yYm0FsgvLkJFfhrMZBTh9PR+nrhUgPf82zqYX4mx6IT6MvQD/lrYYFuGJcd194eNsa+qnS1amqldkTDcfONqyR84a9GnjAqVMiqt5pbiUXYw27g5il0RNgGHEAuWXluPnhAxsPZ2BY1du1bjk4t/SFj2DXNAr0AVRgS3hbNe4WSrtVXLYu9ojwNUevdu4VG/PLirDnqRs/H4uC/sv5iLtZilWxKVg5R8pGBjijkk9/dA7yIV/zdBDO59RiH0XcyGVAM/34u281sJOJUfPoJaIS87BzsRshhErwTBiIfQGAQcu5eLbY9cQcy4L5XeNNO/k64Th4Z4YGu5h8t4JNwc1xnTzxZhuvijWViAuORvfHL2GfRdzsTMxCzsTsxDoaoe/DGyDke297tv7QnQ/a/ZV9ooMj/Bkr5uVGRjqjrjkHOxKzMLL/QLFLoeaAMOImcsrKcfmo1ex6fBVpOffrt4e6qnBk529MTzCE15ONqLUZq+S49H2Xni0vRcuZRdj46E0fH/8OlJySvDXzQlY+Ucq/jEkGP2CXdlTQg1yo+A2fjmVAQB44ZEAkauhpjYwxA1voXLhvJvFWrS0V4ldEpkYw4iZOptegA0H0/DzqQyUV1T2gmjUcozq5I1nuvog3NtR5AprCnKzx4LHw/HakGB8eegKVsalIPFGIaasP4ru/s54Y0QoOvg4iV0mWYj1B9JQYRAQ2doZ7Vs5iV0ONTEvJxuEeWpw/kYh9iTn4KkurcQuiUyMYcSMGAwC4i5kY9XeVBxOzaveHuHtiEk9/fFoe0+oFTIRK3wwB7UCM/sHYVx3X6z4IwXrD6YhPi0Poz47gMk9/fHa4GDOX0L3VVSmw9dHrgJgr4g1iw51w/kbhdiVmMUwYgX4rWAGtBV6/HwyA6v3peJiduXtuHKpBMMiPDG5pz86+zpZ3GWOFnZKvD48FFN6+WPJb0n4OSED6w6kIeZcFt4dFY7+IW5il0hmanP8NRRpKxDkZo/+wXydWKuBoe74z+5L2HshB9oKPVRy8/5DjB4Ow4iIirUV+O7QVazZdxnZRVoAleMwxkX6Ykovf3g6ijMWxJg8HW3wybOdMLqTN9786Syu37qNKeuP4vGOXlg4KhwaTqBGd9FW6LFmf+XA1el9WnMAtBWL8HaEq4MKOUVaHEnNwyNtXcUuiUyI0xmK4GaxFtuuStF36V4s2p6E7CItPDRqvD48BAfnDsDrw0ObRRC5W79gN8S88kjlF4wE+DkhA8M/2YeTV2+JXRqZkS0n0pFVWPl+GN2JXfPWTCqVIDq0smdsV2KWyNWQqTGMNLGDKbno++99iEmXorCsAgGudvjgqfbY+4/+eOGRwGbdU2CrlOONEWH44eWeaNXCBtdv3cbTKw9h5R8pMBjqMT89NWt6g4DP70xyNq1Pa079ThgYUrlw3s7EbAj1WcOCLBbf7U2so48TbBQy+NoJ+PTZDoh9pS+e6epjVR+8nXxbYPtf+2BEe09UGAQs+S0Jk9bF41ZJudilkYh+P5eJy7klcLRRYGx3X7HLITPQK8gFKrkU6fm3kZxVJHY5ZELW8w1oJmyVcvw8owdmR+gxpJ07ZFZ6TVyjVuDTsZ2w5IkIqBVS7LuYi8eW78cFfuBYJUEQ8FncJQDApJ7+vOOKAAA2Shl6B1XOAL3zPC/VNGcMIyLwcrKBhd0cYxISiQTPdvfFzzN7w9fZFtfybmP08gP80LFC+y/l4mx6IdQKKSb39Be7HDIjA0P/d6mGmi+GERJdsIcDfp7ZCz0CnFFSrsf0jcfwWdwlXiO2IiviUgAAz3bzbfR6StQ8DbwziPXU9Xzk3LnrkJofhhEyCy3slNg4NRITevhCEIAPdiTjnz+cRsVda/BQ85RwLR8HU25CLpVgOic5oz9x16gR4e0IQQDiktk70lwxjJDZUMikeHdUBBY+3g5SCfDtset4ceNx3C7Xi10amdBneyrHijze0RveIq2zROatapLE3UkMI80VwwiZnYlR/vh8Yleo5FLsSsrG+DWHeadNM5V4oxAx57MgkQAv92OvCN3bwDthZO+FnOq1uqh5YRghszQozB2bpkXC0UaBE1fz8fTnh2qsWkzNw7LdFwEAIyI8EeTmIHI1ZK4ivB3hYq9CSbke8ZfzHnwAWRyGETJbXf2d8d1LUfB0VONSdjGeWXkIV26WiF0WGUlyZhG2n8kEAPzfgDYiV0PmTCqVYEBI5XTwvFTTPDGMkFlr6+6AH17uiQBXO6Tn38Yznx9CSk6x2GWREXx6Z6zIsHAPBHuwV4Tub8Cd2Vh3JWXxTrtmiGGEzJ6Xkw02v9ADbdzskVWoxZjPD+MiJ0ezaJeyi7H1dAYA9opQ/fRu4wKlTIorN0uRmsse0uaGYYQsgpuDGptf6IEQDwfkFmvx7KrDSLxRKHZZ1Eif7r4IQQAGh7kjzEsjdjlkAexVckQGOAMAdnMCtGaHYYQsRkt7Ff47vQfCvTW4WVKOsasZSCxRak4xfjlV2Svyl4HsFaH6G3DnrppdSZylublhGCGL0sJOiU3TeqCDjxPyS3WYsOYILmVzDIklWb4nBQah8nbNcG9HscshC1IVRo6m3ULhbZ3I1ZAxMYyQxXG0UeDL57tX95BMWn8c2bzr1yKk5BTjp4R0AMD/sVeEGsivpR2C3OyhNwjYd+mm2OWQETGMkEVytFFg4/ORCPFwQHaRFsvPy3DtVqnYZdEDfBR7AXqDgOhQN3T0cRK7HLJAVROg7UnOEbkSMiaGEbJYLeyU+GpaJAJd7ZBfLsFza4/hRgG7SMzVuYwCbD19AwDw6uBgkashS1V1qWbvxVwYeIdvs8EwQhbNxV6FL6d0hatawPX8MkxYcwQ3i7mypzn6d8wFAMBjHbwQ6sk7aKhxuvi1gEYtx61SHdJ4h3+zYXFhZPny5fD394darUZkZCTi4+Pvu/93332HkJAQqNVqREREYPv27U1UKTUVNwcVZobp4emoRkpOCSavO4qiMg5uMyfH0vKwOykbMqkErwxqK3Y5ZMHkMin6Blf2jpzLt7ivMKpDg/9PTpo0CXv37jVFLQ/0zTffYPbs2Zg3bx5OnDiBDh06YMiQIcjOvvc95wcPHsTYsWMxdepUnDx5EqNGjcKoUaNw9uzZJq6cTK2FClg/qQta2ilxJr0AUzccQ5mOq/2aA0EQ8MHvyQCAZ7q2QmsXO5ErIktXNTX8+VsSkSshY2lwGCkoKEB0dDTatGmDRYsWIT093RR13dOHH36I6dOnY8qUKQgLC8PKlStha2uLtWvX3nP/Tz75BEOHDsXf//53hIaGYuHChejcuTM+/fTTJquZmk6Aqx02PN8dDio54i/nYcamE9DpucKn2PZdzEX85Two5VLOtkpG0betG6QSIKNUggwuoNksyBt6wE8//YScnBxs3LgRGzZswLx58xAdHY2pU6fi8ccfh0KhMEWdKC8vx/HjxzF37tzqbVKpFNHR0Th06NA9jzl06BBmz55dY9uQIUPw008/1fk4Wq0WWu3/xhwUFlZOqqXT6aDTGafrv+o8xjqftbu7PYPdbPH5hE6YsuE4didlY/Y3J7H0yQhIpfwLqiGM9RoVBAEf7EgCAIzv7gNXO7nVvu75vjceB6UEHbw1OHm9ELsSszDRyUbskpoFU7xG63uuBocRAHB1dcXs2bMxe/ZsnDhxAuvWrcPEiRNhb2+PCRMmYMaMGWjTxrh/AeXm5kKv18Pd3b3Gdnd3dyQlJd3zmMzMzHvun5mZWefjLF68GAsWLKi1PSYmBra2to2ovG6xsbFGPZ+1u7s9JwVJsCZZil9PZ6IwJwOj/QyQMI802MO+Rk/mSnA2QwaVVEBQeQq2b08xUmWWi+974/CSSHASMvxwKAktb50Xu5xmxZiv0dLS+k250KgwUuXGjRuIjY1FbGwsZDIZhg8fjjNnziAsLAwffPABXnnllYc5vSjmzp1bozelsLAQPj4+GDx4MDQa49wBoNPpEBsbi0GDBpmsJ8ma3Ks9hwNom5CB1344iz9uSNEtPBgvPtJa3EItiDFeo1qdHv/6zwEAZXixbxCeGRBo3CItDN/3xuV7/Ra2fX4UKUVyDBjUH2qFTOySLJ4pXqNVVxcepMFhRKfT4ZdffsG6desQExOD9u3b429/+xvGjRtX/WW9ZcsWPP/880YNIy4uLpDJZMjKqrkmQVZWFjw8PO55jIeHR4P2BwCVSgWVSlVru0KhMPoHiCnOac3+3J5PdfNDfpke725LxNLYi3DT2OCZbj4iVmh5HuY1uubAVVzPL4OHRo2X+gdBoXiov32aDb7vjaOdtxOclALyyw04drUQ/e/MP0IPz5iv0fqep8EDWD09PTF9+nT4+fkhPj4ex44dw0svvVSj16B///5wcnJq6KnvS6lUokuXLti1a1f1NoPBgF27diEqKuqex0RFRdXYH6jsfqprf2p+pvUJwEt9K/8in/PjacSe5wJbTSGnSIvley4BAP4xNBi2SgYRMi6JRIKwFpWznu1O4iq+lq7BYeSjjz5CRkYGli9fjo4dO95zHycnJ1y+fPlha6tl9uzZWL16NTZs2IDExES8/PLLKCkpwZQpUwAAzz33XI0Brn/961+xY8cO/Pvf/0ZSUhLmz5+PY8eOYdasWUavjczXP4cG4+kurWAQgFlfn8CxtDyxS2r2Ptp5AcXaCrRv5YhRHb3FLoeaqXZ3hRFB4HSslqzBYWTixIlQq9WmqOWBxowZg6VLl+Ltt99Gx44dkZCQgB07dlQPUr169Spu3LhRvX/Pnj3x9ddfY9WqVejQoQO+//57/PTTTwgPDxelfhKHRCLB4iciMDDEDdoKA6ZuOIYLWZy60VSSMguxOf4qAODNEWG8k4lMpq1GgEouRXr+bVzI4urdlszipq+bNWsWrly5Aq1WiyNHjiAyMrL6d3FxcVi/fn2N/Z9++mkkJydDq9Xi7NmzGD58eBNXTOZALpPi03Gd0dnXCQW3dZi0Np7zE5iAIAh4b1siDAIwLNwD3Vs7i10SNWNKGdDjzmuMl2osm8WFEaLGslHK8MWkbghys8eNgjJMWhuP/NJysctqVuKSc7DvYi6UMinmDAsRuxyyAv2DXQAAexhGLBrDCFmVFnZKbHi+Ozw0alzMLsY0ThtvNGU6Peb9cg4AMLmXP/xactp3Mr1+wZVTwx+7ksc/LiwYwwhZHW8nG2x4vjs0ajmOXbmFWV+fQAWnjX9on+6+hKt5pfDQqPGXgZz2nZqGt5MN2rrbwyAAf1zIEbscaiSGEbJKwR4O+GJyN6jkUuxMzMbrW85wNP5DuJRdhM/3Vs6uOv+xMNireCsvNZ0BIZU3MXDciOViGCGr1c3fGZ+O6wypBPj22PXqlWWpYQRBwBtbzkKnFzAwxA1D2tU9qSCRKQy4M+HZHxdy2MtpoRhGyKoNCnPH4iciAAAr4lLwxX7jz4/T3P14Ih1HLudBrZBi/mPtIOEiQNTEOvs6wdFGgfxSHU5eyxe7HGoEhhGyemO6+eLvQ4IBAAu3nseWk9dFrshy5JeW473tiQCAvw5sCx9n4y4mSVQfcpm0eiArL9VYJoYRIgAz+gVick9/AMBr353GTk4bXy+Lticir6Qcbd3tMa0PFyIk8VRdqtmdyDBiiRhGiFA5S+vbj4bhiU7e0BsEzPj6BA6l3BS7LLO2OykL3x67DokEeG90BBQyfpyQePq2dYVUAiRnFeH6rfotW0/mg58eRHdIpRK8/1R7RIe6o7zCgOlfHsPp6/lil2WW8krK8Y/vzwAApvZqjW7+nGmVxOVkq0RXv8rXISdAszwMI0R3Ucik+HRcJ0QFtESxtgKT1sbjUjbXsblb5d0zZ5BbrEUbN3u8dme8DZHY+t+5VLOLYcTiMIwQ/YlaIcPqSV3RvpUjbpXqMG71EaTllohdltn4OSEDv53NhFwqwYfPdIRaIRO7JCIAwMDQyjByMOUmSssrRK6GGoJhhOge7FVyrJ/SHcHuDsgu0mLc6sO4lsfr0Bn5t/HWz2cBAH8Z2AYRrRxFrojof9q42aNVCxuUVxhw8BLHfFkShhGiOjjbKfHVtEgEutoho6AM49YctuqVfg0GAX///hSKyirQwccJM/oFil0SUQ0SiaT6rhpeqrEsDCNE9+HqoMLX03vAv6UtruXdxvg1R5BdWCZ2WaL4eNdFHLh0E2qFFB8+0wFy3j1DZqgqjOxJyuYSDxaEnyZED+CuUePr6T3QqoUNLueWYOzqw8iyskCyKzEL/9l1EQDw3qgIBLrai1wR0b31CGgJG4UMmYVlOH+jUOxyqJ4YRojqwcvJBv+d3gNejmqk5JTgmc8PId1KLtmk5Zbgb98kAAAm9vDDk11aiVsQ0X2oFTL0CnIBwAnQLAnDCFE9+Tjb4psXo+DjbIMrN0vxzMpDuHKzed9lU1pegZe+Oo6isgp09nXCW4+GiV0S0QNV3VXDcSOWg2GEqAF8nG3x7YtRCHCxQ3r+bTzz+SGk5BSLXZZJCALw5s/nkZRZBBd7JT4b3wVKOT8yyPxVjRs5dT0fOUVakauh+uAnC1EDeTraYPOLPdDW3R5ZhVqM+fwQzqYXiF2W0cWkS/Dr6UzIpBJ8Oq4zPBzVYpdEVC/uGjUivB0hCMCeZPaOWAKGEaJGcHNQY/MLUQjz1CC3uBzPfH4Icc3oQ29T/DVsv1Y5mdm8kWHoEdBS5IqIGqb6Uk0iF720BAwjRI3kbKfE5hd7oFdQS5SW6zF1wzF8e+ya2GU9tK2nM7BgayIAYGa/ADwX5S9uQUSNMDDEHQCw72IuynR6kauhB2EYIXoIGrUC6yZ3x+g7q/3+4/vT+GTnRYud32DvhRy88k0CBAHo5W7AXwdwYjOyTOHeGrhrVCgt1+NwKmdjNXcMI0QPSSmvnASsakbSj3ZewN++SbC4tTGOX8nDS18dh04vYHi4O55qbYBEIhG7LKJGqZyNtbJ3ZDfvqjF7DCNERiCRSPCPoSF4d1Q4ZFIJfk7IwBOfHbSYBfZ2JWZh/JojKC3Xo08bF/zryQhImUPIwkVXjxvhbKzmjmGEyIgm9PDD19Mi4WKvQlJmEUZ+uh87z5v3ALrN8Vcx/ctjKNMZ0C/YFZ9P5C281Dz0CnKBWiFFev5tJGUWiV0O3Qc/cYiMLDKgJbb9pTc6+zqhqKwC0748hsW/JUJbYV6D6ARBwCc7L2LOj2dgEICnurTC6ue6wlYpF7s0IqNQK2ToFVg5GyvvqjFvDCNEJuCuqbz1d1KUHwDg8z9S8eh/9uPUtXxxC7ujRFuBv39/Gh/tvAAAmNU/CP96qj0UXPyOmpmBoZXjRjgbq3njJw+RiSjlUix4PBwrJ3SBi70SF7OL8cSKg/hgR5KovSQnr97CiP/sw/fHr0MiARY+3g6vDQnmYFVqlqrmG0m4lo/cYs7Gaq4YRohMbGi4B2Je6YuRHbygNwj4LC4Fwz7Zh9/PZTbpoLoKvQEfxV7AUysPIe1mKTwd1dg0NRITOY8INWN3z8bKu2rMF8MIURNwtlNi2dhOWDmhM1zslUjNKcGLG4/jyRUHccTEcyAIgoD9F3PxxIqD+GTXRegNAh7r4IUdf30EPe+sbkrUnFWtVcNxI+aLI9WImtDQcE/0DHLBqj9S8cX+yzhxNR9jVh1Gv2BXTOrpj0fauEJmxHtqD6XcxEexFxCflgcAcFDL8e6ocDze0dtoj0Fk7qJD3fHJrovVs7GqFTKxS6I/YRghamIatQKvDQnGc1F++M/ui9gcfw1xyTmIS86Bp6MaT3VphWe6+sDH2bZR579ZrMWupGz8cPw6jlyuDCFKuRTjuvtiRr9AuGm44B1Zl6rZWLMKtTiUehP9g93ELon+hGGESCRuGjXeHRWBab0DsOFQGracTMeNgjIs230Jy3ZfQpCbPbr4tkAXvxbo7NcCfi1ta93tUqE34EZBGa7fuo1zGQWIOZ+FY2l5MNwZiqKUSfFsdx/M6BfEVXfJakkkEkSHumPTkauIPZ/FMGKGGEaIRObvYod5I9thzrAQxJzLwrfHrmH/pVxcyi7GpexifHPX4nsquRQOagUc1HKUVxiQWVgGvaH2INhwbw0GhXrgqa6t4O1k05RPh8gsDQqrDCM7z2fh3cfDIeUUw2aFYYTITKjkMozs4IWRHbyQV1KOk1dv4diVWzh+5RZOXcuHtsJQ+a9YW+MWRaVMilYtbODb0hZ927piUJg7WrVo3CUeouYqKrAl7FVyZBdpcTq9AB19nMQuie7CMEJkhpztlBgY6l49YZPeIKC4rAKFZToUaytQVFYBmRRo1cIWrvYq/pVH9AAquQx927pi25kbiD2fyTBiZnhrL5EFkEklcLRVwMfZFqGeGnRv7Ywufs5w16gZRIjqaVBYZbiPNfP1oqwRwwgREVmF/sFukEkluJBVjCs3LWNFbWvBMEJERFbB0VaByNbOANg7Ym4YRoiIyGpUXaqJYRgxKwwjRERkNaLvDAo/lpaHvJJykauhKgwjRERkNXycbRHi4QADF84zKwwjRERkVQZX31WTKXIlVIVhhIiIrMqgMA8AwN4LlQvnkfgsJozk5eVh/Pjx0Gg0cHJywtSpU1FcXHzf/f/v//4PwcHBsLGxga+vL/7yl7+goKCgCasmIiJzE+6tgaejGrd1ehy4lCt2OQQLCiPjx4/HuXPnEBsbi61bt2Lv3r144YUX6tw/IyMDGRkZWLp0Kc6ePYv169djx44dmDp1ahNWTURE5kYikVTfVfP7OV6qMQcWMR18YmIiduzYgaNHj6Jr164AgGXLlmH48OFYunQpvLy8ah0THh6OH374ofrnwMBAvPfee5gwYQIqKiogl1vEUyciIhMYGu6BLw9dQez5LFToDZDLLOZv82bJIr6RDx06BCcnp+ogAgDR0dGQSqU4cuQIRo8eXa/zFBQUQKPR3DeIaLVaaLX/W4SssLAQAKDT6aDT6Rr5DGqqOo+xzmft2J7GxzY1PrapcT1se3bydkALWwVulepw6FIOegQ4G7M8i2SK12h9z2URYSQzMxNubm41tsnlcjg7OyMzs35dbLm5uVi4cOF9L+0AwOLFi7FgwYJa22NiYmBra9yVUGNjY416PmvH9jQ+tqnxsU2N62Has62dFEdKpVj1WzzyWhuMWJVlM+ZrtLS0tF77iRpG5syZg/fff/+++yQmJj704xQWFmLEiBEICwvD/Pnz77vv3LlzMXv27BrH+vj4YPDgwdBoNA9dC1CZFGNjYzFo0CAoFAqjnNOasT2Nj21qfGxT4zJGe9ok5+DIVydxodQGQ4c+YvWLTpriNVp1deFBRA0jr776KiZPnnzffQICAuDh4YHs7JqT01RUVCAvLw8eHh73Pb6oqAhDhw6Fg4MDtmzZ8sAGVqlUUKlUtbYrFAqjf4CY4pzWjO1pfGxT42ObGtfDtOcjwe6wV8mRVajF+awSdPJtYeTqLJMxX6P1PY+oYcTV1RWurq4P3C8qKgr5+fk4fvw4unTpAgDYvXs3DAYDIiMj6zyusLAQQ4YMgUqlwi+//AK1Wm202omIyLKpFTL0D3HDr6cysONcJsOIiCxi+HBoaCiGDh2K6dOnIz4+HgcOHMCsWbPw7LPPVt9Jk56ejpCQEMTHxwOoDCKDBw9GSUkJvvjiCxQWFiIzMxOZmZnQ6znJDRERAUPbVfau7zibCUEQRK7GelnEAFYA2LRpE2bNmoWBAwdCKpXiySefxH/+85/q3+t0OiQnJ1cPljlx4gSOHDkCAAgKCqpxrsuXL8Pf37/JaiciIvPUL9gVSrkUV26WIimzCKGexhkbSA1jMWHE2dkZX3/9dZ2/9/f3r5Fq+/Xrx5RLRET3ZaeS45E2rtiZmIUdZzMZRkRiEZdpiIiITGVoeOWlGs7GKh6GESIismrRoW6QSSVIyizC5dwSscuxSgwjRERk1ZxslYgKaAmgciArNT2GESIisnpVl2p+O3tD5EqsE8MIERFZvaHhHpBKgNPXC3DlJi/VNDWGESIisnou9ipEBVZeqtl2hr0jTY1hhIiICMCj7Ssn0dx6imGkqTGMEBERoXI2VplUgvM3CpGaUyx2OVaFYYSIiAhACzslegW5AAC2nWbvSFNiGCEiIrrj0faeAICtDCNNimGEiIjojiFhHlDIJEjOKsLFrCKxy7EaDCNERER3ONoq0KeNKwD2jjQlhhEiIqK7VF2q2XbmBhdcbSIMI0RERHeJDnOHUibFpexiJPNSTZNgGCEiIrqLRq1A3+DKSzW8q6ZpMIwQERH9SdWlml9PZfBSTRNgGCEiIvqT6FB32ChkSLtZioRr+WKX0+wxjBAREf2JnUqOIe3cAQBbTqaLXE3zxzBCRER0D6M7twJQeammvMIgcjXNG8MIERHRPfQKbAlXBxVuleqw90KO2OU0awwjRERE9yCXSfFYh8qVfHmpxrQYRoiIiOowupM3ACA2MQsFt3UiV9N8MYwQERHVoZ2XBm3d7VFeYcBvZzjniKkwjBAREdVBIpFgdKfKgaw/8lKNyTCMEBER3cfjHb0gkQDxl/Nw/Vap2OU0SwwjRERE9+HlZIMerVsCAH5OyBC5muaJYYSIiOgBRneuHMj644nrnB7eBBhGiIiIHmBYuAdUcilScko4PbwJMIwQERE9gINageERlYvnfXP0msjVND8MI0RERPXwbDcfAMAvpzJQrK0QuZrmhWGEiIioHrq3dkaAqx1Ky/X4hQNZjYphhIiIqB4kEgnGdvMFAGw+elXkapoXhhEiIqJ6eqKzNxQyCU5fL8DZ9AKxy2k2GEaIiIjqqaW9CoPbeQBg74gxMYwQERE1wLjulZdqfj6ZgdJyDmQ1BoYRIiKiBogKaAlfZ1sUaSuw7TQXzzMGhhEiIqIGkEolGHPnNt/NnHPEKBhGiIiIGujpLq0gk0pw/MotJGcWiV2OxWMYISIiaiA3jRrRoW4AgI2H08QtphlgGCEiImqEST39AQA/HE9Hfmm5uMVYOIYRIiKiRogKaIlQTw1u6/T4Op63+T4MhhEiIqJGkEgkmNq7NQDgy4NXoNMbRK7IcjGMEBERNdLIDp5wsVchs7AM28/wNt/GYhghIiJqJJVchuei/AAAa/dfhiAIIldkmRhGiIiIHsL4SF8o5VKcul6A41duiV2ORbKYMJKXl4fx48dDo9HAyckJU6dORXFxcb2OFQQBw4YNg0QiwU8//WTaQomIyKq0tFfhiU7eAIAv9l8WuRrLZDFhZPz48Th37hxiY2OxdetW7N27Fy+88EK9jv34448hkUhMXCEREVmr5+8MZP39XCau5ZWKXI3lsYgwkpiYiB07dmDNmjWIjIxE7969sWzZMmzevBkZGRn3PTYhIQH//ve/sXbt2iaqloiIrE1bdwf0aeMCgwCsO5AmdjkWRy52AfVx6NAhODk5oWvXrtXboqOjIZVKceTIEYwePfqex5WWlmLcuHFYvnw5PDw86vVYWq0WWq22+ufCwkIAgE6ng06ne4hn8T9V5zHW+awd29P42KbGxzY1LnNszylRvth3MRdfx1/BtF6+cHVQiV1Sg5iiTet7LosII5mZmXBzc6uxTS6Xw9nZGZmZmXUe98orr6Bnz554/PHH6/1YixcvxoIFC2ptj4mJga2tbf2LrofY2Fijns/asT2Nj21qfGxT4zKn9hQEwN9ehrRiA+Zu3IMn/C1z3hFjtmlpaf0uWYkaRubMmYP333//vvskJiY26ty//PILdu/ejZMnTzbouLlz52L27NnVPxcWFsLHxweDBw+GRqNpVC1/ptPpEBsbi0GDBkGhUBjlnNaM7Wl8bFPjY5sal7m2pyY4F89vOIHDOXK8N6E33DVqsUuqN1O0adXVhQcRNYy8+uqrmDx58n33CQgIgIeHB7Kzs2tsr6ioQF5eXp2XX3bv3o2UlBQ4OTnV2P7kk0+iT58+iIuLu+dxKpUKKlXtrjWFQmH0F7wpzmnN2J7GxzY1PrapcZlbe/YP8UBXvxY4duUW1hy4ivmPtRO7pAYzZpvW9zyihhFXV1e4uro+cL+oqCjk5+fj+PHj6NKlC4DKsGEwGBAZGXnPY+bMmYNp06bV2BYREYGPPvoII0eOfPjiiYiI/kQikeCVQW0xfs0RfB1/FS/2DYCno43YZZk9i7ibJjQ0FEOHDsX06dMRHx+PAwcOYNasWXj22Wfh5eUFAEhPT0dISAji4+MBAB4eHggPD6/xDwB8fX3RunVr0Z4LERE1bz0DW6J7a2eUVxjw2Z4UscuxCBYRRgBg06ZNCAkJwcCBAzF8+HD07t0bq1atqv69TqdDcnJyvQfLEBERmYJEIsEr0W0BAN8cvYb0/NsiV/RgFXoD1h5IQ5lenMe3iLtpAMDZ2Rlff/11nb/39/d/4JoAXDOAiIiaQlRgS0QFtMSh1JtYvucSFo2OELuk+/rv0WtYvOMCPGxkGDWi6b8rLaZnhIiIyJK8Mqiyd+Tbo9dwIatI5GrqVnBbh49iLwAAerkbIJU2/YzlDCNEREQm0L21MwaHuaPCIODtn8+abe/8p7svIq+kHIGudujlLk6NDCNEREQm8tajYVArpDicmodfTt1/+RIxpOWWYP3BNADA3KFtIRMpFTCMEBERmYiPsy1m9Q8CALy3LRFFZeYzfT0ALNqeCJ1eQN+2rujb9sFTbZgKwwgREZEJTX8kAK1d7JBdpMXHOy+KXU61gym5iDmfBZlUgjdHhIpaC8MIERGRCanksuqZWNcfTENSZv2mSDclvUHAwq2Vy62Mj/RFG3cHUethGCEiIjKxvm1dMSzcA3qDgLd/Oif6YNZvj11D4o1COKjl+NudOVHExDBCRETUBN58NAw2Chni0/Kw9kCaaHVcuVmC97ZV9or8dWAbONspRaulCsMIERFRE/B2ssHc4SEAgMXbE3Hi6q0mr6G8woC//PckirUV6O7vjMk9/Zu8hnthGCEiImoiE3v4YUSEJyoMAv7v65PILy1v0sf/MPYCTl0vgEYtx0fPdoRcrHt5/8Q8qiAiIrICEokES56MgH9LW6Tn38ar356CwdA040f2X8zFyj8qF+57/8n28HYyn9WEGUaIiIiakINagc/Gd4FSLsWupGys2pdq8se8WazFK98mAADGRfpiWISnyR+zIRhGiIiImliYlwbv3Lnd91+/J2P/xVyTPZa2Qo+/fZOAnCIt2rjZ460RYSZ7rMZiGCEiIhLBmG4+eKKTN/QGAVM3HEVccrbRH0NbocfLX53Avou5UMmlWDauE2yUMqM/zsNiGCEiIhKBRCLBoiciEB3qBm2FAdO/PIbfz2Ua7fxlOj1e3Hgcu5OyoVZI8cWkbgjx0Bjt/MbEMEJERCQStUKGz8Z3wYgIT+j0AmZsOmGUBfXKdHpM//IY4pJzoFZIsXZSN/Ru42KEik2DYYSIiEhESrkUnzzbsfqSzV83n8Safamo0Bsadb6swjJM3XAU+y7mwlYpw/op3dEzyHyDCMAwQkREJDq5TIqlT3fAuEhfCALw7rZEPLpsPw6n3qz3OcorDFj5RwoGLI3DgUs3YXcniPQIaGnCyo1DLnYBREREBEilErw3KhyhHg5YGnMBSZlFeHbVYYzs4IXXBreFr7MtJBJJreMMBgF/XMzBwl/PIzW3BADQydcJ742KQJiXeY4R+TOGESIiIjMhkUgwMcofI9p74d8xyfg6/ip+PZWBX09lwMVehfatHBHu7QgPjRrJmYU4m1GIxBuFKC3XAwBc7FWYMywET3TyhlRaO7iYK4YRIiIiM+Nsp8R7oyMwtrsvFm1PxOHUm8gt1mJ3UjZ2J9W+BdhGIcO4SF/8NboNNGqFCBU/HIYRIiIiMxXu7Yivp/fA7XI9zt8oxJnr+TiTXoicYi3autkj3NsR7bw0CHC1h8yCekL+jGGEiIjIzNkoZeji1wJd/FqIXYpJ8G4aIiIiEhXDCBEREYmKYYSIiIhExTBCREREomIYISIiIlExjBAREZGoGEaIiIhIVAwjREREJCqGESIiIhIVwwgRERGJimGEiIiIRMW1aR5AEAQAQGFhodHOqdPpUFpaisLCQigUlre6orlhexof29T42KbGxfY0PlO0adV3Z9V3aV0YRh6gqKgIAODj4yNyJURERJapqKgIjo6Odf5eIjworlg5g8GAjIwMODg4QCIxzvLMhYWF8PHxwbVr16DRaIxyTmvG9jQ+tqnxsU2Ni+1pfKZoU0EQUFRUBC8vL0ildY8MYc/IA0ilUrRq1cok59ZoNHwTGRHb0/jYpsbHNjUutqfxGbtN79cjUoUDWImIiEhUDCNEREQkKoYREahUKsybNw8qlUrsUpoFtqfxsU2Nj21qXGxP4xOzTTmAlYiIiETFnhEiIiISFcMIERERiYphhIiIiETFMEJERESiYhgxkeXLl8Pf3x9qtRqRkZGIj4+/7/7fffcdQkJCoFarERERge3btzdRpZahIe25evVq9OnTBy1atECLFi0QHR39wPa3Rg19jVbZvHkzJBIJRo0aZdoCLUxD2zM/Px8zZ86Ep6cnVCoV2rZty/f9nzS0TT/++GMEBwfDxsYGPj4+eOWVV1BWVtZE1Zq3vXv3YuTIkfDy8oJEIsFPP/30wGPi4uLQuXNnqFQqBAUFYf369aYrUCCj27x5s6BUKoW1a9cK586dE6ZPny44OTkJWVlZ99z/wIEDgkwmEz744APh/PnzwptvvikoFArhzJkzTVy5eWpoe44bN05Yvny5cPLkSSExMVGYPHmy4OjoKFy/fr2JKzdfDW3TKpcvXxa8vb2FPn36CI8//njTFGsBGtqeWq1W6Nq1qzB8+HBh//79wuXLl4W4uDghISGhiSs3Xw1t002bNgkqlUrYtGmTcPnyZeH3338XPD09hVdeeaWJKzdP27dvF9544w3hxx9/FAAIW7Zsue/+qampgq2trTB79mzh/PnzwrJlywSZTCbs2LHDJPUxjJhA9+7dhZkzZ1b/rNfrBS8vL2Hx4sX33P+ZZ54RRowYUWNbZGSk8OKLL5q0TkvR0Pb8s4qKCsHBwUHYsGGDqUq0OI1p04qKCqFnz57CmjVrhEmTJjGM3KWh7blixQohICBAKC8vb6oSLU5D23TmzJnCgAEDamybPXu20KtXL5PWaYnqE0b+8Y9/CO3atauxbcyYMcKQIUNMUhMv0xhZeXk5jh8/jujo6OptUqkU0dHROHTo0D2POXToUI39AWDIkCF17m9NGtOef1ZaWgqdTgdnZ2dTlWlRGtum77zzDtzc3DB16tSmKNNiNKY9f/nlF0RFRWHmzJlwd3dHeHg4Fi1aBL1e31Rlm7XGtGnPnj1x/Pjx6ks5qamp2L59O4YPH94kNTc3Tf29xIXyjCw3Nxd6vR7u7u41tru7uyMpKemex2RmZt5z/8zMTJPVaSka055/9s9//hNeXl613ljWqjFtun//fnzxxRdISEhoggotS2PaMzU1Fbt378b48eOxfft2XLp0CTNmzIBOp8O8efOaomyz1pg2HTduHHJzc9G7d28IgoCKigq89NJLeP3115ui5Ganru+lwsJC3L59GzY2NkZ9PPaMULO2ZMkSbN68GVu2bIFarRa7HItUVFSEiRMnYvXq1XBxcRG7nGbBYDDAzc0Nq1atQpcuXTBmzBi88cYbWLlypdilWay4uDgsWrQIn332GU6cOIEff/wR27Ztw8KFC8UujeqBPSNG5uLiAplMhqysrBrbs7Ky4OHhcc9jPDw8GrS/NWlMe1ZZunQplixZgp07d6J9+/amLNOiNLRNU1JSkJaWhpEjR1ZvMxgMAAC5XI7k5GQEBgaatmgz1pjXqKenJxQKBWQyWfW20NBQZGZmory8HEql0qQ1m7vGtOlbb72FiRMnYtq0aQCAiIgIlJSU4IUXXsAbb7wBqZR/ezdEXd9LGo3G6L0iAHtGjE6pVKJLly7YtWtX9TaDwYBdu3YhKirqnsdERUXV2B8AYmNj69zfmjSmPQHggw8+wMKFC7Fjxw507dq1KUq1GA1t05CQEJw5cwYJCQnV/x577DH0798fCQkJ8PHxacryzU5jXqO9evXCpUuXqkMdAFy4cAGenp5WH0SAxrVpaWlprcBRFfYELsHWYE3+vWSSYbFWbvPmzYJKpRLWr18vnD9/XnjhhRcEJycnITMzUxAEQZg4caIwZ86c6v0PHDggyOVyYenSpUJiYqIwb9483tp7l4a255IlSwSlUil8//33wo0bN6r/FRUVifUUzE5D2/TPeDdNTQ1tz6tXrwoODg7CrFmzhOTkZGHr1q2Cm5ub8O6774r1FMxOQ9t03rx5goODg/Df//5XSE1NFWJiYoTAwEDhmWeeEespmJWioiLh5MmTwsmTJwUAwocffiicPHlSuHLliiAIgjBnzhxh4sSJ1ftX3dr797//XUhMTBSWL1/OW3st0bJlywRfX19BqVQK3bt3Fw4fPlz9u759+wqTJk2qsf+3334rtG3bVlAqlUK7du2Ebdu2NXHF5q0h7enn5ycAqPVv3rx5TV+4GWvoa/RuDCO1NbQ9Dx48KERGRgoqlUoICAgQ3nvvPaGioqKJqzZvDWlTnU4nzJ8/XwgMDBTUarXg4+MjzJgxQ7h161bTF26G9uzZc8/Pxao2nDRpktC3b99ax3Ts2FFQKpVCQECAsG7dOpPVJxEE9l8RERGReDhmhIiIiETFMEJERESiYhghIiIiUTGMEBERkagYRoiIiEhUDCNEREQkKoYRIiIiEhXDCBEREYmKYYSIiIhExTBCREREomIYISIiIlExjBCRxcnJyYGHhwcWLVpUve3gwYNQKpW1lj0nIvPHhfKIyCJt374do0aNwsGDBxEcHIyOHTvi8ccfx4cffih2aUTUQAwjRGSxZs6ciZ07d6Jr1644c+YMjh49CpVKJXZZRNRADCNEZLFu376N8PBwXLt2DcePH0dERITYJRFRI3DMCBFZrJSUFGRkZMBgMCAtLU3scoiokdgzQkQWqby8HN27d0fHjh0RHByMjz/+GGfOnIGbm5vYpRFRAzGMEJFF+vvf/47vv/8ep06dgr29Pfr27QtHR0ds3bpV7NKIqIF4mYaILE5cXBw+/vhjbNy4ERqNBlKpFBs3bsS+ffuwYsUKscsjogZizwgRERGJij0jREREJCqGESIiIhIVwwgRERGJimGEiIiIRMUwQkRERKJiGCEiIiJRMYwQERGRqBhGiIiISFQMI0RERCQqhhEiIiISFcMIERERier/AXUpHfqTNz2BAAAAAElFTkSuQmCC", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "plt.plot(x_norm.detach().numpy(), y_norm.detach().numpy())\n", "plt.xlabel('x')\n", @@ -734,21 +647,10 @@ }, { "cell_type": "code", - "execution_count": 88, + "execution_count": null, "id": "cc7f9018", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "torch.Size([100, 1])" - ] - }, - "execution_count": 88, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "x_norm.shape" ] @@ -786,22 +688,10 @@ }, { "cell_type": "code", - "execution_count": 89, + "execution_count": null, "id": "4bb6bff9", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Sequential(\n", - " (0): Linear(in_features=1, out_features=5, bias=True)\n", - " (1): LeakyReLU(negative_slope=0.01)\n", - " (2): Linear(in_features=5, out_features=1, bias=True)\n", - ")\n" - ] - } - ], + "outputs": [], "source": [ "n_hidden_01 = 5\n", "\n", @@ -828,22 +718,10 @@ }, { "cell_type": "code", - "execution_count": 90, + "execution_count": null, "id": "56967956", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Sequential(\n", - " (0): Linear(in_features=1, out_features=5, bias=True)\n", - " (1): Tanh()\n", - " (2): Linear(in_features=5, out_features=1, bias=True)\n", - ")\n" - ] - } - ], + "outputs": [], "source": [ "n_hidden_11 = 5\n", "\n", @@ -870,24 +748,10 @@ }, { "cell_type": "code", - "execution_count": 91, + "execution_count": null, "id": "64f4f426", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Sequential(\n", - " (0): Linear(in_features=1, out_features=10, bias=True)\n", - " (1): Tanh()\n", - " (2): Linear(in_features=10, out_features=5, bias=True)\n", - " (3): Tanh()\n", - " (4): Linear(in_features=5, out_features=1, bias=True)\n", - ")\n" - ] - } - ], + "outputs": [], "source": [ "n_hidden_21 = 10\n", "n_hidden_22 = 5\n", @@ -923,150 +787,12 @@ }, { "cell_type": "code", - "execution_count": 92, + "execution_count": null, "id": "02e4fbbb", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "['T_destination',\n", - " '__add__',\n", - " '__annotations__',\n", - " '__call__',\n", - " '__class__',\n", - " '__delattr__',\n", - " '__delitem__',\n", - " '__dict__',\n", - " '__dir__',\n", - " '__doc__',\n", - " '__eq__',\n", - " '__format__',\n", - " '__ge__',\n", - " '__getattr__',\n", - " '__getattribute__',\n", - " '__getitem__',\n", - " '__getstate__',\n", - " '__gt__',\n", - " '__hash__',\n", - " '__iadd__',\n", - " '__imul__',\n", - " '__init__',\n", - " '__init_subclass__',\n", - " '__iter__',\n", - " '__le__',\n", - " '__len__',\n", - " '__lt__',\n", - " '__module__',\n", - " '__mul__',\n", - " '__ne__',\n", - " '__new__',\n", - " '__reduce__',\n", - " '__reduce_ex__',\n", - " '__repr__',\n", - " '__rmul__',\n", - " '__setattr__',\n", - " '__setitem__',\n", - " '__setstate__',\n", - " '__sizeof__',\n", - " '__str__',\n", - " '__subclasshook__',\n", - " '__weakref__',\n", - " '_apply',\n", - " '_backward_hooks',\n", - " '_backward_pre_hooks',\n", - " '_buffers',\n", - " '_call_impl',\n", - " '_compiled_call_impl',\n", - " '_forward_hooks',\n", - " '_forward_hooks_always_called',\n", - " '_forward_hooks_with_kwargs',\n", - " '_forward_pre_hooks',\n", - " '_forward_pre_hooks_with_kwargs',\n", - " '_get_backward_hooks',\n", - " '_get_backward_pre_hooks',\n", - " '_get_item_by_idx',\n", - " '_get_name',\n", - " '_is_full_backward_hook',\n", - " '_load_from_state_dict',\n", - " '_load_state_dict_post_hooks',\n", - " '_load_state_dict_pre_hooks',\n", - " '_maybe_warn_non_full_backward_hook',\n", - " '_modules',\n", - " '_named_members',\n", - " '_non_persistent_buffers_set',\n", - " '_parameters',\n", - " '_register_load_state_dict_pre_hook',\n", - " '_register_state_dict_hook',\n", - " '_replicate_for_data_parallel',\n", - " '_save_to_state_dict',\n", - " '_slow_forward',\n", - " '_state_dict_hooks',\n", - " '_state_dict_pre_hooks',\n", - " '_version',\n", - " '_wrapped_call_impl',\n", - " 'add_module',\n", - " 'append',\n", - " 'apply',\n", - " 'bfloat16',\n", - " 'buffers',\n", - " 'call_super_init',\n", - " 'children',\n", - " 'compile',\n", - " 'cpu',\n", - " 'cuda',\n", - " 'double',\n", - " 'dump_patches',\n", - " 'eval',\n", - " 'extend',\n", - " 'extra_repr',\n", - " 'float',\n", - " 'forward',\n", - " 'get_buffer',\n", - " 'get_extra_state',\n", - " 'get_parameter',\n", - " 'get_submodule',\n", - " 'half',\n", - " 'insert',\n", - " 'ipu',\n", - " 'load_state_dict',\n", - " 'modules',\n", - " 'named_buffers',\n", - " 'named_children',\n", - " 'named_modules',\n", - " 'named_parameters',\n", - " 'parameters',\n", - " 'pop',\n", - " 'register_backward_hook',\n", - " 'register_buffer',\n", - " 'register_forward_hook',\n", - " 'register_forward_pre_hook',\n", - " 'register_full_backward_hook',\n", - " 'register_full_backward_pre_hook',\n", - " 'register_load_state_dict_post_hook',\n", - " 'register_module',\n", - " 'register_parameter',\n", - " 'register_state_dict_pre_hook',\n", - " 'requires_grad_',\n", - " 'set_extra_state',\n", - " 'share_memory',\n", - " 'state_dict',\n", - " 'to',\n", - " 'to_empty',\n", - " 'train',\n", - " 'training',\n", - " 'type',\n", - " 'xpu',\n", - " 'zero_grad']" - ] - }, - "execution_count": 92, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ - "dir(model0)" + "# dir(model0)" ] }, { @@ -1084,21 +810,10 @@ }, { "cell_type": "code", - "execution_count": 93, + "execution_count": null, "id": "59410e54", "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "" - ] - }, - "execution_count": 93, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "model0.parameters()" ] @@ -1113,29 +828,10 @@ }, { "cell_type": "code", - "execution_count": 94, + "execution_count": null, "id": "95e9f71a", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Parameter containing:\n", - "tensor([[ 0.6627],\n", - " [ 0.1130],\n", - " [ 0.9284],\n", - " [-0.0028],\n", - " [ 0.3684]], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([-0.6180, -0.9190, -0.1378, 0.7278, 0.5515], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([[-0.3507, 0.3441, 0.4413, -0.2675, -0.4247]], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([0.0901], requires_grad=True)\n" - ] - } - ], + "outputs": [], "source": [ "for element in model0.parameters():\n", " print(element)" @@ -1171,29 +867,10 @@ }, { "cell_type": "code", - "execution_count": 95, + "execution_count": null, "id": "1ef5f2e7", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Parameter containing:\n", - "tensor([[ 0.6627],\n", - " [ 0.1130],\n", - " [ 0.9284],\n", - " [-0.0028],\n", - " [ 0.3684]], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([-0.6180, -0.9190, -0.1378, 0.7278, 0.5515], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([[-0.3507, 0.3441, 0.4413, -0.2675, -0.4247]], requires_grad=True)\n", - "Parameter containing:\n", - "tensor([0.0901], requires_grad=True)\n" - ] - } - ], + "outputs": [], "source": [ "for element in model0.parameters():\n", " print(element)" @@ -1228,7 +905,7 @@ }, { "cell_type": "code", - "execution_count": 96, + "execution_count": null, "id": "fc2ceaf9", "metadata": {}, "outputs": [], @@ -1251,18 +928,17 @@ }, { "cell_type": "code", - "execution_count": 97, + "execution_count": null, "id": "85354d61", "metadata": {}, "outputs": [], "source": [ - "batch_size = 64 # how many points to pass to the model at a time\n", "learning_rate = 1e-2" ] }, { "cell_type": "code", - "execution_count": 98, + "execution_count": null, "id": "90b75147", "metadata": {}, "outputs": [], @@ -1274,7 +950,7 @@ }, { "cell_type": "code", - "execution_count": 99, + "execution_count": null, "id": "f92a6ed6", "metadata": {}, "outputs": [], @@ -1302,19 +978,20 @@ }, { "cell_type": "code", - "execution_count": 100, + "execution_count": null, "id": "43bdff9a", "metadata": {}, "outputs": [], "source": [ - "batch_size = 64\n", + "batch_size = 64 # how many points to pass to the model at a time\n", + "# batch_size = len(x_norm) # uncomment to pass all data at once\n", "dataset = TensorDataset(x_norm, y_norm)\n", "dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True, pin_memory=False, drop_last=True)" ] }, { "cell_type": "code", - "execution_count": 105, + "execution_count": null, "id": "84e1212f", "metadata": {}, "outputs": [], @@ -1322,7 +999,6 @@ "# Define the training loop NEW\n", "def training_loop(dataloader, model, optimizer, epochs):\n", " losses = []\n", - " print(epochs)\n", " for _ in range(epochs):\n", " for id_batch, (x_batch, y_batch) in enumerate(dataloader):\n", " x_batch = x_batch.to(device)\n", @@ -1332,63 +1008,40 @@ " loss = loss_function(pred_y, y_batch)\n", " loss.backward() # Back-prop\n", " optimizer.step()\n", - " # Every 50 batches, print the loss for this batch\n", - " # as well as the number of examples processed so far \n", - " if id_batch % 50 == 0:\n", - " loss, current = loss.item(), (id_batch + 1)* len(x_batch)\n", - " losses.append(loss)\n", + " losses.append(loss.item())\n", " return losses" ] }, { "cell_type": "code", - "execution_count": 106, + "execution_count": null, "id": "8455e259", "metadata": { "slideshow": { "slide_type": "slide" } }, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "10000\n", - "10000\n", - "10000\n" - ] - } - ], + "outputs": [], "source": [ "# Run the training for all the models\n", + "# epochs = 2000\n", + "epochs = 500\n", "\n", - "losses0 = training_loop(dataloader, model0, optimizer0, epochs=10000)\n", - "losses1 = training_loop(dataloader, model1, optimizer1, epochs=10000)\n", - "losses2 = training_loop(dataloader, model2, optimizer2, epochs=10000)" + "losses0 = training_loop(dataloader, model0, optimizer0, epochs=epochs)\n", + "losses1 = training_loop(dataloader, model1, optimizer1, epochs=epochs)\n", + "losses2 = training_loop(dataloader, model2, optimizer2, epochs=epochs)" ] }, { "cell_type": "code", - "execution_count": 108, + "execution_count": null, "id": "a14805b5", "metadata": { "slideshow": { "slide_type": "slide" } }, - "outputs": [ - { - "data": { - "image/png": "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", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "plt.plot(losses0, label='Model 0', color='green')\n", "plt.plot(losses1, label='Model 1', color='blue')\n", @@ -1400,6 +1053,23 @@ "plt.show()" ] }, + { + "cell_type": "markdown", + "id": "c0bac345", + "metadata": { + "slideshow": { + "slide_type": "slide" + } + }, + "source": [ + "

Interpreting the loss curves

\n", + "

$\\implies$ Have the NNs learned?

\n", + "

$\\implies$ Why is model 0 learning faster than model 1?

\n", + "

$\\implies$ Why is model 2 better than models 0 and 1?

\n", + "

$\\implies$ Train for more epochs. How does the loss curve change?

\n", + "

$\\implies$ Change the number of minibatches to pass all data at once. How does the loss curve change? which method is more effective

" + ] + }, { "cell_type": "markdown", "id": "846d5b90", @@ -1416,7 +1086,7 @@ }, { "cell_type": "code", - "execution_count": 109, + "execution_count": null, "id": "436395cb", "metadata": {}, "outputs": [], @@ -1438,7 +1108,7 @@ }, { "cell_type": "code", - "execution_count": 110, + "execution_count": null, "id": "38a68ec4", "metadata": {}, "outputs": [], @@ -1450,29 +1120,18 @@ }, { "cell_type": "code", - "execution_count": 111, + "execution_count": null, "id": "114f53af", "metadata": { "slideshow": { "slide_type": "slide" } }, - "outputs": [ - { - "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhAAAAFfCAYAAAAf5vWtAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB3t0lEQVR4nO3deVxU5f7A8c+ZYVcBN5BlEJfcctc0FxLSEjPTEM0l0zJt0dQsTetXttyrllZYWbaZdq87orZ4MTcMtzQVM8UdFBBwQVlEtpnz+2NgYAQUjGH9vu9rLs05z5l55qic75zneb5fRVVVFSGEEEKIUtBUdAeEEEIIUfVIACGEEEKIUpMAQgghhBClJgGEEEIIIUpNAgghhBBClJoEEEIIIYQoNQkghBBCCFFqVhXdgbJmMBi4dOkSderUQVGUiu6OEEIIUWWoqkpqairu7u5oNHe+x1DtAohLly6h0+kquhtCCCFElRUTE4Onp+cd21S7AKJOnTqA8cM7OjpWcG+EEEKIqiMlJQWdTme6lt5JtQsg8oYtHB0dJYAQQggh7kFJpgDIJEohhBBClJoEEEIIIYQotXIJIBYvXoy3tzd2dnZ0796dAwcO3LH9jRs3mDRpEm5ubtja2tKiRQs2b95cHl0VQgghRAlYfA7EmjVrmD59OkuWLKF79+4EBQXRv39/Tp06hYuLS6H2WVlZPPLII7i4uBAcHIyHhwcXLlzA2dnZ0l0VQghRhvR6PdnZ2RXdDXEbGxubuy7RLAlFVVW1DPpTrO7du/PAAw/wxRdfAMY8DTqdjldeeYVZs2YVar9kyRIWLFjAyZMnsba2vuvrZ2ZmkpmZaXqeN4M0OTlZJlEKIUQFUFWVhIQEbty4UdFdEUXQaDQ0adIEGxubQvtSUlJwcnIq0TXUoncgsrKyOHToELNnzzZt02g09OvXj3379hV5zE8//USPHj2YNGkSmzZtomHDhowaNYo33ngDrVZbqP28efN47733LPYZhBBClE5e8ODi4oKDg4Mk9atE8pItxsfH4+Xl9Y/+bCwaQFy9ehW9Xo+rq6vZdldXV06ePFnkMefPn2fHjh2MHj2azZs3c/bsWV5++WWys7OZM2dOofazZ89m+vTppud5dyCEEKKi6Q16wi+GE58aj1sdN3y8fNBqCn8Rqk70er0peKhfv35Fd0cUoWHDhly6dImcnJwS3ekvTqXLA2EwGHBxceGbb75Bq9XSpUsX4uLiWLBgQZEBhK2tLba2thXQUyGEKF5IZAhTQ6cSmxJr2ubp6Mki/0UEtA6owJ5ZVt6cBwcHhwruiShO3tCFXq//RwGERVdhNGjQAK1WS2Jiotn2xMREGjVqVOQxbm5utGjRwmy4onXr1iQkJJCVlWXJ7gohRJkIiQwhcG2gWfAAEJcSR+DaQEIiQyqoZ+VHhi0qr7L6s7FoAGFjY0OXLl3Yvn27aZvBYGD79u306NGjyGN69erF2bNnMRgMpm2nT5/Gzc2tyAkfQghRmegNeqaGTkWl8Px0Nfd/00KnoTfoK6B3QpQdi+eBmD59Ot9++y3Lly8nMjKSl156iZs3b/Lss88C8Mwzz5hNsnzppZdISkpi6tSpnD59ml9//ZW5c+cyadIkS3dVCCH+sfCL4YXuPNwuJiWG8Ivh5dQjISzD4gHEU089xcKFC3nnnXfo2LEjERERhIaGmiZWXrx4kfj4eFN7nU7Hli1bOHjwIO3bt2fKlClMnTq1yCWfQghR2cSnxt+9EbDp5CYL90RURmFhYSiKUqolrt7e3gQFBVmsT/eqXDJRTp48mQsXLpCZmckff/xB9+7dTfvCwsJYtmyZWfsePXqwf/9+MjIyOHfuHG+++WaRSziFEKKycavjVqJ2K46tkGGMu9Ab9IRFh7Hq2CrCosMsfr7GjRuHoii8+OKLhfZNmjQJRVEYN26cRftwr9atW0erVq2ws7OjXbt25ZK9WWphCCFEGfLx8qGBQ4O7truSfkWGMe4gJDIE70Xe+C33Y1TIKPyW++G9yNviE1B1Oh2rV6/m1q1bpm0ZGRmsXLkSLy8vi773vdq7dy8jR45k/PjxHDlyhCFDhjBkyBD+/vtvi76vBBBCCFGGtBotT7d/ukRtSzrcUdNU5CqWzp07o9PpCAnJf4+QkBC8vLzo1KmTWdvMzEymTJmCi4sLdnZ29O7dm4MHD5q12bx5My1atMDe3h4/Pz+io6MLvefu3bvx8fHB3t4enU7HlClTuHnzZon7vGjRIvz9/ZkxYwatW7fmgw8+oHPnzqYM0JYiAYQQQpSxwS0Hl6hdSYc7apK7rWIBLL6K5bnnnuOHH34wPV+6dKlp4n9BM2fOZP369SxfvpzDhw/TvHlz+vfvT1JSEgAxMTEEBAQwaNAgIiIieP755wvN5zt37hz+/v4MHTqUv/76izVr1rB7924mT55c4v7u27ePfv36mW3r379/sRmfy4oEEEIIUcZ8vHzwdPREofj19vXt66M36GUexG3utopFRbX4Kpann36a3bt3c+HCBS5cuMCePXt4+mnzu0o3b97kq6++YsGCBQwYMIA2bdrw7bffYm9vz/fffw/AV199RbNmzfj4449p2bIlo0ePLjSHYt68eYwePZpp06Zx33330bNnTz777DN+/PFHMjIyStTfhISEIjM+JyQk3PtJKAEJIIQQooxpNVoW+S8CKDaIuHbrGv3+069cxvWrkpIO61hy+Kdhw4YMHDiQZcuW8cMPPzBw4EAaNDCf13Lu3Dmys7Pp1auXaZu1tTXdunUjMjISgMjISLNFA0ChHEhHjx5l2bJl1K5d2/To378/BoOBqKgoC33CslHpUlkLIUR1ENA6gODhwYXSWd8ub1w/eHhwtU5xXVIlHdax9PDPc889ZxpGWLx4scXeJy0tjRdeeIEpU6YU2lfSSZuNGjUqVcbnsiJ3IIQQwkICWgcQPTWabWO2Uc++XpFtymtcv6q42/CPgoLOUYePl49F++Hv709WVhbZ2dn079+/0P5mzZphY2PDnj17TNuys7M5ePAgbdq0AYxlGA4cOGB23P79+82ed+7cmRMnTtC8efNCj5JmX+7Ro4dZxmeArVu3FpvxuaxIACGEEKVUmvwEWo0WrUZL0q2kYtuUx7h+VXGn4Z+850H+QRavaqrVaomMjOTEiRNF5iGqVasWL730EjNmzCA0NJQTJ04wYcIE0tPTGT9+PAAvvvgiZ86cYcaMGZw6dYqVK1cWynv0xhtvsHfvXiZPnkxERARnzpxh06ZNpZpEOXXqVEJDQ/n44485efIk7777Ln/++WepXuNeSAAhhBClcC/5CSrDuH5Vkjf84+HoYbbd09GzXId6HB0dcXR0LHb//PnzGTp0KGPGjKFz586cPXuWLVu2ULduXcA4BLF+/Xo2btxIhw4dWLJkCXPnzjV7jfbt27Nr1y5Onz6Nj48PnTp14p133sHd3b3E/ezZsycrV67km2++oUOHDgQHB7Nx40batm17bx+8hBRVVQuvlanCUlJScHJyIjk5+Y5/8EIIUVp5+QluX2KY9824uItbWHQYfsv97vr6O8fuxNfbt0z6WlEyMjKIioqiSZMm2NnZ/aPX0hv0hF8MJz41Hrc6bvh4+Vj8zkNNcKc/o9JcQ2USpRBClMDd8hMoKEwLncbgloMLXeTyxvXjUuKKPF5BwdPR0+Lj+lWNVqOt8gFVdSZDGEIIUQJm+QkMGojqA8dGGH8aNEXOY9Bn6Yn4ZDuRQ99lxU4/fM+C1mD+uuU5ri9EWZI7EEIIUQKm+QknnoTQRZCiy9/pGAP+U6HNBlO7/TNDaPHxRDoarpma7QCSbGz5tGcmZ+tDfG2IbufBJ48tkiWcosqRAEIIIUrArY6bMXhYG4wGPT6E4UY88bgRntILw9pgGB6I21g39s8ModuCoUUuRKyblckHYfnPVU9QmgCty+dzCFFWJIAQQogS6Onhg3ZLC54ghEW8io785FAxeDKVT/npt8/p/G19Dqzox8qOA8jWWqGioCqgotDk+iW6xp7AKbNAoaS4OAgMhOBgCJC7EKLqkABCCCFuo9dDeDjEx4ObG/j4wN49Wp5I3k8ww+G2iZCZ9cCvxTrOeNnS9d8assbMLva1FdVAq8vRdI/5G7/zf/JQ1GHjFMxp02DwYCgi54AQlZEEEEIIUUBICEydCrEFsk97esKwAD2LmAqoptnnf7s248sHh/G/lj1Rlfw56U63Uul86SS1sm6hqCoaVUWv0XDCpSnn63sS6dqUSNemLOv6BPcnnGXK3tU8cuYPNOHh4Otbnh9XiHsmAYQQQuQKCTGOJtyeHScuDg5/Fm4atjjk0YrPeo5gV9Oupja+5/7E9/yfaO8bwOg1k9EUsVwT4HItZw563s/exh3Y2MaX442a80LA/9HqchSTT1xhYB8VRSm+iqcQlYUs4xRCCIzDFlOnFg4ewLjNnXgyrGx4/+HnGfr0QnY17YrGoGfI8Z1s+X4Sy4LfZdzhXxg5wIl4xb2Y8AFcbt5g4Kk9/Pu3L9m9ZDyv7F1NncybnHRpwuSLDry84jDJt7It+llFxQkLC0NRFG7cuFHiY7y9vQkKCrJYn+6VBBBCCIFxzkPesIUGPX0IYwSr6EMYGvREubgwaOynLH1gCACBx7YS9s1Egn75mJZXL5hex6qxBxemf4bK7TMlCqubkcpr4f/l96/GM3bPr2gVhf/9ncDAz8KJiLlhiY8p7mDcuHEoisKLL75YaN+kSZNQFIVx48aVf8fu4vjx4wwdOhRvb28URSm3YEMCCCGEwDhhEuBJQojGmzD8WMUodvAw/+r+AgnP3OJMg8Y0SLvOD+veZeHmRXgl55dQVlFApwMfH3ouDOC/g9dzjfqF3uf2oMKAglPmTY7t7sf/PdgTr3oOxF6/ReBXe/ku/DxqTg6EhcGqVcaf+ppTsVOvL/+PrtPpWL16Nbdu3TJty8jIYOXKlSUur13e0tPTadq0KfPnz7d4Ce+CJIAQQgiMqy2eJIRgAvHIneuQqbViyhMz+Nr3SVStQpPTVwldOhnfqENmx6qKgqIAQUGmVRTPbAwgbHUiAY7beJ//433+jznMIRbzAlGxeDKMYP7UBfDM4878MqU3j7VrRI5B5V+/RvLaqHfRP9wXRo0CPz/w9jZO1qjmQkKMH9XPr3w/eufOndHpdIQUeKOQkBC8vLzo1KmTWdvMzEymTJmCi4sLdnZ29O7dm4MHD5q12bx5My1atMDe3h4/Pz+io6MLvefu3bvx8fHB3t4enU7HlClTuHnzZqF2xXnggQdYsGABI0aMwNbWtnQf+B+QAEIIIYCeD2axSJlM3iqLVBt7ngt8l19aP4S1Ppt5//ucZRvfxGnp1yge5kGA4ulZZB6HwKe0rEvqy0M7P+D6tA94n3fx5gK+7GQkK/FlJ02JYoMSYIo9HO2sWTyqMx80zkZr0BPStAevPzYNfd4qj7y8EdU4iMibzFpwJQyU30d/7rnn+OGHH0zPly5dyrPPPluo3cyZM1m/fj3Lly/n8OHDNG/enP79+5OUZCzdHhMTQ0BAAIMGDSIiIoLnn3+eWbNmmb3GuXPn8Pf3Z+jQofz111+sWbOG3bt3W7wUd5lQq5nk5GQVUJOTkyu6K0KIKuTIqk9V1ThfUr3s4Kw+NjZIbfzGL2rrV9epu7w7mfYdWfWpqubkqOrOnaq6cqXxZ05Oid5j/XpV9fQ0vZQKqqrTGbebyclRVU9PdXOLnmqz1zeqjd/4RZ3y+OtqtqIxHqQoxgNL+L7l6datW+qJEyfUW7du3dPxuR/d7BwVfFjyo48dO1YdPHiwevnyZdXW1laNjo5Wo6OjVTs7O/XKlSvq4MGD1bFjx6qqqqppaWmqtbW1umLFCtPxWVlZqru7u/rRRx+pqqqqs2fPVtu0aWP2Hm+88YYKqNevX1dVVVXHjx+vTpw40axNeHi4qtFoTOewcePG6qefflqiz1CStnf6MyrNNVSWcQohBJB+4RwAF51cefqpf3Gxrhv1b97gh+B3aZ9w1rydVntP+RoCAoy5om5PUlUod1TujM4BxPLFpg+ZPPgNNt3vi0FR+PSXj7FSDRATY2xXzfJGFJzMWhRVtfxHb9iwIQMHDmTZsmWoqsrAgQNp0KCBWZtz586RnZ1Nr169TNusra3p1q0bkZGRAERGRtK9e3ez43r06GH2/OjRo/z111+sWLHCtE1VVQwGA1FRUbRuXXlznEsAIYQQgEPjZiTWrseoEf8m1rkRuhsJ/Lj2HZpcv1So3T9Rotgjb0Yn4H9mH19unMekIbP4uU0ftKqBT3/52Fhno0C76qKkH8nSH/25554zDSMsXrzYYu+TlpbGCy+8wJQpUwrtq6yTNvPIHAghhAC8HpvAyBHvE+vcCO+kS6z/7wyz4MEAxDlraTf0Zct3xs3N7OmjZ//gy43zsNLnsPF+P758cFiR7aqDkn4kS390f39/srKyyM7Opn///oX2N2vWDBsbG/bs2WPalp2dzcGDB2nTpg0ArVu35sCBA2bH7d+/3+x5586dOXHiBM2bNy/0sLGxscAnKzsSQAgharyMbD0vrDzK+freNExLYvnat3G5ed2035D7M+a96Wity+GXuo8PeHqiFkhI+cjZA7y3dQkACx8aw47uA4ztqpncj05xyTiV/NWyFqXVaomMjOTEiRNoi6hPUqtWLV566SVmzJhBaGgoJ06cYMKECaSnpzN+/HgAXnzxRc6cOcOMGTM4deoUK1euZNmyZWav88Ybb7B3714mT55MREQEZ86cYdOmTaWaRJmVlUVERAQRERFkZWURFxdHREQEZ8+evfvB/4AEEEKIGi1Hb2DyysMcjL5OHTsrXm9yHmuumrWJd9ZyYNEMHpzyUfl0Sqtl/4yRqKoxT0Se0UdDGXnkf6iKhkl+L3Au6dYdXqRq0mph0SLjf98eROQ9L7Ba1qIcHR1xdHQsdv/8+fMZOnQoY8aMoXPnzpw9e5YtW7ZQt25dwDgEsX79ejZu3EiHDh1YsmQJc+fONXuN9u3bs2vXLk6fPo2Pjw+dOnXinXfewd3dvcT9vHTpEp06daJTp07Ex8ezcOFCOnXqxPPPP39vH7yEFFUtKnFr1ZWSkoKTkxPJycl3/IMXQghVVZm1/hhr/ozB1krDf8Z3p1uTeuizszi2/kvSL5zDoXEz2g19uXzuPOTSG/S4jn+Rh5b1L1Q6/JzGmydGLOKmTkvTBrXYOLkXjnbW5da3u8nIyCAqKoomTZpgZ2d3z69TVFEznc4YPEjV83/mTn9GpbmGSgAhhKixftwXzTubjqNR4OsxXXmkjWtFdwmA7efC6Ne5GaR4oEHFh3DciCceN8LxAYcs3J4Nx6p2Ng+3cuG7Z7qi0VSOAlxlFUBA0WXVpdr5P1dWAYSswhBCVGvFXYT+jE7i/Z9PADB7QOtKEzwAhO3SQ4oOMM6/2IWveYN0e66s74bnuD3sOHmZH/dFM65Xk3Lvp6Xd42pZUU4kgBBCVFtF3Qb39IQPFmSw+NxhcgwqA9u78bxPJbv4pt19iUFWgjNdHOtzIPka80NP8lCLhjRtWLscOieEkUyiFEJUSyEhMHRo4aREsZcMzPrlMFdSM2nhWpuPhrZHKW7KfwXxbduyRO0mdOtK7+YNyMg28Nq6o+gN1WpEWlRyEkAIIaodvR4mTjT+9+2lues9fBw7z+uoWVZ8OaortWwr341Y3z5a6rumk7+A9HYG6jdK52FfKz4MbE8dWyuOXLzBN7+fL89uihquXAKIxYsX4+3tjZ2dHd27dy+UWKM4q1evRlEUhgwZYtkOCiGqlbAwuHatcGnuN1q8RZ0uFwG48lNHLvxdq2I7WgytFr750iF33eLtQYQBFIVvFjug1YKHsz3vDDImLvp062lOJqSUe39FzWTxAGLNmjVMnz6dOXPmcPjwYTp06ED//v25fPnyHY+Ljo7m9ddfx6caJkoRQlhWWJgxeFjPUDxzl0Am1q7HbH9jcp6X9q3D/9wewsIqro93ExAA64MVPD3Nh1c8dQrrgxWzpYyBXTzp19qFLL2B19YeJSunuDsXQpQdiwcQn3zyCRMmTODZZ5+lTZs2LFmyBAcHB5YuXVrsMXq9ntGjR/Pee+/RtGlTS3dRCFHNKAY932AcwzB+h1d4/bFp3LB3pG3CWabtXsHXTEQx6Cu2o3cREADR0Qo7d8LKlbBzJ0RHKYXyICiKwtyAdjg7WHP8UgrfhstQhrA8iwYQWVlZHDp0iH79+uW/oUZDv3792LdvX7HHvf/++7i4uJjSgd5JZmYmKSkpZg8hRM2l10Pba2E04Joph+PyLo8T3qQzdtkZBP28EFtDDg25xmDnsIrsaonkLWUcOdL4s7g8CC517JiTO5TxxY6zxCdXvyyVonKxaABx9epV9Ho9rq7m66tdXV1JSEgo8pjdu3fz/fff8+2335boPebNm4eTk5PpodPp/nG/hRBVU0gIeHvDya/DTNtON/Binu+zALy1cynNk/KXZXS6EUZ1MqSjB10b1+VWtp65m09WdHdEEcLCwlAUhRs3bpT4GG9vb4KCgizWp3tVqVZhpKamMmbMGL799ttCtdeLM3v2bJKTk02PmJgYC/dSCFEZhYRAYCBcitXTmGgAMrVWTH38dbKsbPA7d5Cnj2w2O0ZTqX4D/nOKovDuE/ejKPDz0UvsP3+tortUpYwbNw5FUXjxxRcL7Zs0aRKKojBu3Ljy79hdfPvtt/j4+FC3bl3q1q1Lv379SrxY4Z+w6D+fBg0aoNVqSUxMNNuemJhIo0aNCrU/d+4c0dHRDBo0CCsrK6ysrPjxxx/56aefsLKy4ty5c4WOsbW1NRU8uVvhEyFE9aTXGxNGDVGNqy7G8l8APu85gkjXptS/eYOPNi+iULaHapjmsK2HE6O6eQHw7k/HydFX4QmVer1xRuyqVcafesvPWdHpdKxevZpbt/KHgDIyMli5ciVeXl4Wf/97ERYWxsiRI9m5cyf79u1Dp9Px6KOPEhcXZ9H3tWgAYWNjQ5cuXdi+fbtpm8FgYPv27fTo0aNQ+1atWnHs2DFTWdKIiAieeOIJ/Pz8iIiIkOEJIUSRwsPhgdgQggnEI3fVRWRDb5Z0DwTg378tpmH6DVN7FbjqoKB/qHqu8nr90ZY4O1hzMiGV/+6/UNHduTd541F+fjBqlPGnt7dxuwV17twZnU5HSIH3CQkJwcvLi06dOpm1zczMZMqUKbi4uGBnZ0fv3r05ePCgWZvNmzfTokUL7O3t8fPzIzo6utB77t69Gx8fH+zt7dHpdEyZMoWbN2+WuM8rVqzg5ZdfpmPHjrRq1YrvvvvOdK21JIvfwJs+fTrffvsty5cvJzIykpdeeombN2/y7LPGMclnnnmG2bNnA2BnZ0fbtm3NHs7OztSpU4e2bdtiY1N+1fCEEFVHQpyeRUwFVDRAjqLhjQFTyNFaMeDUHvxP50/aVnMfEx9XCYsJr6AeW1bdWja89qgxm+UnW09z7bed5fot/h/LG4+6PY1oXJxxu4WDiOeee44ffvjB9Hzp0qWma1ZBM2fOZP369SxfvpzDhw/TvHlz+vfvT1JSEgAxMTEEBAQwaNAgIiIieP7555k1a5bZa5w7dw5/f3+GDh3KX3/9xZo1a9i9ezeTJ0++5/6np6eTnZ1NvXr17vk1SsLiAcRTTz3FwoULeeedd+jYsSMRERGEhoaaJlZevHiR+Ph4S3dDCFGNtboSjo5Y0y+0H7oO5i+3FjhmpPHe1iVmbWPqQOBw2NAGwqLDyr2v5WVUNy/aOBhIycjh44/Xl+u3+H8kbzyqqELRedumTbNoIPT000+ze/duLly4wIULF9izZw9PP/20WZubN2/y1VdfsWDBAgYMGECbNm349ttvsbe35/vvvwfgq6++olmzZnz88ce0bNmS0aNHF5pDMW/ePEaPHs20adO477776NmzJ5999hk//vgjGRkZ99T/N954A3d3d7MVkJZQLjlcJ0+eXGw0FXaXTC7Lli0r+w4JIaqV9g3zv4RccG7Exz6jAXhr5/e43Lxu2ve+D7znB4ZqNnmyKNqNG3j3u3cYPupD1nR4lIkHQvC+EZ//LT44mEIJJSqD8PDCdx4KUlWIiTG2s9AcloYNGzJw4ECWLVuGqqoMHDiw0MT+c+fOkZ2dTa9evUzbrK2t6datG5GRkQBERkbSvXt3s+NuH74/evQof/31FytWrDBtU1UVg8FAVFQUrVu3LlXf58+fz+rVqwkLC/vH5dTvpvIlgRdCiFLSeBirV6rAm/0nk2FtR48LRxn+11azdjuamgcPvt6+5dfJ8pT7Lb5bbCx+5w6ys9kDfOIzms9+Xmi8ACuK8Vv84MHFJ5aoKCW9I23hO9fPPfec6Yvv4sWLLfY+aWlpvPDCC0yZMqXQvtJO2ly4cCHz589n27ZttG/fvqy6WKwaEIcLIao9Hx/SXeuztl0/9nh3xDY7k3mhX5hWXRiAi44Q3jj/kPr29atvAFHgW/xrv/8HgJ/a+HKiYW7Z8oLf4isbt7uXMi9Vu3vk7+9PVlYW2dnZ9O/fv9D+Zs2aYWNjw549e0zbsrOzOXjwIG3aGBN6tW7dutByyv3795s979y5MydOnKB58+aFHqWZ9/fRRx/xwQcfEBoaSteuXUvzUe+ZBBBCiCpPr8DLj9Vifm7CqGl7Vhpv15Nfimqav/ndh28GfYNWU8m+fZeVAt/O214+z8DI3wH4xOfpYttVGj4+4OmZW0isCIoCOp2xnQVptVoiIyM5ceIE2iLu0tSqVYuXXnqJGTNmEBoayokTJ5gwYQLp6emmLMovvvgiZ86cYcaMGZw6dYqVK1cWGpZ/44032Lt3L5MnTyYiIoIzZ86wadOmUk2i/PDDD3n77bdZunQp3t7eJCQkkJCQQFpa2j86B3cjAYQQosoLvxjOz26Pc93BiSZJF3j+4EbTvljH/EmTAJ51PFk/fD0BrSvh+H9Zue3b+Wvh/0Vr0LPtvu4ccm9VbLtKQauFRYuM/317EJH3PCioXIZe7pZbaP78+QwdOpQxY8bQuXNnzp49y5YtW6hbty5gHIJYv349GzdupEOHDixZsoS5c+eavUb79u3ZtWsXp0+fxsfHh06dOvHOO+/g7u5e4n5+9dVXZGVlERgYiJubm+mxcOHCe/vgJaSoalFTXauulJQUnJycSE5OlqRSQtQQC8PW8HmoAwoaLlvPolvM37ilQXxt47BF3p2H//P5P971fbf63nnIo9cbV1vExZlWLrzh/wprOvTnwQt/sWrNWyienhAVVeYX4oyMDKKiomjSpMk/m8QXEmJcjVFwQqVOZwweKuPkzyrkTn9GpbmGyiRKIUSVZjCobD7siIKBm9owbln9za4mRbft27Rv9Q8eIP9bfGCg8Vu7qjJlz2o23P8w+xu3Z3fjjvh8/H+VbwJlQQEBxkme4eHGoRY3N+OwRWXucw0jQxhCiCpt3aEYzl82oCq3uGH9Q5FtFBR0jjp8vKpn5skiBQQYl2p6eADgkXqF0RHGWiALn3sP9cknK7J3JVPSUqSiQkgAIYSosm6kZzH/f8aqk4O7WKNXklBuq3iR9zzIP6hm3H0oKCAAoqNh505YuZJJM0Zib63l6E0N4WeuVnTvRBUnAYQQospa+Nsprqdn08K1Nh8/OYjg4cF4OHqYtfF09CR4eHD1njR5B3q0hOHLKkbyt60fIx4w5hZYvPNsBfdMVHUyB0IIUXXo9aYx8RO1XFj5hzHV7/uD22Kt1RDQOoDBLQcTfjGc+NR43Oq44ePlU/PuPOQqah6i531NsBkazR9RSfwZnURXb8vUS6hm8/OrlbL6s5EAQghRNYSEoE6dihIbiwr866l/Y/DuwMB6eh5sWt/UTKvRVt8EUaWQV4/KeLHIH9aJO2tHvQhPaneI4cuwcywdV7YBhLW1NWAs6GRvb1+mry3KRlZWFkCR+S1KQwIIIUTlFxKCOnQoeZfCbc27sde7AzY5Wcz690vQPEeW9hWQX4/KPHgAUFWF5ANNqdUuhh0nL3P8UjL3uzuV2XtrtVqcnZ25fPkyAA4ODijFJYUS5c5gMHDlyhUcHBywsvpnIYAEEEKIyk2vJ3PsRGwwTtrK0ljxb7/nAXj+4EY8Ui6TPnEaDpWxrkMFyc9kXfSFOyepNukn3anV5hJfhZ3ji1Gdy/T9GzVqBGAKIkTlotFo8PLy+seBnQQQQohKTf/Bv7FNu2Z6/mOXx4mu506DtOu8vH8dGlQcrsWgDwtH29e34jpaicTFGbjbHPnk/c2o1eYSvx6LZ/qvO2iaklhmuRYURcHNzQ0XFxeys7P/0WuJsmdjY4NG88/XUEgAIYSovPR6Mj9ahEPu02v2jizqOQKAmb//SO2sW6amp8LiadO3AvpYCV3R/gV0vGOb7CuOtDWk8bemNl9/sZEPQz837vD0NCahKoMhIa1W+4/H2UXlJcs4hRCVlj4sHIdbSabnQb1HkWpXmzaJ5xj693aztvFUwroOFaRhq5PgGEN+KbHbGXjS/mveW/kuACFtHya+Tu5E1Lg44+zLkJDy6KqowiSAEEJUWqfC8qtFnqmvY0XHAQC8vf07tGr+xfEq9eChnuXev8rKw7kR+E/NfXZ7EGFAg55FTKNz3Em6XzxGttaa5Z0fN+7OW+I3bZpxNqYQxZAAQghRacUZXE3/Pd/3WQwaLY+e3kePmGNm7RZZTYCme8u7e5WWj5cPng8ehOHDwDHOfKdjLD6+fdHdykABxudWLl3VwZ90a1tjG1WFmBjjbEwhiiFzIIQQldaJ+51ohScxurpsb94NrUHPrLD8ehcqcJX6zO2lpU16fPEvVMNoNVoW+S8iMCUQtdUmuNAb0tygdjw03o378fw7C33PHaTx9UtcqOvO+vsfZkzE//JfKF7OqSie3IEQQlRaDdqcYYr928zzew6AURH/o+n1S4DxxryKwgs2CzD0mY9bHZkDUVBA6wCChwfj6ewGTXZBu9XQZBc6Z3deHPSeqZ1WNTDu0M8A/NB1MIaCSz/d5JyK4ilqNcs3Wppa5kKIyi0sOoyB7/9CQxc/7DMzCP9mPA3SkwG4iI5pfMKG4Stp2HU38a/F19iU1XeiN+gLp/ZWAW9v44RJVSXNxp4eLy8j1bYWP6x7F7+oQ8bVGFFRklujhinNNVSGMIQQlVY39140aHQBDJAQ4UVg+kbciCceN8LrNMYw4DVos4HFj62V4KEYxab2XrTIuNpCUaiddYunjv7Gd92e5PsHBhsDiKAgCR7EHUkAIYSotFb+EYNiaEAO10j2CWRX4+4FxvLDQWNgRs8ZDLt/WEV3teoJCIDgYFO1rbGHfmZp1yfY7d2JUz+up2XAkxXdQ1HJSQAhhKiUktOz+XyHseT0qJ6OrDrXgFjtLtP+hg4N+fKxLwm8P7Ciulj1BQTA4MEQHo4uPp7+CTb8L1HPUrtmfFjRfROVngQQQohK6cuwsyTfyqaFa23mPf4Q8xgkZbotQI+WcHyJBzp7J/G/xH1siIhjpuM16l9LKLP01qL6kQBCCFFp6LOzOLb+S2KiYvkhpQ+gYdaAVmg1CiBlustaSAhMnaoSG5u38qIu7Z41kOICK9/9mlf2rTFuLsP01qL6kGWcQohKYf9nM0ls6EDHka+y+/csslQNHS4dx37LxxXdtWopJASGBqrExuYvxHuSDbz7x6cArOrYnxwl9xIh6a1FESSAEEJUuP2fzaTb1AU0StZztr4n69r1A+Dt7T/QfdoC9n82s4J7WL3o9TDx5fTctNXGy4AxvfVUBpzaTd30ZC45urCj2QPGAyS9tSiCBBBCiAqlz87C651PAOMvpAU+z2DQaHnk9D66XjoJgG7OJ+izsyqwl9VL2C491xIdKHgJ8CEcHbHY67MZfmwbAP/t9Fj+QZLeWtxGAgghRIU6tv5L3JP1aIDD7i3Z0rInGoOemb//CBh/SXnc0HNs/ZcV2s/qJOzvU4W2uZGftnpUxP9QVAO/N+3CBedG5g0lvbXIJQGEEKJCpV84BxjrWnzYZxwAQ//ewX3XYopsJ8pA7cJBQMFy6I1vJPBQ1GEAUwVUE0lvLXJJACGEqFC2ns0ACGvahT+82mGTk8Wru1cU2078c759tOAYQ8FS3+H4EIOnqRbGmMObAVjb/hEyrGxAUUCnMy7pFAIJIIQQFeyGy8tcQMdHD40FYOzhX3BPvWrab0DhIjpuuLxcUV2sdnyb+FA/4P3cZ4bc/9cylUW5/63gd/5PPJIvc8PekV9b9TY2lfTWogAJIIQQFeryVRvGtv6MSNem1M68ycv71pn25X0bnkYQl6/aVFQXqx2tRss3MwfA8GHgGGfavoEAAu2/Is7eDq1qYOTRUAD+232IMe215IEQBUgiKSFEhWroauCMjxPWpDPij23UzUg17YvFk2kEsYEApsjQe5kKaB3A+ndhSrfexB1rYqoxcrBdNPsf+YFrJ+JpGhWDVYrKkQZN+bt7b9pWdKdFpVIu5bwXL17MggULSEhIoEOHDnz++ed069atyLbffvstP/74I3///TcAXbp0Ye7cucW2v52U8xaiavnP3gu8/dPf6G/aEP/1Q/TO3pdfcRMfVEUrlaUt6PZy31dvXmVa6GumoKKBRxdq2bfG1/kKyxrbSmrraq4011CLD2GsWbOG6dOnM2fOHA4fPkyHDh3o378/ly9fLrJ9WFgYI0eOZOfOnezbtw+dTsejjz5KXFxcke2FEFXXrSw9n+88A0DyvuYYcmzZhS+rGckufFEV40VKht4tJ6/c98h2I0m6lcSw91cR9/5uWB4G61fRcYNx8uqBy7VJGzce/PzA21uyUgrLBxCffPIJEyZM4Nlnn6VNmzYsWbIEBwcHli5dWmT7FStW8PLLL9OxY0datWrFd999h8FgYPv27ZbuqhCinC3bG83l1Ew869rz3SwvPDzM93t6ytB7edEb9Ez86H+wdh2kGP8gniSE0JiheCfFkW5jnz+ZUlJbCywcQGRlZXHo0CH69euX/4YaDf369WPfvn0leo309HSys7OpV69ekfszMzNJSUkxewghKr/kW9ks2WXM7fBqvxYMD9QSHQ07d8LKlcafUVESPJSXsKhwroW8k/tMY0ptraAy4uhvAKxu39+4W1JbCywcQFy9ehW9Xo+rq6vZdldXVxISEkr0Gm+88Qbu7u5mQUhB8+bNw8nJyfTQ6XT/uN9CCMv75vdzpnLdQzoZv/FqteDrCyNHGn/KsEX5CdulhxQdeZeFvNTWGiDg+Has9Dkc8WjFqQaNjQdIausar1Iv45w/fz6rV69mw4YN2NnZFdlm9uzZJCcnmx4xMTFFthNCVBy9QU9YdBirjq0iLDqMhOR0lu6OBuD1R1vmlusWFSrNfJlLwdTWLjdv0PfsAQDWtH/E/DhJbV1jWXQZZ4MGDdBqtSQmJpptT0xMpFGjRsUcZbRw4ULmz5/Ptm3baN++fbHtbG1tsbW1LZP+CiHKXkhkCFNDpxKbEmvapuN1NNm+dPJy5pE2rnc4WpQX37Yt+VeB5wVTWwOM+Os3trTsSUjbh3lj1zJs9TnGHZLausay6B0IGxsbunTpYjYBMm9CZI8ePYo97qOPPuKDDz4gNDSUrl27WrKLQggLCokMIXBtoFnwYGVwRbllnIzXo/VlFEXuPlQGvn201HdNJy8z5e2prR+KOkyj1KvcsHfkt/t6SGprYfkhjOnTp/Ptt9+yfPlyIiMjeemll7h58ybPPvssAM888wyzZ882tf/www95++23Wbp0Kd7e3iQkJJCQkEBaWpqluyqEKEN6g56poVNRUdEYoE8UjDgGLW48jYIVtzSH+CJiKnqDTMKrDLRa+OZLB2NggKFQamutamD4X1sBWN3hUVRVheefr8Aei4pm8QDiqaeeYuHChbzzzjt07NiRiIgIQkNDTRMrL168SHyBMbSvvvqKrKwsAgMDcXNzMz0WLlxo6a4KIcpQ+MVwYlNiefIERAdB2HJ453dv0u36ANAp5kdiUmIIvyiT8CqLgABYH6zg6Wm867CBAAIJJk4xDjkP+2srimpgj3cnYp1cYc4cyQlRg5VLKuvJkyczefLkIveFhYWZPY+OjrZ8h4QQFhefGs+TJyB4bf62hQ89g6poGBj5O5//dI7A4cZ2ovIICIDBgxXCdukZ/sNUNmj/ZpPHZd78tTHvHb1Ar+gIdjfpzNr2j/Ba+H9RY2NRhg6F9etlzW0NU6lXYQghqi43BxcWGWsxoQH+9GjN9ubd0Br0vBb+XwCCQo3tROWi1YK2aThJzRfDrXrw+VkmHjUONY34y5gTYl27fugVDaYZLBMnSk6IGkYCCCFEmdMb9NT5IwJdivGXjAp82GccAMP/2krT65fQAF4p4HOhAjsqihWfGg8nnoS1wfiknjflhHjkzH7qpieTUKcB4d4d8w+4dg3+/e+K6q6oABJACCHKVEhkCN6LvFkY8rppW1jTrhzU3Y9tdiZT9q4ya69NLLoujqhYLg5uEGqcROlG/lJ8W30Og0/sAiC43W0J/j77TO5C1CASQAghykzBZZvxtY3b9IqGD/uMBWDs4V9wS71mfpDkEaicLviYMlPenhMi8G/j0vzf7nuQZNta+TuuXZPMlDWIBBBCiDJRcNkmQHhjiHGEjW36cNKlCXUy0nh5/7r8AySPQKV2OTE/j3g4Plwjvx7R/YnnaHU5iiwrG35u/ZD5gZKZssaQAEIIUSbylm3mMWjgFX8rPvF5GoCX9gfjnGHM52JAMdZjkjrdlVbBG0MGtAQx1fRcIf8uRHC7vsUfKKo1CSAqyO21ASSZjqjqCi7HzEscdUP/AnFOrjRIvc6zh3427Y/Fk0CCCUGW/VVWPj7Gcup5iULn8hZXqZ97fwkGHw9Da9AT4d6Ks/U9MQCxThr0vXpWVJdFOZMAogLkTTLzW+7HqJBR+C33w3uRNyGRkoxFVF1udYzfPPMSR/28yp4Ud+PF5Nk9vzI3501GshJfdtKEKDYoAVINuhLTamHRotwniooBLRP5BhUFA9Aw/QZ+5/4EYG1b42TKKf0NhMftrZgOi3InAUQ5K6o2AEBcShyBawPvLYjQ6yEsDFatMv6U38iiAvh4+fB8dH2C14JHCnzbLYAkByeaXotlwl/reI93ycSWXfhiQCvVoKuAgAAIDoaGrplAgcyUeAIQ+Pc2ANbf34+hPbuzoY0kBqtJJIAoR8XVBugTBYrBeGNwWui00g1nhIQYU8n6+cGoUcafklpWVACtCov+Z/zvaw7OfPfAEABeD/8PNqrx73QQ09Bg/vdb5txVbgEBcCHKGqXWFcDABgLwJhpfdvLD2VfQpMO1Os6EJv0HDBrTnShR/ZVLKmthVLA2wKJQ0KXk74txhKn+KhvaGGsD+Hr7FvkaqqoaJ58BbNgAw4ahqAbM6hnGxUFgoPGrQ3GpZfV641e/+HjjpCcfH5nMJv6Z8HAcEo1LND/rNYJ0G3s6XDrFgFN7ANCg4kUMPoSzC1/TYTLnrvL7Y78W9WZD03MDWuOfoQHqnjiOY9doardIxeaPD/GZI6tqagoJIMpRUbUB8nikwLq1MGRkXfacvcylxIucv5JGzPV0km5mkXQzi+vp2dxIz8KQF0BgCzN/wjY7k/q3kqmXnkL99Bu4piXR4upFWsz/mpZ+/XFxdjAvmRwSAlOnQmyBYRRPT+OAp+SyF/cq91bCuXoerOw4AIA3di3j9mLdbhjbKYrxr52s4qz87nSXKO2YJ45do3G4L5HM/7wIqnwRqSkkgChHt9cGSLWx56hbC464tyLCvSVH3FuS5ODEsW0Ax0r8upnWtlyyduGSYxE1BT4Mw9nBmge869GrWX16RUfQfEwgiuk2Rq6S3LUQ4k5ybyV89NBY9BotD589QM+Lhf8ex+NmmtkvqzirhjvdJcq+7EjW5TrYuKSiul8jPLw2vr7l1jVRgSSAKEcPRqn8Xac169p3Yrd3JyLcWmDQmP/21Bj0eNayoqlXA5o2qI13Awca1LalroMN9WrZUNfBGmutxngXYeJEVEXhprUd12o5k2TvyNVazsQ6uXCmQWNONfAiur4nN9Kz2Xoika0nEgEbGr68nH5nD/DEiV10iz2OVjWAqhq/Ek6bBo8/Dnv3yvCGKB0fHw526sOWlj3RGPTMDvvBbLcBhVg8CccHT09j8CCxatXg4wP16kFSUlF7FdL+9qTew5HUahtLfHzj8u6eqCASQFhYWmYOYacus+V4ImHHMkh9eoHZft2NBDpdOknHS6fpGH+KNonnsfvPchj52J1fWNcIMlIBqHcrBV1K0fUEMrbtIPK+juw7f429B89wMOEWV2rXY1VHf1Z19Mc19RqDIn8n4O8dtLkSZZwW7+EBV6/mv4gMb4gSUDUa5j75GtyE4ce2ct+1mPx9ioKiwoVpQWwfrJWYtIrRao2jnnPmFL0//YQ7dX0jsfO4gcbxJlCr6IaiWlFU9fZ72VVbSkoKTk5OJCcn4+joWCF9SM3I5rfjifx6LJ7dZ6+SlWMw7XO+lULv6Ah8oo/QOyoCj9QrhY7Xb9uJtq/vnd9Erwdvb9TY2EJjzGCsfqjodBAVlf+betUqMsaM5aDufn5p5cP/WvYixa626ZjuF4/x3J+b6Hf2gPGuRJ68+80yvCHuYPOxeF5ecRhbVWXdkjdon3LCtC9OqyNmehAPfiR/f6oqvR5cXY3lLoriMuwA9k2v8IrffbzWv0X5dk6UmdJcQyWAKCOZOXrCTl3hp4hLbItMJLNA0OBd34H+bRvRMN2FfsM64aXGoqHwac+7xXt+WxS+fe/+9ezH8fN5eumbAGavl7cm47/PzeWZ72flHxAWZlzmmddnrRW7mnRh4/1+bGnRA33ucIrX9XiePfQTI49uwS4ny9g4b8ZbwYBEiFxZOQYe/XQX0dfSSd5zHym7m+FDOG7EE48bu/HBoGglBq3iQkKMU6Vuv2ooCji0jqPBoAi86jmwa4av+cRtUWVIAFFOAYSqqhyLSyb4UCybIi6RfCvbtK9pw1o80cGdAW3daOFaG0VRWLUK1o0KIZhAoOiLfiDBDFsZwMiRd37vrGw9Dg0TeSJ5L4t4FR35KyouomMan/CTcw/SLzfCxjr3gp9714K4uEK/AeLr1OfHTo+zsqM/yfZ1AHBNvcaUPasYfmwr1nm5KXbuRGZIidst2xPFuz+fQL1lS8wSX9SswqOjEoNWD0Ut4tLp4KOP9XxwbCs3s/QEv9iDrt71in8RUWlJAGHhAOJqWiYbDscRfCiWU4mppu2ujrY80cGdwR09uN/dsVAEnncD4ElCWMTUIi76QWwgoETX6KDVEbw6siMAGvRm3/bC8cGA8Tf0p6simDaiY/6BuV8hVDBbiaFiLJCTbm1LyP0P89WDw4hzMq7qaHz9Eq/uXskTJ3ahWbmCu0Y3okZJvpWN74KdXE/P5tqWtqRF3HkSncSgVV9xaWReX3eU4EOxjHjAE/8u14lPjcetjhs+Xj5oNRI1VgWluYbKJMoS0htUfj99hTUHY9gWmUhObjIGWysN/e9vRGAXT3o1b4BWU/xtu7ziNBvjAtikDi500VcVLboSros/dyHd9N+mpC53aQdAQAD7Xw/G65OpuOvzA5hrmoY0MFzBITuTpyP+x7BjW1nVwZ8vej7FhbruTBv0Oj90GcS/HFxod/fuiRrk8+1nuJ6ejYttbS4c1d21vWSerPq02qKDwIBOHgQfimXVn6f48NjToBjvyno6erLIfxEBrWX8qjqRAKIEvt8dxXfh54lPzjBt66BzZnhXTx5v746TvXWJXievOE1gIKiKll2qr2lfadfFN2vsUKL3vL1dSAgELgxAUZ/Ah92mAGavoQfnaI4ncSio2OpzGHf4F4Yd28ayLoP46sFhHHVvyRP7M3ha/ZvXH22Jk0PJPreovqKu3mT5vmgAxrRtw0H17tnxJfNk9XUp63dylBSs1IY4GLqRrjVmIc2r9RM8PFiCiGpEhjBKYP7/TrJk1zmcHawJ6OTJUw/oaNmozj2/XnFjiKVZF583B0Kf3IiiS5oY0DrHm82ByJsCERubN2Bh7knWE8wwoPD8jMu16zH37aX8lGR8rfq1bHj78TYM7uguk6VqsAk//snWE4n4tmzI9890K26KDSBzIKo7vUGP9yJv0q71xSlnGPYZ++kY+y/ia0N4Y1A1Cp6OnkRNjZLhjEqsNNdQKaZVAqO7e/HFqE788WZf3hnU5h8FD2AMEqKjjWPBK1caf0ZFlW52uo21lunvX8x9Zrhtr/H59Pdi8idQYhyzNAYtRV/wNzA0t9Keh9n2WDyZlPYNvs2fYOXz3WnWsBbXbmYxbU0EL/73EFfTMkvecVFt7D17la0nEtFqFP5vYGuz8s+3x5SSebL6y6v182D0DgCyrbuyeLMjYcuN5d2HnFCJSTHW+hHVgwQQJaCr58Dj7d2xtSq733x5Y4gjRxp/3ssv1Y+mPMiMRQfQOiWYv7ZzPDMWHeCjKQ+abY+Luz3QKKxgpb2RrMSXnTQhihDlSaZNg+5NGvC/qQ/x2iMtsNIobDmeyKOf/k7o37kD21JavEbQG1Te/8WY52HMg41p7mIMqvPKP3uYx6B4ekoakeour9bPrytiaBt/hhytFT+3fggw1voJXgtPnpBy39WJDGFUA1nZer5cf4xzF9Jp1tiBl4e2M7vzkKfgyo17VXAG/fFLyby29ignE4wrUZ6sr+f9L1+jTvTZ/AMki2W1tOrARWaHHMPJ3pqw132pW8vGbL8Ue615ws5tp1nnfnikwLIuT/B+v4l0vHSSjf95HTDeF411hPOHt+HbrG/FdlYUS4Ywahgbay3TRnTk8zd6Mm1ExyKDB4CGrU6CYwyFhzxKruBdjPvdndg0uReT/JqhQWXDNS2DHpnJ3y5NCx5gnDUaEnLP7ykql5SMbD7+7RQAU/veVyh4gLK5wyaqFp8LoEsxXlQGRf6O1qAnwr0VUXXdAeN2rxRjO1E9SABRg3g4NwL/qbnPip43cTdXtH+ZPbe10jKj332s27IA95TLRNdzJ2DMQv7TcYBxGqaqGh8TJ8L27TKkUQ0EbT3D1bQsmjasxZgeUjhJGGkT8+vxNEy/Qe/oIwBsatOn2HaiapMAogbx8fLB88GDMHwYOMaZ73SMBfurFB9IGMDxovEuxu3Cw+kS8Tubf5hCvzN/kGVlw9v9JzH5iZmk2dgb21y7Bv36GZeByN2IKutUQqpp2ea7g+43VoYVAgqtz33yeBgAG9v4mSful3W81Yb8669BtBoti/wXobTZANOawFhfGDrS+HNaExg0MbdlMXcn/KcZ72LcLjczkHNGGt+GfMD/7fgOK30Ov7Z+iCfHLDTdwgSMy0CGDoX335e7EVWMqqq8s+lv9AYV//sb8VCLhhXdJVGZ+PiQXt/TlJb/0TP7cMi6RXQ9dyLcWmBAIb2+rmSZ8kSVIAFEDRPQOoDg4cF4OrtBk13QbjU02YWnszv1u/xe/N2J4cPQPfgnPl5F/OMv8I1CAZ4/uJG1K9/ANfUaZxo05olnPmFn067mx8yZI3cjqpifjl7ij6gk7Kw1/N/jrSu6O6KS0aNlKsZ1vAYUHLIz6X96HwAb7jcW8ZtGEHpkQkx1Iaswaii9QU/4xXCzXPWbTm0icG0gqkEDF3pDmhvUjofGu1E0huKzyOn1pLt6Y3ctziwB1eVazrw05E0OebZBUQ28/vt/eHn/OvMsFIoi6/uqgLTMHPp+HEZiSiavPdKCV/reV9FdEpVMUbV+djXpzNjh7+OUnkrTxTZsNJSs1o+oOLIKQ9yVVqPF19uXke1G4uvti1ajLfbuhM7Z/Y4paG//5pHH5eYNVq16k1FH/oeqaFjQZyyvPDGTDKvbZu1PmybDGZXcZ9vPkJiSSeP6Dkx4qOndDxA1Tl6Nk4K5ZL6Kfh3NTUh2qMMW715m7UTVJ7UwhJmA1gEMbjm40N2JO6WeDQ+H764FcI3gQlVGbQw5zP1tMW0Tz/LOIy/xS+uHiHN04ZuQf9Ew/YZxhUZMjPFF5GtJpXT2cipLd0cBxomTdsUsExY1W8G5kaYCfyrUjTyOY9doat1/iVvnXWUOZTUidyBEIUXdnbiTvNwQed88HmYb16hnNhVz1NEt/Lj2bZxupXLEoxVDnvmY0w288hvI15JKyWBQmR1yjByDSr/WLvi1cqnoLolKKq/a8O1pzG8eN6Yltb8vAV2THJlDWY1IACH+sYK5IQxo2UlfJvAtoJgFET0vHiPkv6/T+Pol4pxcGfr0An737gSA3kW+llRGqw/GcDD6Og42Wt594v6K7o6oxIqrhZKV4ET2tVporA2MnpUgScWqkXIJIBYvXoy3tzd2dnZ0796dAwcO3LH9unXraNWqFXZ2drRr147NmzeXRzfFPSoqw+UGAooszNUsKY4N/3mdbjF/k2pbi2eHvcvXbYcRjnwtqWwup2Qw73+RALz2aEs865ashLyouYquhaKgjTVuiNHEFXmcqJosHkCsWbOG6dOnM2fOHA4fPkyHDh3o378/ly8XnY1s7969jBw5kvHjx3PkyBGGDBnCkCFD+Pvvvy3dVXGPistwaRzSuMDbvIdaYE+9Wyn8Z83/Mfj4TvQaLfMGjuU/R89RzRYEVXnv/XyC1Iwc2ns6Ma6nd0V3R1QRRVUb3vWjMYDYe+4ql1MyKraDosxYfBln9+7deeCBB/jiiy8AMBgM6HQ6XnnlFWbNmlWo/VNPPcXNmzf55ZdfTNsefPBBOnbsyJIlS+76frKMs/zpDXq8F3kTu78bhAZBiq5Qm4JLu/JE48VTD31GYg/jXN6R3XR8MLgtVpLdsMJtO5HI8z/+iVaj8NPkXtzv7lTRXRJVXMCXezh88Qb/N7A1z/vISp7KqtIs48zKyuLQoUP069cv/w01Gvr168e+ffuKPGbfvn1m7QH69+9fbPvMzExSUlLMHqJ8FZnh8sFPMN5zMJ9gWbBMeDPOc+D3QVz73R1FUVl1IIaJ/zlEelZORX6cGi8tM4e3Nxnv+D3v00SCB1EmhnQy3oXYFHGpgnsiyopFA4irV6+i1+txdXU12+7q6kpCQkKRxyQkJJSq/bx583BycjI9dLrC336F5RXKIeH/GgwPNMtqmbe0azUj2YWvKWdEmu4DXnnUDlsrDTtOXmb0d39w/WZWRX2UGu+j0JPEJ2fgVc+BaX1bVHR3RDUxsJ0bVhqFY3HJnL2cVtHdEWWgyt8rnj17NsnJyaZHTExMRXepxgpoHUD01Gh2jt3JtAenQZsNMM07/46Ew23zXgqkyJ7ax4+VE7rjZG/NkYs3GPb1Pi7duFUBn6Jm23P2Kj/uM9Zb/veTbbG3kSnzomzUr21rqp+yKUImU1YHFg0gGjRogFarJTEx0Wx7YmIijRoVUZQJaNSoUana29ra4ujoaPYQFScvh8Sn/T9l/fD1eDq759+ReN3ttgJeTVHabCDIPwitRkuXxvUIfrEHbk52nL2cxtCv9nImMbWiP1KNkZKRzYx1RwEY3d0Ln/ukWJYoW4M7GgvrbYq4JJOmqwGLBhA2NjZ06dKF7du3m7YZDAa2b99Ojx49ijymR48eZu0Btm7dWmx7UXmZ3ZHoPo2GtevfNUX2fa51WP9ST5q71CY+OYPAJfs4dOF6BX6KmuP9n09wKXfo4s3HpFiWKHuPtHHFwUbLxaR0Dl+8UdHdEf+QxVdhrFmzhrFjx/L111/TrVs3goKCWLt2LSdPnsTV1ZVnnnkGDw8P5s2bBxiXcfbp04f58+czcOBAVq9ezdy5czl8+DBt27a96/vJKozKq6gCXsVlubx+M4vnlh/kyMUb2Flr+OrpLvi1lCyIlrL1RCITfvwTRYG1L/TgAe96Fd0lUU29uiaCDUfieKZHY94ffPff6aJ8VZpVGGBclrlw4ULeeecdOnbsSEREBKGhoaaJkhcvXiS+QBrjnj17snLlSr755hs6dOhAcHAwGzduLFHwICq30qTIrlvLhhXPd8e3ZUMysg1MWP4nG47EFtte3LtraZnMDjFmE53o01SCB2FReasxfvkrnmy94S6tRWUm5bxFpZatNzAz+C82HDFOunr78TaM792kgntVfaiqykv/PUzo8QRauNbmp8m9pViWsKgcvYEH523naloWS8d15eFWrnc/SJSbSnUHQoh/wlqr4eNhHXiulzFo+OCXE3wUelImYJWR//5xkdDjCVhpFD4Z3lGCB2FxVloNj7c3TqbceERyQlRlEkCISk+jUXj78dbM6N8SgC/DzjFr/TFy5PbnP/J3XDIf/HwCgFkDWtHWQxJGifKRN4zx24kEbmZK4riqSgIIUSUoisIkv+bMD2iHRoE1f8bw4n8Pk5Gtr+iuVUmpGdlMXnmYLL2Bfq1dZFhIlKsOnk40aVCLjGwDv50oOkmgqPwkgBBVyohuXnz1dBdsrDRsi0xkzPd/kJyeXdHdqlJUVWV2yDGir6Xj4WzPwmEdUArWXxbCwhRFMeWE2CDDGFWWBBCiyul/fyP+81w36thZcTD6OsO+3kt88p2zVmblZBG0P4hXNr9C0P4gsnJqbqrslQcu8stf8VhpFD4b2QlnB5uK7pKogYZ0NA5j7D5zhSupmRXcG3EvJIAQVVL3pvVZ+0IPXOrYcjoxjYAv93IqoeislTO3zsRhrgOvbnmVLw5+watbXsVhrgMzt84s516XD71BT1h0GKuOrSIsOgy9IX+Y52jMDd7Lnfcw078lXRrXrahuihrOu0EtOuqcMajw81G5C1EVSQAhqqzWbo6EvFwwa+Ve9p27ZtZm5taZLNi7AH22AvumwubPYN9U9NkKC/YuYNDKQYUuslVZSGQI3ou88Vvux6iQUfgt98N7kTchkSEkJGcw4cc/ycoxznt4vreUVBYV68ncyZQbpTZGlSR5IESVdyM9i4k/HuJAdBI2Wg0fD+/AoA7uZOVk4TDXAf2Wf8O+10C1yj9IyYEeH8OjswCoZ1+Pqd2n8pbPW3dMcFWZhUSGELg2EBXzf9IKCopqSw+HtcQmaWjhWpv1L/Wkjp11BfVUCKOraZl0n7sdvUFl+2t9aNawdkV3qcaTPBCiRnF2sOHH8d14rF0jsvQGXll1hC/DzrL44JfG4GHvTFBvCwpUrXH7b/MBSLqVxJywObgudCUkMqQCPsU/ozfomRo6tVDwAKCqUD9rGrFJGuo5WPP92AckeBCVQoPatjx0XwMANh2RuxBVjQQQolqws9byxcjOpoRTH4WeInivA/zxam6L21cZ5D7f+zqcfRgMxn8K125dY+jaoby/6/0i5xBUVuEXw4lNMab61higTxSMOGb86Zw9EgdDb1SyefERLbp6DhXcWyHyDTENY0iFzqrG6u5NhKgaNBqFdwa1wbuBA+/+dJy4Kx64Dj/ElQ1d0GTAJL6kGec4RzMW8zI52ABa+O92sL8G3YPgobmgMTAnbI7pdd1ru/P5Y5+bVQ2tbOJTjfVknjwBi0JBl2Lcvv7+h3mt0SgArll/QZ3a4yuqi0IU6fYKnTKxt+qQOxCi2nmmhzdLxz2AlarFziuJtmM2cbzefQTxKq/wBUG8yi0cmE+BVRi36kPYB7AgEU48afZ6l9IuMXTt0Eo9tOFWx40nT0DwWvDIDR7+16InMx6bCsDEP9bz6OntuNVxq8BeClGYg40V/e9vBMBGGcaoUiSAENWSb0sXRjToRe3kDJLrOTD0mQVsa9bNtF+LnpksMA8iwBhIrA0uFEQAjN04lu3nt1fKoQ0fj558scU4z0MD7GzahSlPzMCg0TLsr63MDFvG579p8fHoWbEdFaII+RU6L0mFzipEAghRbb3zki3bfpxI15jjpNrW4vnAdwjqNRIDimlGxOss5GG2oiEvGMjd88tX8NcoiOpjmh+RlpVGv//0K7Q8sjLQ7tmLe7IeDbBf15YXh7xJttaagZG/Mz/0c6xQ8bihR7tnb0V3VYhCejWrT4PatlxPz2bXqSsV3R1RQhJAiGrL5rsvaZSexMrVb/HMoV8ACOo9mokBb5Fi44ACaFHZzqNcxoX/4/3cQEID6a4QsgKWh8En0UXekYhLiSNwbWDlCCLijXMgjri1YPzQd8i0tqXv2QN8+ssnaFVDoXZCVCZWWg2DOhiH1zZITogqQwIIUX2dOweAjSGH97ct4aPNQdjkZLHtvgcZNC6Ivxo1NzWtTxIfMIdEXHmS2wKCNM8ihzXylkxOC51W8cMZbm787t2J0SP+zU1bB3pGH2XxxnnYGHIKtROiMgro5AnA1hOJpGRIfZuqQAIIUX01a2b2dPixbaxb8QbuKZe5UNedoU8v4LsHhmAosMSzPtcIZuhtQYRifGz8wTSckUdFJSYlhvCL4Rb8IHe3oU4znhv2Luk29vSKjuDbkA+w0xf4JawooNOBj0/FdVKIO2jr4Uhzl9pk5Rj43zG5U1YVSAAhqq+XXwateQKpDgln2PzDFPqf2ku21pp/Pfw8zwXO4aqDE5CfLeIrXmQUK+hDWO6whgJZTsZ02MdGmM2NgPxllOVNVVW++f0cr677ixyNlidO7GJp8HvUys7Ib5NXaTMoqND5EKKyUBTFlNo65LAMY1QFEkCI6svGBqZPL7TZOSONJRvn8q8ti7HNziSsWVf8n/uCzS17AcZ/FK5cYQVPE4Yf0Xjl35HY+gmsX2WcGxEUbRrWqIjlkdl6A+//coK5m08C8HzvJoxo9SDXcDVrd0njyf7XgyGg8uaxEALyV2P8EZVE7PX0Cu6NuBuphSGqv5kz4ZNPQF94nsKpBo155YmZnG7YGID+p/bywdavcLl53dRGzX0Esp4NFLwIGycn1h/3Ionff1WuNTRir6czZdURDl+8AcBbj7Wm/tWmBAaCourxIRw34onHjd34YFC0BEsMIaqAEd/sY//5JGb0b8kkv+Z3P0CUqdJcQyWAEDVDVhZ88QXMmQNpaWa7MrVWLO7xFF8+OIwcrRWOGWm8tfN7hv21DU3uREkVSKE2T7IRVy4Tjxvh+GBAoX6jDBJjHcptdGDL8QRmrDtKSkYOdeysWBDYnkdau+HtDbGxRR+jKODpCVFRMoohKre1B2OYuf4vmjWsxbbpfVCU29PQC0uSAEICCFGckBAIDDRWmLpNZENvZg6YyjG3+wBoH3+aN3cu5cGYv4t8qRg8mcoiNhDAzp3g62vJjsPNzBw+Cj3J8n0XAOigc+aLkZ3Q1XMgLAz8/O7+GuXRTyH+idSMbLr+axuZOQZ+ntybdp5OFd2lGkWqcQpRnIAACA6G+vUL7Wp9JZoN/3mNN3d+T63MdP5ya8GIUfN5PuBtztb3LNTegziCCeRJQiyaXsFgUNlwJJaHPw4zBQ8TH2rKuhd6mApjlfT9JQ2EqOzq2FnzSBvjPJ6QI8XcUhOVggQQouYJCIDERHjvPahXz2yXlWpg4oEN7PpmAmMO/4LWoGfbfd3p/9xipj7+ulnuiLzhjSCm4eZSujwQ+uwsIlYHsffDV4hYHYQ+O6vIdhExNxi6ZC+vrjlKYkomXvUcWPbsA7z5WGtsrPL/+ZY0vYOkgRBVQUBn42TKn49eIkdSW1daMoQhaja9HsLD4bvvYMWKQrvP1vPkwz5j2dqih2lbt5i/ee7gJh45+4cpy6N+2060fX1L9Jb7P5uJ1zuf4J6cH3RcctJy8f3pPDjlIwwGlV2nr/Djvmh25qb1dbDRMvnh5jzXqwl21oUnMej14O0NcXFFjs7IHAhRpWTrDTw4dzvXbmbxw7gH8Gvlcsf2eoOe8IvhxKfG41bHDR8vn3Kd1FydyBwICSBEaWVlgb09GIr+tvO3azO+7zqYn1s/RI7WCoCGaUk8emY//qf38uC/Z2I9auRd32b/ZzPpNnUBYH77z4DC5VrOfP7mu4RbteNiUv4StoDOHrzh3wpXR7s7vnbe9A4wDyLy5qDJKgxRlbz703GW7Y3m8fZufDGqc7HtQiJDmBo6ldiU/OEOT0dPFvkvIqC1/IUvLQkgJIAQ92LmTFiw4I5NEmrX58fOA1nZ0Z8b9vl/v5yswae1G+08nLjf3Yk27o7Uq2Vjdqw+O4sE53q4pd9EVTRcqVWXCPcW7PXqwN7G7TnbwMvU1tHOimFddTz9YGOaNKhV4o8QEgJTp5qvxtDpjDmkJHgQVcnfcck8/vlubKw0HHyrH0721oXahESGELg20JRWPk9eubzg4cESRJSSBBASQIh7dYecEQVlaazY27g9W1r25LeWvbhmV6dQmwa1bXC0t6aWjRUONlqyrsRgOB1HYp16XKlVF/1tt1gV1UD7+DP0fMCZKS+Mxd7m3m7B5o3KxMcb5zz4+Miwhah6VFXFPyicU4mpzH2yHaO6e5nt1xv0eC/yNrvzUJCCgqejJ1FTo2Q4oxQkgJAAQvwTWVnw5ZfGYlzNmhmvwiNzhyeKGBvQrwvmz059+PPCdU5cSuH4pWSir909i57GoKfJ9Uv0ij5Kz4tHefDiMZwz0lg78U2Gf/1vS3wyIaqUb38/z783R9LZy5mQl3uZ7QuLDsNv+d3XLu8cuxNfb18L9bD6Kc011Kqc+iRE1WFjA9OmmW+zti48NuDpCUFBaAMC6A50b5q/NDQtM4foqze5mZlDepaem1k5bPlhF4N+WYRrWhKN0q5R/+YNrNTCcy5Sbe6zzOcSoooZ3Mmd+aEnOXzxBuevpNG0YW3TvpLWn6moOjU1gQQQQpREQAAMHlzisYHatla09TBPgGP/2FBaBM3EgzjTEtCCDCjE4on3E6Ms8hGEqGpc6tjRp0VDdpy8zPrDsczo38q0r6T1ZyqiTk1NIXkghCgprdaYxnHkSOPPUk4s8H3Yhv9z+gjArIR4wedvO3+I78M2hY4VoqYa2tmYxC3kcBx6Q37g7ePlg6ejp2nC5O0UFHSOOny8pIS9pUgAIUQ50Wph8NKRBLKOODzM9sXiSSDrGPz9SJnwKEQBfVu74GRvTXxyBvvOXTNt12q0LPJfBFAoiMh7HuQfJBMoLUgCCCHKUUAAPL1+KL09ovBlJyNZiS878fE8z9Prh8pSSyFuY2etZVAH4zBE8KEYs30BrQMIHh6Mh6N5QO7p6ClLOMuBrMIQogLIUkshSi4i5gZDFu/BztqYE6KOnXlOCMlEWXZkFYYQlVzedAohxN118HSiWcNanLtyk83H4nnqAfOcEFqNVpZqVgCLDmEkJSUxevRoHB0dcXZ2Zvz48aSlpd2x/SuvvELLli2xt7fHy8uLKVOmkJycbMluCiGEqMQURSGwiw6A4EMlr9CpN+gJiw5j1bFVhEWHoTeUruiduDOL3oEYPXo08fHxbN26lezsbJ599lkmTpzIypUri2x/6dIlLl26xMKFC2nTpg0XLlzgxRdf5NKlSwQHB1uyq0IIISqxJzt5sGDLSQ5GXy+UE6IoUiPD8iw2ByIyMpI2bdpw8OBBunbtCkBoaCiPPfYYsbGxuLu7l+h11q1bx9NPP83Nmzexsioc72RmZpKZmWl6npKSgk6nkzkQQghRzTy37CA7Tl7mxT7NmDWgVbHtpEbGvSvNHAiLDWHs27cPZ2dnU/AA0K9fPzQaDX/88UeJXyfvQxQVPADMmzcPJycn00On0/3jvgshhKh8hnc1/n5ffziWbH1+Fle9HsLCYNUq2L5Dz5TNrxYKHgDTtmmh02Q4owxYLIBISEjAxcW8hruVlRX16tUjISGhRK9x9epVPvjgAyZOnFhsm9mzZ5OcnGx6xMTEFNtWCCFE1dW3tQsNattwJTWTnScvA8YKtN7e4OcHo0ZBv75a4t7fDSeeLPI1VFRiUmIIvxhejj2vnkodQMyaNQtFUe74OHny5D/uWEpKCgMHDqRNmza8++67xbaztbXF0dHR7CGEEKL6sdZqTJkp1/4ZQ0gIBAaal6gBIMUD1gYXG0SA1MgoC6WeRPnaa68xbty4O7Zp2rQpjRo14vLly2bbc3JySEpKolGjRnc8PjU1FX9/f+rUqcOGDRuwti5cB14IIUTNM6yrjq9/P8+Ok5fZsj4DVbUropUGMEBoELTaBJrCReukRsY/V+oAomHDhjRs2PCu7Xr06MGNGzc4dOgQXbp0AWDHjh0YDAa6d+9e7HEpKSn0798fW1tbfvrpJ+zsivrLIYQQoiZq7lKbB7zrcjD6Oin1Y+FM82JaaiDFCy74QJNdpq0KCp6OnlIjowxYbA5E69at8ff3Z8KECRw4cIA9e/YwefJkRowYYVqBERcXR6tWrThw4ABgDB4effRRbt68yffff09KSgoJCQkkJCSg18uEFyGEEPmTKWu3j4EiJkuaScu/0yA1MsqWRRNJrVixglatWtG3b18ee+wxevfuzTfffGPan52dzalTp0hPTwfg8OHD/PHHHxw7dozmzZvj5uZmesjkSCGEEAAD27thp7XCum46trqkOzeunT/XQWpklC2phSGEEKLKmbX+GKsPXuTmcQ+u/tKx0H5FAQ9PlWU7d3E5XWpklJTUwhBCCFGtjeymY/XBi9i3iEdjdz+GjPzJ9kpude9FQQp9m/kWOlaKb5UNCSCEEEJUOe09nWjVqA4nE1Lx6BVLzPYmpn2enhAUBAFFjFRIiuuyY9E5EEIIIYQlKIrC6AcbA9BiwEV27FBZuRJ27oSoqOKDh8C1gWbBA0BcShyBawMJiQwpj65XGxJACCGEqJKGdHTHwUbL2Stp2CXtYCSr8CUMLYVX7ekNeqaGTpUU12VIAgghhBBVUh07a4Y4ZQHw3x9/M+ay9vMz5rYOMb+bEH4xvNCdh4IkxXXpSQAhhBCiagoJ4emPpgGwpUVPLtdyNm6Pi4PAQPTBIWZFtjDc/ZK36dQmi3W3upEAQgghRNWj18PUqbS5HEXnuEhytFasbf+ocZ+qoqqQMGIaff30jBoF/xrfF4Ki71gfAyBof5DMhSghCSCEEEJUPeHhpipaTx/ZDMCqDv7oFeNlTUHFQx+DDwWGJEpQZAtkLkRJSQAhhBCi6onPzzD52MndON9KIc7JhbCmXcyauVGw6mbuJS806I7DGTIXomQkgBBCCFH1uOXXuLDTZzPs2DYA/tvpMbNm8dxedbNAka07kHLfdycBhBBCiKrHx8eYMSo37eSoiFAAwpp2IcbJFQMKF9ERTjGBQtqdy3lLue+7kwBCCCFE1aPVwqJFxv9WFJpcv4RP1GFURcN/cu9CTCMIA8WkqK6dUORmBQWdo07KfZeABBBCCCGqpoAACA4GDw8Axh76BYBVHQYQYB3MBgqno1QUqN8oHRqHm8p7m/ZJue9SkQBCCCFE1RUQANHRsHMnD78/jcYOCql2Duxo29lUVCtP3vNvFjuwfsRaPBw9zPZLue/SkXLeQgghqo1le6J49+cTNLSrRcL3fYiNzY8idDrzIltSlbMwKecthBCiRhrWVcfHW09zJeMmP2y5gtVlF+LjjYs2fHyMUyf0emMaifh4LW5uvgzP3S5KRwIIIYQQ1UYtWytGPKDj2/Aolu2N4r/Pu5jtDwmBqVNNOagA42KORYuKruApiidzIIQQQlQrz/TwRqPA7rNXOZWQatoeEgKBgebBA5hKZ9xef0vchQQQQgghqhVdPQf82zYC4IfgvbBqFfrtYbw6RU9Rs/5U1fiYOtU4vCFKRgIIIYQQ1c5zORcBCIlO59r4F9H282N3nDdPUvxththY+Pe/y6uHVZ8EEEIIIaqXkBC6PDuU9vGnybKyYUVuYikP4ggm8I5BxJw5MpRRUhJACCGEqD5yy3wrqsr4g5sAWN75cW5Z2aLBOH4RxDQ0FD9WMW2aDGWUhAQQQgghqo8CZb4HngzH80YC12o5s6b9IwBoUPHitjLft4mJMb6MuDMJIIQQQlQfBcp8W6kGXvxjPQBfdx9KliY/c4F5me87vgxgTDoVFh3GqmOrCIsOQ2+QWxQSQAghhKg+3MyraAYe24ZL6jXiHRuyoa2faXvhMt/Fv0xIZAjei7zxW+7HqJBR+C33w3uRNyGRNXuyhAQQQgghqo/bynzb6bOZcHAjAF91H0aORkuctvgy34piTHntk7s7JDKEwLWBxKaYJ4+IS4kjcG1gjQ4iJIAQQghRfdxW5htgVMT/cL6VQnQ9dza37EXM9CBURVtssa2goNyU1wY9U0OnolI4eUTetmmh02rscIYEEEIIIaqX28p818rO4Lk/fwLgy6dep/uHTxbcbeLpaTwsL6V1+MXwQnceClJRiUmJIfxizZxxKbUwhBBCVD8BATB4cF7VLMY2aMQ3uzM5eQu2R14mIMC14G6zYlt54lPvPNGytO2qGwkghBBCVE9aLfj6AuAEPJ1zkiW7zvHFzjP0TTyBNiEBXzc3iivH6VbnzhMtS9uuupEhDCGEEDXC+N5NsNOoRMQks33CLBg1Cvz8wNu7yPSTPl4+eDp6oqAUfrFcDR0a0tOzpwV7XXlJACGEEKJGaLj1V8btNeaFWNDnGfRK7iWwmHKcWo2WRf7GCZnFBRFX0q/g+aknwceDLdfxSkoCCCGEENVfborrl/avwzEjjVMNvdnYxte4L69EZxE5rANaBxA8PBgPx9tmXBZwJf0Kw4KHMXPrTMv0vZKSAEIIIUT1l5vi2inzJi/tXwfAJz6jydTmTgVU1WJzWAe0DuDcK+do6NDwjm+xYO8C1h1fV+Zdr6wkgBBCCFH9FchNPe7QL7imXiPOyZUVHR8zb7dpU5GHh1/Yy5XjbeDYCIjqA4aiL5+TNk+qMXkhLBpAJCUlMXr0aBwdHXF2dmb8+PGkpaWV6FhVVRkwYACKorBx40ZLdlMIIUR1VyA3tX1OJtP2rATgi55PkWpjn98uKKjQXIiQEBje6wFYHgbrVxl/BkXDiScLvc2V9Cs1Ji+ERQOI0aNHc/z4cbZu3covv/zC77//zsSJE0t0bFBQEMrtacKEEEKIe3Fbiuthf22l6bVYkhyc+O6B2wKBAnMhQkKM8yuTEh3M26R4wNrgIoOImpIXwmIBRGRkJKGhoXz33Xd0796d3r178/nnn7N69WouXbp0x2MjIiL4+OOPWbp0qaW6J4QQoibJS3GdO2HSSjXwevh/APjugSFccXDOb5s7FyJ33mXuIbd/oc29fIYGFRrOqCl5ISwWQOzbtw9nZ2e6du1q2tavXz80Gg1//PFHscelp6czatQoFi9eTKNGje76PpmZmaSkpJg9hBBCiEICAox3F3INOLWHDpdOc9PWgXl+z5m3jY/Pm3d5BxpI8YIL+YW5dI46fLyKLtRV3VgsgEhISMDFxcVsm5WVFfXq1SMhIaHY41599VV69uzJ4MGDS/Q+8+bNw8nJyfTQ6XT/qN9CCCGqsQLXFgV4b9sSFNVASNuH2adrl9/Oza3gvMs7S3XPfT2FIP8gtJrCWS2ro1IHELNmzUJRlDs+Tp48eU+d+emnn9ixYwdBQUElPmb27NkkJyebHjExMff03kIIIWqA2+ZCdIw/zaiIUADefvQlsrTWpnrebiUdifh1MfWjJhI8PJiA1gEW6njlU+paGK+99hrjxo27Y5umTZvSqFEjLl++bLY9JyeHpKSkYocmduzYwblz53B2djbbPnToUHx8fAgLCyt0jK2tLba2tqX5CEIIIWqqvLkQgYHGIEJVmblrOVta9OBsAy++fWAIk2aMAK3WFGvExeXnmipSZl2uLV/CfheFgI/K7ZNUOEVV73ha7llkZCRt2rThzz//pEuXLgD89ttv+Pv7Exsbi7u7e6FjEhISuHr1qtm2du3asWjRIgYNGkSTJk3u+r4pKSk4OTmRnJyMo6Nj2XwYIYQQ1UtIiHGGZO4kh5D7/Zj++GvYaVS2vv4wunoOpmaBgXcJIApYuxaGDbNUpy2vNNdQiwUQAAMGDCAxMZElS5aQnZ3Ns88+S9euXVm50rj+Ni4ujr59+/Ljjz/SrVu3ojuoKGzYsIEhQ4aU6D0lgBBCCFEier2pnrfaqBEjT9uxPyqJfq1d+G7sA6ZmISHwwgtw2/fbIjVsaMxZVURxzzLvc6jSkMh6nkzu2wJrbdlMaSzNNdSieSBWrFhBq1at6Nu3L4899hi9e/fmm2++Me3Pzs7m1KlTpKenW7IbQgghRGF55b5HjkTx8+NfT7bFWquwLfIyoX/nz6AMCDDmlyqJK1eKzIZdNkJCjJVD/fy4POFlZu+9zKId51jx9U8WesM7K/UciNKoV6+e6W5DUby9vbnbDRAL3iARQgghTJq71GHiQ01ZvPMcs0OO0VFXl0ZOdgB4FF9Lq5ASr94ojQJjKSow238K1x2caJ14nlELX4NGq42RTjmSWhhCCCFEril97+N+d0eup2czbc0R9Abjl1gfH+PwREmUePVGSZlntGJt+0fY3rwbNjnZfPrLx9gYcoqsJGppEkAIIYQQuWyttHw+shMONlr2n0/iy51nAeNox5df3v343BWgZatARqsYJ1fef3gCAK+F/4dWVy/csZKoJUkAIYQQQhTQtGFtPhjcFoCg7Wf4MzoJ9HoCG4Sx4vFV9CEMDYW/7SuKca5EmU+gzB0T0SsaXhv4KjdtHXgg5jjPH9xYZLvyIgGEEEIIcZuAzh4M6eiO3qAydeleklu0AT8/Rv0yijD8uKjx5knyq3bqdBAcbKFpCLljIku7DuaAri0OWbf4+NdP0KqGItuVFwkghBBCiNsoisIHQ9rS2FYlLkvhtY5PkaPkXzLd1TjWE8jv00LYuROioiw4h9HHhwOdfFnw0DMAvL3jO7ySEwt21kJjJ3cmAYQQQghRhDrWGj7bOB+bnCy23dedt/pPIm9doKKqKAr4rJ+Gr4/ecnkfgLPX0pkw4DWyrKwZcGoPI45uyd+Zm5LbMmMndyYBhBBCCFGU8HA6/LWHz35agMagZ02H/qa7AED+5MVFiyy2AuJyagZjlx4kWa/QqZaBT4+sNi8s7ulpwbGTO5MAQgghhChK7qRE/zP7mLtlMQBf9hjOd11vqxb92mvg6mrM1fBP6fUQFgarVnFz206e++EgcTdu4V3fge9efRS7c2dg505YuRLLj53cmUUTSQkhhBBVVoFJiSP++o1rDk4s6DOWf/WdQN1bqQw9viO/7bVrMHQorF9/7xf0devg5Zfh6lWyNVomB/wffzd7gHpWKsue7Ub92rmFI3197/0zlSG5AyGEEEIU5bbS3y/vX8f43KWTrw+cxuIHh1EoV/LUqaUfztDrYcQIGD4crl7lul0dnhn+PjubPYBddgbf//A63r9vufvrlDMJIIQQQoii5JX+zqUAb+34njGHf0FVNCzoM5aXhswmzcY+/5jY2JIldMobqnj1VahXD9asAeBkg8Y8MfZT9jXuQK3MdL7aOI9Ol05VSKbJu5EAQgghhChOQIBxkmK9egBoUPlg6xLmhX6OTU42oS17MWTMx5yrV6BYxt0SOhUoikVQEKSkABB6Xw8CxiwkxrkRja9fYsN/X8fv/CHjMRWQafJuZA6EEEIIcScBAeDkBP36mTaNPLqFlleieWnIm5xt4MXgZz5l4oEQxh76Gac7JXTKLYqlqqppNcVFJ1cW9xjOmg79AegdfYQvNn2Ic0aa+bHlnGnybhS1mpW7LE0tcyGEEKJE9HrjXYPcmhR5LtdyZvLgWRzQGVNf18m6xTOP3M/4h5pRr5ZNka+hxsaiAGfrefJlj+FsatMHvcaYw+HZPzfx1o7vsbo9yyQYV11YeAJlaa6hEkAIIYQQJRESYlxpcRu9omFzy1580fMpTjX0BsDeWksnL2dauNahuUtt7nOpza1DRzj70Recq+/J2fpe/OnZGjU3u2Wf838yZe9qusSdLPq9GzY03oGwcLIoCSAkgBBCCGEJISEwcaJx2WZB9etj+PobtrbqyRc7znIsLrlEL/fI6X1M3reWDgln7txw7VoYNuweO11yEkBIACGEEMJS8lZQhIUZn/v6Gh+5dwdUVeXvuBROJqRw5nIaZxJTOXslDVIyaHt8N82vxdD8agztEs7Q9Pqlu7/fjBnw0UcW+jDmJICQAEIIIUQls3qFnl5Pe+NBHJrCGSQKScaR49O/p+fHgeXQO6PSXENlGacQQghRDhp5aJmKMa+EwbyihZmr1Odt3qMeSfT6JJCZM8urh6UjAYQQQghRDnr20vOT04MEspY4PMz2JdKQT5iGLztxJZF/8Q4GjEMiCxYYs1xXNjKEIYQQQpSDsOgw/N74DNYGo0GPD3twI5543AjHxxQwFKWcFmGU6hoqiaSEEEKIchCfGg9tNsDwQAy/LGFXum+Jj71yxZiIspLU0QJkCEMIIYQoF251cjNUttkA0z3A4TKUYDJlnkqWiFICCCGEEKI8+Hj54OnoiYICVjnw+IsYA4iSBRFubvkrSFetMv6syPpaEkAIIYQQ5UCr0bLIP7+6Z95wBvbXij8ol05nHMbIq8E1apTxp7e3MbdVRZAAQgghhCgnAa0DCB4eTAOHBsYNbTbADFe4fyVF34lQURQYMQKeeqpQKQ7i4iAwsGKCCAkghBBCiHIU0DqAoP5B+Rs0Bhg2GgKH5c6LyFe7wQ3WrjUOWRS1ZjJv27Rp5T+cIaswhBBCiHLm4ehReGPb9cY7Ehd8IM0NaseT1jic4OjFxMa+WOxrqSrExJT/Kg0JIIQQQohyljehMi4lDrXg0IXGAE12mbVds38XUHwAkae8V2nIEIYQQghRzvImVKolWYFRu2SRgZvbP+xUKUkAIYQQQlSAgNYBTOs+7e4NG4eDYwwoRQcbimJcpeHjU7b9uxsJIIQQQogKMrjV4Ls30hjAfypgDBYKynseFGT5NNeFulW+byeEEEKIPD5ePjR0aHj3hm028N6Xx/G4be6lpycEB0NAgGX6dycSQAghhBAVRKvR8uVjX961nc5Rx1sTWxMdDTt3wsqVxp9RURUTPICswhBCCCEqVOD9gcy4NIMFexcUuV9BIcg/CK3GOEZRWQpqWewORFJSEqNHj8bR0RFnZ2fGjx9PWlraXY/bt28fDz/8MLVq1cLR0ZGHHnqIW7duWaqbQgghRIX76JGPWBu4Nj9DZS6do47g4cEEtK6g2wx3oKhqUbmt/rkBAwYQHx/P119/TXZ2Ns8++ywPPPAAK1euLPaYffv24e/vz+zZsxk0aBBWVlYcPXqUwYMHY2trW6L3LU0tcyGEEKIy0Rv0hF8MJz41Hrc6bvh4+ZjuPJSH0lxDLRJAREZG0qZNGw4ePEjXrl0BCA0N5bHHHiM2NhZ3d/cij3vwwQd55JFH+OCDD+75vSWAEEIIIe5Naa6hFhnC2LdvH87OzqbgAaBfv35oNBr++OOPIo+5fPkyf/zxBy4uLvTs2RNXV1f69OnD7t277/hemZmZpKSkmD2EEEIIYVkWCSASEhJwcXEx22ZlZUW9evVISEgo8pjz588D8O677zJhwgRCQ0Pp3Lkzffv25cyZM8W+17x583BycjI9dDpd2X0QIYQQQhSpVAHErFmzUBTljo+TJ0/eU0cMBgMAL7zwAs8++yydOnXi008/pWXLlixdurTY42bPnk1ycrLpERMTc0/vL4QQQoiSK9Uyztdee41x48bdsU3Tpk1p1KgRly+blyTNyckhKSmJRo0aFXmcW24S7zZt2phtb926NRcvXiz2/WxtbUs8wVIIIYQQZaNUAUTDhg1p2PDuGbN69OjBjRs3OHToEF26dAFgx44dGAwGunfvXuQx3t7euLu7c+rUKbPtp0+fZsCAAaXpphBCCFEtVfQqjYIskkiqdevW+Pv7M2HCBJYsWUJ2djaTJ09mxIgRphUYcXFx9O3blx9//JFu3bqhKAozZsxgzpw5dOjQgY4dO7J8+XJOnjxJcHCwJbophBBCVBkhkSFMDZ1KbEqsaZunoyeL/BdVSJ4Ii2WiXLFiBZMnT6Zv375oNBqGDh3KZ599ZtqfnZ3NqVOnSE9PN22bNm0aGRkZvPrqqyQlJdGhQwe2bt1Ks2bNLNVNIYQQotILiQwhcG1gofLfcSlxBK4NrJBkUxZLJFVRJA+EEEKI6kRv0OO9yNvszkNBCgqejp5ETY36x8MZFZ4HQgghhBBlI/xieLHBA4CKSkxKDOEXw8uxVxJACCGEEJVafGp8mbYrKxJACCGEEJWYWx23Mm1XViSAEEIIISoxHy8fPB09UVCK3K+goHPU4ePlU679kgBCCCGEqMS0Gi2L/BcBFAoi8p4H+QeVez4ICSCEEEKISi6gdQDBw4PxcPQw2+7p6FkhSzhBlnEKIYQQVYalM1GW5hpqsURSQgghhChbWo0WX2/fiu4GIEMYQgghhLgHEkAIIYQQotQkgBBCCCFEqUkAIYQQQohSkwBCCCGEEKUmAYQQQgghSq3aLePMS2uRkpJSwT0RQgghqpa8a2dJUkRVuwAiNTUVAJ1OV8E9EUIIIaqm1NRUnJyc7tim2mWiNBgMXLp0iTp16qAoRRceKa2UlBR0Oh0xMTGS3bKMyDkte3JOy5acz7In57RsWeJ8qqpKamoq7u7uaDR3nuVQ7e5AaDQaPD09LfLajo6O8pe+jMk5LXtyTsuWnM+yJ+e0bJX1+bzbnYc8MolSCCGEEKUmAYQQQgghSk0CiBKwtbVlzpw52NraVnRXqg05p2VPzmnZkvNZ9uSclq2KPp/VbhKlEEIIISxP7kAIIYQQotQkgBBCCCFEqUkAIYQQQohSkwBCCCGEEKUmAYQQQgghSk0CiFyLFy/G29sbOzs7unfvzoEDB+7Yft26dbRq1Qo7OzvatWvH5s2by6mnVUdpzum3336Lj48PdevWpW7duvTr1++ufwY1UWn/nuZZvXo1iqIwZMgQy3awiint+bxx4waTJk3Czc0NW1tbWrRoIf/2b1PacxoUFETLli2xt7dHp9Px6quvkpGRUU69rdx+//13Bg0ahLu7O4qisHHjxrseExYWRufOnbG1taV58+YsW7bMch1Uhbp69WrVxsZGXbp0qXr8+HF1woQJqrOzs5qYmFhk+z179qharVb96KOP1BMnTqj/93//p1pbW6vHjh0r555XXqU9p6NGjVIXL16sHjlyRI2MjFTHjRunOjk5qbGxseXc88qrtOc0T1RUlOrh4aH6+PiogwcPLp/OVgGlPZ+ZmZlq165d1ccee0zdvXu3GhUVpYaFhakRERHl3PPKq7TndMWKFaqtra26YsUKNSoqSt2yZYvq5uamvvrqq+Xc88pp8+bN6ltvvaWGhISogLphw4Y7tj9//rzq4OCgTp8+XT1x4oT6+eefq1qtVg0NDbVI/ySAUFW1W7du6qRJk0zP9Xq96u7urs6bN6/I9sOHD1cHDhxotq179+7qCy+8YNF+ViWlPae3y8nJUevUqaMuX77cUl2scu7lnObk5Kg9e/ZUv/vuO3Xs2LESQBRQ2vP51VdfqU2bNlWzsrLKq4tVTmnP6aRJk9SHH37YbNv06dPVXr16WbSfVVFJAoiZM2eq999/v9m2p556Su3fv79F+lTjhzCysrI4dOgQ/fr1M23TaDT069ePffv2FXnMvn37zNoD9O/fv9j2Nc29nNPbpaenk52dTb169SzVzSrlXs/p+++/j4uLC+PHjy+PblYZ93I+f/rpJ3r06MGkSZNwdXWlbdu2zJ07F71eX17drtTu5Zz27NmTQ4cOmYY5zp8/z+bNm3nsscfKpc/VTXlfm6pdNc7Sunr1Knq9HldXV7Ptrq6unDx5sshjEhISimyfkJBgsX5WJfdyTm/3xhtv4O7uXugfQ011L+d09+7dfP/990RERJRDD6uWezmf58+fZ8eOHYwePZrNmzdz9uxZXn75ZbKzs5kzZ055dLtSu5dzOmrUKK5evUrv3r1RVZWcnBxefPFF3nzzzfLocrVT3LUpJSWFW7duYW9vX6bvV+PvQIjKZ/78+axevZoNGzZgZ2dX0d2pklJTUxkzZgzffvstDRo0qOjuVAsGgwEXFxe++eYbunTpwlNPPcVbb73FkiVLKrprVVZYWBhz587lyy+/5PDhw4SEhPDrr7/ywQcfVHTXRAnU+DsQDRo0QKvVkpiYaLY9MTGRRo0aFXlMo0aNStW+prmXc5pn4cKFzJ8/n23bttG+fXtLdrNKKe05PXfuHNHR0QwaNMi0zWAwAGBlZcWpU6do1qyZZTtdid3L31E3Nzesra3RarWmba1btyYhIYGsrCxsbGws2ufK7l7O6dtvv82YMWN4/vnnAWjXrh03b95k4sSJvPXWW2g08h23NIq7Njk6Opb53QeQOxDY2NjQpUsXtm/fbtpmMBjYvn07PXr0KPKYHj16mLUH2Lp1a7Hta5p7OacAH330ER988AGhoaF07dq1PLpaZZT2nLZq1Ypjx44RERFhejzxxBP4+fkRERGBTqcrz+5XOvfyd7RXr16cPXvWFIgBnD59Gjc3txofPMC9ndP09PRCQUJegKZKncdSK/drk0WmZlYxq1evVm1tbdVly5apJ06cUCdOnKg6OzurCQkJqqqq6pgxY9RZs2aZ2u/Zs0e1srJSFy5cqEZGRqpz5syRZZy3Ke05nT9/vmpjY6MGBwer8fHxpkdqampFfYRKp7Tn9HayCsNcac/nxYsX1Tp16qiTJ09WT506pf7yyy+qi4uL+q9//auiPkKlU9pzOmfOHLVOnTrqqlWr1PPnz6u//fab2qxZM3X48OEV9REqldTUVPXIkSPqkSNHVED95JNP1CNHjqgXLlxQVVVVZ82apY4ZM8bUPm8Z54wZM9TIyEh18eLFsoyzPHz++eeql5eXamNjo3br1k3dv3+/aV+fPn3UsWPHmrVfu3at2qJFC9XGxka9//771V9//bWce1z5leacNm7cWAUKPebMmVP+Ha/ESvv3tCAJIAor7fncu3ev2r17d9XW1lZt2rSp+u9//1vNyckp515XbqU5p9nZ2eq7776rNmvWTLWzs1N1Op368ssvq9evXy//jldCO3fuLPL3Yt45HDt2rNqnT59Cx3Ts2FG1sbFRmzZtqv7www8W65+iqnKfSAghhBClU+PnQAghhBCi9CSAEEIIIUSpSQAhhBBCiFKTAEIIIYQQpSYBhBBCCCFKTQIIIYQQQpSaBBBCCCGEKDUJIIQQQghRahJACCGEEKLUJIAQQgghRKlJACGEEEKIUvt/Qsc0w6M+4fsAAAAASUVORK5CYII=", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], + "outputs": [], "source": [ "plt.plot(x_norm.detach().numpy(), y_norm.detach().numpy())\n", - "plt.scatter(x_test_torch.detach().numpy(), y0_test_torch.detach().numpy(), color='green', label='Model 0')\n", - "plt.scatter(x_test_torch.detach().numpy(), y1_test_torch.detach().numpy(), color='blue', label='Model 1')\n", + "plt.scatter(x_test_torch.detach().numpy(), y0_test_torch.detach().numpy(), color='green', marker='*', label='Model 0')\n", + "plt.scatter(x_test_torch.detach().numpy(), y1_test_torch.detach().numpy(), color='blue', marker='v', label='Model 1')\n", "plt.scatter(x_test_torch.detach().numpy(), y2_test_torch.detach().numpy(), color='red', label='Model 2')\n", "plt.legend()\n", "plt.show()" @@ -1480,42 +1139,51 @@ }, { "cell_type": "markdown", - "id": "31ca70bb", + "id": "703024a9", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ - "## Play with the notebook!\n", - "Some ideas:\n", - "- Change the number of epochs in `Section 5` to 5000 and re-train the models. What happens?\n", - "- Change the random seed in the `Reproducibility` cell at the very top. How do the results change?\n", - "- Change the optimizer in `Section 4` from `Adam` to `SGD` and re-train the models. What happens?\n", - "- [**if time allows, takes several minutes**] Change the epochs in `Section 5` to 1000000. What happens?\n", - "- Go back to 1000 epochs and the Adam optimizer. Change the learning rate in `Section 4` to 0.05. How do the results change? what does it tell us about our previous value?\n", - "- Change the learning rate to 0.5. What happens now?\n" + "

Comment on the NN predictions

\n", + "

$\\implies$ Why does the prediction of model 0 have that particular shape?

\n", + "

$\\implies$ Which activation function would be more appropriate to fit this function, the one from model 0 or model 1?

\n", + "

$\\implies$ Which NN gets the best prediction and why?

" ] }, { "cell_type": "markdown", - "id": "703024a9", + "id": "412c4eb3", "metadata": { "slideshow": { - "slide_type": "-" + "slide_type": "slide" } }, "source": [ - "

Comment on the results

" + "

Bonus

\n", + "

$\\implies$ Change the seed at the top of the notebook. How do the predictions change?

\n", + "

$\\implies$ Change the optimizer in Section 4 from Adam to SGD and re-train the models. What happens? How did the loss curves change? Did the NNs learn? Change the number of epochs and try to make it learn.

" ] }, { - "cell_type": "code", - "execution_count": null, - "id": "dac1d9a3", - "metadata": {}, - "outputs": [], - "source": [] + "cell_type": "markdown", + "id": "31ca70bb", + "metadata": { + "slideshow": { + "slide_type": "slide" + } + }, + "source": [ + "## Play with the notebook!\n", + "Some ideas:\n", + "- Change the number of epochs in `Section 5` to 5000 and re-train the models. What happens?\n", + "- Change the random seed in the `Reproducibility` cell at the very top. How do the results change?\n", + "- Change the optimizer in `Section 4` from `Adam` to `SGD` and re-train the models. What happens?\n", + "- [**if time allows, takes several minutes**] Change the epochs in `Section 5` to 1000000. What happens?\n", + "- Go back to 1000 epochs and the Adam optimizer. Change the learning rate in `Section 4` to 0.05. How do the results change? what does it tell us about our previous value?\n", + "- Change the learning rate to 0.5. What happens now?\n" + ] } ], "metadata": {