Last active
May 21, 2023 11:49
-
-
Save jmcalvomartin/db432899d0ff67ad1069f4fb3db03f23 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
{ | |
"cells": [ | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"## Funciones de activación\n", | |
"### Redes neuronales" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 2, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"import numpy as np\n", | |
"import math, random\n", | |
"import matplotlib.pyplot as plt\n", | |
"%matplotlib inline \n", | |
"from IPython.display import display, Math, Latex" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"#### Función Sigmoide\n", | |
"La razón principal por la que usamos la función sigmoide es porque existe entre (0 a 1). Por lo tanto, se usa especialmente para modelos en los que tenemos que predecir la probabilidad como un resultado. Dado que la probabilidad de cualquier cosa existe solo entre el rango de 0 y 1" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 3, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"#Creamos la función sigmoide con lambda , donde lambda[0] es la sigmoide y lambda[1] es su derivada\n", | |
"sigm = (lambda x:1/(1+np.e**(-x)),lambda x:x * (1-x))" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 4, | |
"metadata": {}, | |
"outputs": [ | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle sigmoide(x) = \\frac{1}{1+e^{-x}} $" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle sigmoide'(x) = x{(1-x)}$" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"image/png": "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\n", | |
"text/plain": [ | |
"<Figure size 1080x720 with 2 Axes>" | |
] | |
}, | |
"metadata": { | |
"needs_background": "light" | |
}, | |
"output_type": "display_data" | |
} | |
], | |
"source": [ | |
"#Latex\n", | |
"display(Math(r'sigmoide(x) = \\frac{1}{1+e^{-x}} '))\n", | |
"display(Math(r\"$$sigmoide'(x) = x{(1-x)}\"))\n", | |
"\n", | |
"v = np.linspace(-5,5,100)\n", | |
"plt.figure(1,figsize=(15,10))\n", | |
"plt.subplot(221)\n", | |
"plt.plot(v,sigm[0](v))\n", | |
"plt.title(\"Funcion Sigmoide\")\n", | |
"plt.subplot(222)\n", | |
"plt.plot(v,sigm[1](v), \"--\", color=\"red\")\n", | |
"plt.title(\"Derivada función Sigmoide\")\n", | |
"plt.show()" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"#### Función Tangente hiperbólica o Gaussiana\n", | |
"Es una función similar a la Sigmoide pero produce salidas en escala de [-1, +1]. Además, es una función continua. En otras palabras, la función produce resultados para cada valor de x." | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 5, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"#Tangente Hiperbólica y su derivada\n", | |
"cosh = (lambda x: (np.e**(x) + np.e**(-x))/2) #Saco la función de coseno hiperbólico para la derivada de la tangh\n", | |
"tanh = (lambda x: (np.e**(x) - np.e**(-x))/(np.e**(x) + np.e**(-x)), lambda x: 1/cosh(x)**2)" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 6, | |
"metadata": {}, | |
"outputs": [ | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle cosh(x) = \\frac{e^{-x} + e^{-x}}{2}$" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle tanh(x) = \\frac{e^{-x} - e^{-x}}{e^{-x} + e^{-x}} $" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle 'tanh(x) = \\frac{1}{\\cosh^2{x}} $" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"image/png": "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\n", | |
"text/plain": [ | |
"<Figure size 1080x720 with 3 Axes>" | |
] | |
}, | |
"metadata": { | |
"needs_background": "light" | |
}, | |
"output_type": "display_data" | |
} | |
], | |
"source": [ | |
"#Latex\n", | |
"display(Math(r'cosh(x) = \\frac{e^{-x} + e^{-x}}{2}'))\n", | |
"display(Math(r'tanh(x) = \\frac{e^{-x} - e^{-x}}{e^{-x} + e^{-x}} '))\n", | |
"display(Math(r\"'tanh(x) = \\frac{1}{\\cosh^2{x}} \"))\n", | |
"\n", | |
"v = np.linspace(-5,5,100)\n", | |
"plt.figure(1,figsize=(15,10))\n", | |
"plt.subplot(331)\n", | |
"plt.plot(v,cosh(v))\n", | |
"plt.title(\"Funcion Coseno hiperbólico\")\n", | |
"plt.subplot(332)\n", | |
"plt.plot(v,tanh[0](v))\n", | |
"plt.title(\"Función Tangente hiperbólica\")\n", | |
"plt.subplot(333)\n", | |
"plt.plot(v,tanh[1](v), \"--\", color=\"red\")\n", | |
"plt.title(\"Derivada función Tangente hiperbólica\")\n", | |
"plt.savefig('images\\sigmoide.png')\n", | |
"plt.show()" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"#### Función RELU (Rectified Lineal Unit)\n", | |
"ReLU es la función de activación más utilizada en el mundo en este momento. Desde entonces, se utiliza en casi todas las redes neuronales convolucionales o el aprendizaje profundo.\n", | |
"Como puedes ver, ReLU está medio rectificado (desde abajo). f(z) es cero cuando z es menor que cero y f(z) es igual a z cuando z es superior o igual a cero.\n", | |
"Es una función usada en las capas ocultas de nuestra red neuronal, NO en las de salida" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 7, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"#Relu Rectified Lineal Unit\n", | |
"relu = (lambda x: np.maximum(0,x), lambda x: 1. * (x > 0))" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 8, | |
"metadata": {}, | |
"outputs": [ | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle relu(x) = \\max(0,x) $" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"text/latex": [ | |
"$\\displaystyle 'relu(x) = 1.(x>0) $" | |
], | |
"text/plain": [ | |
"<IPython.core.display.Math object>" | |
] | |
}, | |
"metadata": {}, | |
"output_type": "display_data" | |
}, | |
{ | |
"data": { | |
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA2AAAAEmCAYAAAAEDPP8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3dd5hU1f3H8c+XZWHpdQHpHUSkiz0aK2LD2AH9aaKGZotdY0yiJpbEaCygUWOUotgbdmOMUVRYehVXOlKkl4Ut5/fHnYXZmYHdhd05d3bfr+fZZ+7MuXP3M3d35sz3lnPNOScAAAAAQPmr4jsAAAAAAFQWFGAAAAAAkCQUYAAAAACQJBRgAAAAAJAkFGAAAAAAkCQUYAAAAACQJBRgwD6Y2VYza+87BwAAZnasmS0op2U7M+u4n88dbmarI31mo7LOFvkdCV+7mWWa2XQz610evxcoDxRgSClmttjMdkQ+5At/mpfX73PO1XbOZZf1cmNex49m9pyZ1Y5qf87MdsW8zhmRtraRjrJqzDKPN7PlCX7XZ2Z2RVm/BgDA3kV9zm8xs41m9qWZDTOz/f7u5Zz7r3OuS1nmPFBmli7pIUmnRPrMn8rj9yR67ZHf/S9Jw51z0/ZnuVF9amFfu9jMbo2ZJ9F3j8cibZeZ2RcJlvt7Mxub4PH9LnRRcVCAIRWdGfmQL/xZ6TvQfjrTOVdbUi9JvSXdFtP+QMzr7Jn8iACAA3Cmc66OpDaS7pN0i6Rn9mdBsRvdQqSppAxJc5L9i51zuc65gc65r8pgcfUjffJ5ku40s5Nj2mO/e4wqg9+JSooCDBVCZOvUSVH3d295itq69X9mttTM1pnZHVHzppnZ7Wb2fWRL5VQzaxVp272lyszqmdnzZrbWzJaY2W8Lt2QWbgEzs7+Y2QYz+8HMTitJdufcj5I+UFCIAQAqGOfcJufcW5IulPR/ZtZdksyseqTfWBo5hG+MmdWItB1vZsvN7BYz+1HSP6OPdDCzW83slejfY2aPmNnfI9OXm9m8SL+WbWa/jpn3JjNbZWYrzeyXMW2nm9k0M9tsZsvM7PeJXpeZdZZUeFjgRjP7NNFRGtFHYhTXX5pZQzP7ZyTXBjN7I3p9RM13cGS5G81sjpmdFdX2nJk9bmbvRl7/12bWoYR/qykKikn6ZJQbCjBUJsdI6iLpREm/M7ODI4//RtLFkgZKqivpl5K2J3j+o5LqSWov6ThJl0q6PKr9cAUdUWNJD0h6xsysuFBm1lLSaZIWlf4lAQBShXPuG0nLJR0beeh+SZ0VfNnvKKmFpN9FPaWZpIYK9qBdFbO4CZIGmlldKdiYKOkCSeMj7WsknaGgX7tc0t/MrE9k3gGSbpR0sqROkk4qumhtU9DH1Zd0uqThZjYowetZKOmQyN36zrkTSrIetO/+8gVJNSPLbSLpb7FPjhx6+LakDyPzXC1pnJlFH6J4saQ/SGqgoH+9tyTBzOwISd1Fn4xyRAGGVPRGZIvXxsItYyX0B+fcDufcDEkzJBUe0neFpN865xa4wIzYY9gjHduFkm5zzm1xzi2W9FdJl0TNtsQ59w/nXL6CY9IPUnBoxr5exxZJyxR0lHfFtN8Y9To3mtm/SvFaAQDhtFJSw0jBcaWk651z651zWyT9SdJFUfMWSLrLObfTObcjeiHOuSWSsiQVFkYnSNrunJscaX/XOfd9pF/7j4JipbDwu0DSP51zs51z2yT9PmbZnznnZjnnCpxzMxUUe8eV2RrYS39pZgcp2CA5zDm3IXKI4X8SPP8ISbUl3eec2+Wc+1TSOwqKrkKvOee+cc7lSRqn4vdorTOzHZK+kvSEpNjvF2/E9MlXlvZFA4UowJCKBjnn6kd+4rbI7cOPUdPbFXx4S1IrSd8X89zGkqpJWhL12BIFWyvjlu+cK9yDVlt7NyhybsDxkrpGfke0v0S9zvrOuf8rJmOepPQEj6dLyi3muQCA5Gghab2kTAV7eqYWfqmX9H7k8UJrnXM5+1jWeO0pOgZrz94vmdlpZjbZzNZHlj1Qe/qZ5go2/hWK7ttkZoeb2b8jh9xvkjRM8X3Ugdhbf9lK0nrn3IZint9c0jLnXEHUY3vtk1W0z9+bxpF5blTQL8f2p4Ni+uR/FLO8uD45sudOok+u9CjAUFFsU9CRFWpWiucuk1TcseHrFHxgtol6rLWkFaX4PQlFtu49J+kvB7iopZIaW9HRFE1B5iV7fRYAICnM7DAFRcIXCvqVHZIOifpSXy8yEEQhV8wiX5Z0fORQ9nMUKcDMrLqkVxX0K02dc/UlTZJUeJjfKgXFTqHWMcsdL+ktSa2cc/UkjYl6bnG2RW73p09epmDvYP1i5lspqZUVHVHygPtk51y+c+6vknIkjTiQZSnok9vGPNZOUr7K4LsDUhsFGCqK6ZIuMrN0M+unYBSjknpa0t1m1skCPSzmOiaRwyQmSrrXzOqYWRsF547FDTG7nx6WdLKZleak3+pmllH4o+C8gq8l3W9mtSMd8E0KtsJNLqOcAIBSMrO6ZnaGpBcljS08vE/SPxScm9UkMl8LMzu1pMt1zq2V9Jmkf0r6wTk3L9JUTVJ1SWsl5UUGuTgl6qkTJV1mZt3MrKbiD4Gvo2BPVI6Z9Vewd600mVZIGmrBIFe/VPEbOQufu0rSe5KeMLMGkT79Zwlm/VpBoXdzZJ7jJZ2pYP2Whfsiy84o4fwW3R9Hnve+pC5mdkkkY0MFh5i+EjksEpUYBRgqijsVfMBvUHDS7fh9z17EQwo6ow8lbVYwRHCNBPNdreADP1vB1svxkp7d/8h7RDqs5xW8jkI3W9FrjqyLedpWBVtPC39OUHCeWhMFJw+vUDDgyMBiDmEBAJSPt6PO9b1DQX8TPXjTLQo+ryeb2WZJHysYLKo0xisYRGN3vxc5n+waBX3bBgUF1FtR7e8p2PD3aeT3fxqzzBGS/hjJ/rvIckrjSgUbAH9SMJjGl6V47iUKjjiZr+D86OtiZ3DO7ZJ0loLzxdYpOGfrUufc/FLm3Jt3Fay36PO83o7pk1+PajtKRfvjHQoOMx0o6deR1zFb0iZJw8soI1KYOVfc3m0AAAAAQFlgDxgAAAAAJAkFGAAAAAAkCQUYAAAAACQJBRgAAAAAJAkFGAAAAAAkSdXyWGjjxo1d27Zty2PRAIAQmTp16jrnXKbvHKmC/hEAKo+99ZHlUoC1bdtWU6ZMKY9FAwBCxMyW+M6QSugfAaDy2FsfySGIAAAAAJAkFGAAAAAAkCQUYAAAAACQJBRgAAAAAJAkFGAAAAAAkCQlKsDMbLGZzTKz6WbG8E0AgArNzJ41szVmNnsv7WZmfzezRWY208z6JDsjACA1lWYP2M+dc72cc/3KLQ0AAOHwnKQB+2g/TVKnyM9VkkYnIRMAoALgEEQAqITy8gt8Rwg159znktbvY5azJT3vApMl1Tezg5KTDgCQykp6IWYn6UMzc5KedM49VY6ZAADlaNvOPF3w5Ff6v6Pa6oJ+rXzHSVUtJC2Lur888tgqP3GAFDF1qnTFFfGPP/ywdNxx0uefS9deG9/+1FPSYYdJ778v3XZbfPvYsdIhh0ivvirdc098++uvS23bSi+8ID30UHz7Bx9ITZpITz4pjRkT3/7FF1KtWsFzX3ghvj0rSzILfverrxZtq1FD+vLLYPq224LXEK1xY+mjj4Lpa68N1kG01q2lN98Mpq+8Uoq9mPvBB0vjxwfTQ4ZIc+cWbe/TR3rmmWB60CBpScy1gY89Vvr734PpU06R1q4t2n7qqdJ99wXTRx8tbd9etP0Xv5DuvDOY7t1bcYYOlW64IXje0UfHt191lTR8ePB7Tzklvv3666VLLw1yDxoU337HHdJ55wWve8iQ+PY//Uk67bTS/e899ljirGWkpAXY0c65lWbWRNJHZjY/snVwNzO7SsFhGGrdunUZxwQAlAXnnG5/fZbmrtqs5vVq+I6TyizBYy7hjPSPgDR5srRtm9S0aVBQxKpRY89tovaMjOC2Zs3E7dWrB7e1ayduT08PbuvUSdxeNfKVuF69xO1VIgeNNWiQuL1Qw4bx7YXZpaDYim1v0GDPdGZmfHvz5nummzSJb2/WbM9006bS1q1F25s23TN90EGSi/moysws+rtqxPQNjRvvmW7ZUsrJKdresOGe6UTrpvD1mSVur1cvuE1LS9xep05wW7Vq4vbatYPbatUSt9esGdxWr17y/73YdVDGzMX+EYp7gtnvJW11zv1lb/P069fPTYmtzgEA3r3w1WLd+eYc3XhKZ406odMBL8/MplbUc4PNrK2kd5xz3RO0PSnpM+fchMj9BZKOd87tcw8Y/SMqrXPOkbKzpRkzfCcBkmZvfWSx54CZWS0zq1M4LekUSQlHhQIAhNf0ZRv1x3fm6oSuTTTi+I6+46S6tyRdGhkN8QhJm4orvoBKLS9vz14moJIryTuhqaTXzaxw/vHOuff3/RQAQJis37ZLI8ZOVdO6GXrogp6qUiXREXQoZGYTJB0vqbGZLZd0l6R0SXLOjZE0SdJASYskbZd0uZ+kQIqgAAN2K/ad4JzLltQzCVkAAOUgv8Dpupema93WXXpl+JGqX7Oa70ih55y7uJh2J2lkkuIAqY8CDNiNYegBoIJ79NPv9PnCtbrrrG7q0bK+7zgAKiMKMGA33gkAUIH9Z+FaPfLJd/pFnxYa3J8R+AB48uijUgHXHwQkCjAAqLBWbtyh616cps5N6ujeQYcqci4vACRf97jBRIFKi0MQAaAC2pVXoJHjs5Sb7zR6aB/VqJbmOxKAyuydd6RPP/WdAggF9oABQAV077tzNW3pRj0xpI/aZ9b2HQdAZfe73wUX8T3hBN9JAO/YAwYAFcxbM1bqX18t0RXHtNPAQw/yHQcAGIQDiEIBBgAVyHert+jWV2fqsLYNdMtpXX3HAYAABRiwGwUYAFQQ23bmafi4LNWslqbHBvdRehof8QBCggIM2I13AgBUAM453fraLGWv3aqxVxyupnUzfEcCgD3y8qT0dN8pgFCgAAOACuD5r5bo7RkrddOpXXRUh8a+4wBAUe+9J9Wo4TsFEAoUYACQ4rKWbtA9787ViV2baPhxHXzHAYB4Xbr4TgCEBicIAEAKW79tl0aNy1Kzehl66IJeqlKFiy0DCKGnn5Y+/9x3CiAUKMAAIEXlFzhd++I0rdu2S6OH9FW9mpxfASCkbr5ZeuUV3ymAUKAAA4AU9fdPvtN/v1unP5x1iLq3qOc7DgDsHaMgArtRgAFACvpswRr9/dPvdG6flrrosFa+4wDAvlGAAbtRgAFAilm+Ybuue2m6ujSto3sGdZcZ530BCDkKMGA3CjAASCE78/I1cvw05ec7jR7aVzWqpfmOBADFowADduOdAAAp5J535mnGso0aM7SP2jWu5TsOAJRMdrZUt67vFEAoUIABQIp4c/oKvTB5ia76WXsN6H6Q7zgAUHJt2/pOAIQGhyACQApYuHqLbn11lvq3baibTuWCpgBSyK5d0p//LH3zje8kQChQgAFAyG3dmadhY6eqVvWqemxwb6Wn8dENIIXk5Ei33y598YXvJEAo0IsDQIg553TLqzO1eN02PXpxbzWpm+E7EgCUTl5ecMsgHIAkCjAACLXnvlysd2eu0k2ndtWRHRr5jgMApUcBBhRBAQYAITV1yQbd++48nXRwUw07rr3vOACwfyjAgCIowAAghNZt3amR47LUvH4N/fWCnlxsGUDqogADiuCdAAAhk1/gdN2L07V++y69Nvwo1auR7jsSAOy/Fi2kdeukmjV9JwFCgQIMAELm4Y8X6otF63T/uYeqe4t6vuMAwIFJS5MacQ4rUIhDEAEgRP49f40e/XSRzu/bUhce1tp3HAA4cKtXB8PQz57tOwkQChRgABASy9Zv13UvTdfBB9XV3YO6+44DAGVj9ergQswLFvhOAoQCBRgAhEBObr5GjMtSQYHT6CF9lJGe5jsSAJQNBuEAiuCdAAAh8Md35mrWik168pK+atu4lu84AFB2KMCAItgDBgCevZa1XOO/Xqpf/6y9Tj2kme84AFC2KMCAIijAAMCj+T9u1u2vz1L/dg1106ldfMcBgLKXnx/cUoABkjgEEQC82ZKTq+Fjs1QnI12PXdxbVdPYJgagAjrmGGnnzmA4egAUYADgg3NON78yU0vXb9f4Kw5Xk7oZviMBQPkwk6pV850CCA02twKAB8988YPem/2jbj61iw5vzwVKAVRgs2dL11wjLV3qOwkQCiUuwMwszcymmdk75RkIACq6KYvX67735uuUbk111c/a+46DvTCzAWa2wMwWmdmtCdrrmdnbZjbDzOaY2eU+cgKhl50tPfqotG6d7yRAKJRmD9i1kuaVVxAAqAzWbd2pkeOz1KJBDf3lgp4yM9+RkICZpUl6XNJpkrpJutjMusXMNlLSXOdcT0nHS/qrmXGcFRCLURCBIkpUgJlZS0mnS3q6fOMAQMWVX+B0zYRp2rg9V6OH9FXdjHTfkbB3/SUtcs5lO+d2SXpR0tkx8zhJdSyoomtLWi8pL7kxgRRAAQYUUdI9YA9LullSQTlmAYAK7W8fLdSX3/+kuwd1V7fmdX3Hwb61kLQs6v7yyGPRHpN0sKSVkmZJutY5Rz8JxKIAA4ootgAzszMkrXHOTS1mvqvMbIqZTVm7dm2ZBQSAiuDT+av12L8X6cJ+rXRBv1a+46B4iY4NdTH3T5U0XVJzSb0kPWZmcZU1/SMqPeek9HSGoQciSrIH7GhJZ5nZYgWHYJxgZmNjZ3LOPeWc6+ec65eZmVnGMQEgdS1bv13XvThd3Q6qqz+cfYjvOCiZ5ZKiK+WWCvZ0Rbtc0msusEjSD5K6xi6I/hGV3pAh0q5dUocOvpMAoVBsAeacu80519I511bSRZI+dc4NLfdkAFAB5OTma/i4qXKSxgztq4x0tgCniG8ldTKzdpGBNS6S9FbMPEslnShJZtZUUhdJ2UlNCQBIOVwHDADK0R/enqPZKzbroQt6qXWjmr7joIScc3mSRkn6QMEIwBOdc3PMbJiZDYvMdreko8xslqRPJN3inGOcbSDWxx9Ll10mbd7sOwkQCqU6G9I595mkz8olCQBUMK9MXa4J3yzTsOM66ORuTX3HQSk55yZJmhTz2Jio6ZWSTkl2LiDlzJ0r/etf0kMP+U4ChAJ7wACgHMxbtVl3vD5Lh7drqBtP6ew7DgD4wyiIQBEUYABQxjbn5Gr42KmqVyNdjw7urappfNQCqMQowIAieCcAQBlyzunml2dq2YYdmnDlEWpSJ8N3JADwiwIMKILNsgBQhp7+7w96f86PunVAV/Vv19B3HADwr3p1qWFDrgMGRFCAAUAZ+eaH9brv/fkacEgzXXFsO99xACAcbrhB+uknCjAgggIMAMrAmi05GjU+S60b1tSD5/eQmfmOBAAAQogCDAAOUF5+ga6ZME2bc3I1emgf1clI9x0JAMLj+eelwYN9pwBCg7MhAeAA/fWjhZqcvV5/Pb+nujar6zsOAITLjBnS22/7TgGEBnvAAOAAfDR3tUZ/9r0u7t9a5/Zt6TsOAIRPXh4jIAJRKMAAYD8t/Wm7fjNxurq3qKu7zuzmOw4AhBMFGFAEBRgA7Iec3HwNHzdVJmn0kL7KSGd0LwBIiAIMKIJ3AwDsh9+/NUdzVm7Ws5f1U6uGNX3HAYDwatBAatPGdwogNCjAAKCUXp6yTC9+u0wjf95BJ3Rt6jsOAITbfff5TgCECocgAkApzFm5Sb99Y7aO6tBIvzm5i+84AAAgxVCAAUAJbdqRqxHjslSvRroeuai30qpwsWUAKNbdd0tXXOE7BRAaHIIIACXgnNNNL8/Qig079OJVRyizTnXfkQAgNUyfLi1c6DsFEBrsAQOAEnjq82x9OHe1bj2tq/q1beg7DgCkDkZBBIqgAAOAYnyd/ZMe+GCBBh7aTL86pp3vOACQWijAgCIowABgH9ZsztGoCdPUumFN3X9uD5lx3hcAlAoFGFAE7wYA2Iu8/AKNmjBNW3Jy9cKv+qtORrrvSACQetq0kXJyfKcAQoMCDAD24sEPF+ibH9broQt6qmuzur7jAEBqeuop3wmAUOEQRABI4IM5P+rJ/2RryOGt9Ys+LX3HAQAAFQQFGADEWLxum258eYZ6tKyn353ZzXccAEhtl18u3XCD7xRAaHAIIgBEycnN1/BxWapipscH91H1qmm+IwFAapsxQ2rRwncKIDQowAAgyp1vzNa8VZv1z8sOU6uGNX3HAYDUxyiIQBEcgggAES99u1QvT12uq0/oqJ93beI7DgBUDBRgQBEUYAAgac7KTbrzzTk6pmNjXXdSZ99xAKDioAADiqAAA1DpbdqRq+Fjs9SwZjU9clEvpVXhYssAUGZ69ZK6dPGdAggNNkcAqNQKCpxumDhDKzfu0Eu/PlKNalf3HQkAKpaJE30nAEKFPWAAKrUnP8/Wx/NW6/aBB6tvmwa+4wAAgAqOAgxApfXV9z/pwQ/m6/QeB+nyo9v6jgMAFdOxx0p33+07BRAaFGAAKqU1m3N09YRpatu4lu4/t4fMOO8LAMrF3LnSmjW+UwChwTlgACqd3PwCjRyfpW078zT+ysNVuzofhQBQbhgFESiCdwOASueB9+fr28Ub9PCFvdS5aR3fcQCgYqMAA4rgEEQAlcr7s1fpH//9QZcc0UaDerfwHQcAKj4KMKAICjAAlcYP67bpppdnqmer+vrtGQf7jgMAlcNJJ0mducA9UKjYzRFmliHpc0nVI/O/4py7q7yDAUBZ2rErX8PHTlVamunxwb1VvWqa70gIOTMbIOkRSWmSnnbO3ZdgnuMlPSwpXdI659xxSQ0JpIJ33/WdAAiVkuwP3inpBOfcVjNLl/SFmb3nnJtcztkAoEw45/TbN2Zrweot+udlh6llg5q+IyHkzCxN0uOSTpa0XNK3ZvaWc25u1Dz1JT0haYBzbqmZNfGTFgCQSoo9BNEFtkbupkd+XLmmAoAy9OK3y/Rq1nJdfUInHd+F78gokf6SFjnnsp1zuyS9KOnsmHkGS3rNObdUkpxzjLMNxNqxQ2rdWhozxncSIDRKdA6YmaWZ2XRJayR95Jz7unxjAUDZmL1ik+56a46O7dRY157YyXccpI4WkpZF3V8eeSxaZ0kNzOwzM5tqZpcmLR2QKvLypGXLpG3bfCcBQqNEBZhzLt8510tSS0n9zax77DxmdpWZTTGzKWvXri3rnABQapu252r4uKlqXKuaHrmot9KqcLFllFiif5bYoz+qSuor6XRJp0q608ziRhqgf0SllpcX3DIKIrBbqUZBdM5tlPSZpAEJ2p5yzvVzzvXLzMwso3gAsH8KCpx+M3G6ftyUo8eG9FHDWtV8R0JqWS6pVdT9lpJWJpjnfefcNufcOgUDVvWMXRD9Iyo1CjAgTrEFmJllRk40lpnVkHSSpPnlHQwADsTo/3yvT+av0W9P76Y+rRv4joPU862kTmbWzsyqSbpI0lsx87wp6Vgzq2pmNSUdLmleknMC4UYBBsQpybvhIEn/iowIVUXSROfcO+UbCwD235ffr9NfP1ygM3s216VHtvEdBynIOZdnZqMkfaBgGPpnnXNzzGxYpH2Mc26emb0vaaakAgVD1c/2lxoIoWrVpPPOkzp08J0ECI1iCzDn3ExJvZOQBQAO2OrNObpmwjS1z6yt+35xqMw47wv7xzk3SdKkmMfGxNx/UNKDycwFpJRGjaSXX/adAggV9gcDqDBy8ws0clyWtu/K14Qr+6hWdT7iAABAuJRqEA4ACLP735uvKUs26L5ze6hT0zq+4wAAFiyQGjSQXn/ddxIgNCjAAFQIk2at0tNf/KBLj2yjs3o29x0HACBJu3ZJGzdKBQW+kwChQQEGIOVlr92qm1+ZqZ6t6uuO0w/2HQcAUIhREIE4FGAAUtr2XXkaPjZL6WmmJ4b0UfWqab4jAQAKUYABcXg3AEhZzjn99vXZWrhmi567vL9a1K/hOxIAIBoFGBCHPWAAUtb4b5bqtWkrdO2JnXRc50zfcQAAsRo3li6/XGrZ0ncSIDTYHAEgJc1cvlF/eGuuftY5U9ec0Ml3HABAIp06Sc8+6zsFECrsAQOQcjZu36XhY7OUWae6Hr6wl6pU4WLLAAAgNVCAAUgpBQVO1780XWu25OjxIX3UsFY135EAAHvz/vtSerr09de+kwChQQEGIKU88dki/XvBWv3ujG7q1aq+7zgAgH3Jywt+0hihFihEAQYgZfxv0To99NFCnd2ruYYe0cZ3HABAcRgFEYhDAQYgJfy4KUfXTJimDpm19edfHCozzvsCgNCjAAPiUIABCL3c/AKNHJ+lnNx8jR7aVzWr0ZEDQEqgAAPi8G4AEHp/njRfU5ds0KMX91bHJrV9xwEAlFTHjtLVV0sNG/pOAoQGBRiAUHt35io9+78fdNlRbXVmz+a+4wAASqNfv+AHwG4cggggtBat2aqbX5mhPq3r6/aBB/uOAwAordxcKSdHcs53EiA0KMAAhNL2XXkaMW6qqqen6fEhfVStKh9XAJBynn5aqlFDWrPGdxIgNDgEEUDoOOd0+2uz9N2arXr+l/11UL0aviMBAPYHg3AAcdikDCB0xn69VG9MX6nfnNRZx3bK9B0HALC/CgswLsQM7EYBBiBUZizbqLvfnqufd8nUyJ939B0HAHAg2AMGxKEAAxAaG7bt0ohxWcqsU11/u7CXqlThYssAkNIowIA4vBsAhEJBgdP1E6dr7ZademX4kapfs5rvSACAA3XUUdLtt0vp6b6TAKFBAQYgFB779yJ9tmCt7hnUXT1a1vcdBwBQFo47LvgBsBuHIALw7r/frdXfPl6oc3q30JDDW/uOAwAoK1u3SuvW+U4BhAoFGACvVm7coWsmTFOnJrV17zndZcZ5XwBQYdxzj9Sqle8UQKhQgAHwZldegUaMy1JuvtPooX1VsxpHRQNAhZKXxwAcQAzeEQC8+dOkeZq+bKMeH9xHHTJr+44DAChrFGBAHPaAAfDirRkr9dyXi/XLo9vp9B4H+Y4DACgPFGBAHAowAEm3aM0W3frqTPVt00C3DezqOw4AoLzk5lKAATF4RwBIqm078zvP8/AAABg9SURBVDRsbJZqpKfp8cF9lJ7GdiAAqLDOOUfq3t13CiBUKMAAJI1zTre9NkvZa7fqhV8drmb1MnxHAgCUpwEDgh8Au7HpGUDSvDB5id6asVI3nNJFR3ds7DsOAKC8/fijtGyZ7xRAqFCAAUiKaUs36O535urErk00/LgOvuMAAJLhuuukk0/2nQIIFQowAOVu/bZdGjkuS03rZuihC3qpShUutgwAlQKjIAJxeEcAKFf5BU7XvTRd67bu0qvDj1K9mum+IwEAkoUCDIhT7B4wM2tlZv82s3lmNsfMrk1GMAAVw6OffqfPF67V7886RIe2rOc7DlBiZjbAzBaY2SIzu3Uf8x1mZvlmdl4y8wEpgQIMiFOSQxDzJN3gnDtY0hGSRppZt/KNBaAi+GzBGj3yyXf6RZ8Wurh/K99xgBIzszRJj0s6TVI3SRcn6vsi890v6YPkJgRSBAUYEKfYd4RzbpWkVZHpLWY2T1ILSXPLORuAFLZi4w5d99J0dWlaR/cOOlRmnPeFlNJf0iLnXLYkmdmLks5WfN93taRXJR2W3HhAihg5Utq503cKIFRKtUnCzNpK6i3p6/IIA6Bi2JmXrxHjspSX7/TEkD6qUS3NdySgtFpIih47e7mkw6NnMLMWks6RdIIowIDEzjzTdwIgdEo8CqKZ1Vawle8659zmBO1XmdkUM5uydu3asswIIMXc++48zVi2UX85v4faZ9b2HQfYH4l22bqY+w9LusU5l7/PBdE/ojJbuFDKzvadAgiVEhVgZpauoPga55x7LdE8zrmnnHP9nHP9MjMzyzIjgBTy5vQVev6rJbrimHYa0P0g33GA/bVcUvSJiy0lrYyZp5+kF81ssaTzJD1hZoNiF0T/iErt0kulESN8pwBCpdhDEC04ceMZSfOccw+VfyQAqeq71Vt066uzdFjbBrrltK6+4wAH4ltJncysnaQVki6SNDh6Budcu8JpM3tO0jvOuTeSGRIIPQbhAOKUZA/Y0ZIukXSCmU2P/Aws51wAUszWnXkaNnaqalVP02OD+yg9jeu8I3U55/IkjVIwuuE8SROdc3PMbJiZDfObDkghFGBAnJKMgviFEh8LDwCSJOecbn11pn5Yt03jrjhCTetm+I4EHDDn3CRJk2IeG7OXeS9LRiYg5VCAAXHYRA3ggP3ry8V6Z+Yq3XhqFx3ZoZHvOACAsKAAA+LwjgBwQLKWbtC9k+bppIObaNjPOviOAwAIk/vvlxo08J0CCBUKMAD77aetOzVyXJaa1cvQX8/vpSpVOFoZABDl7LN9JwBCh0MQAeyX/AKn616arp+27dLoIX1Vr2a670gAgLCZPFlatMh3CiBUKMAA7JdHPvlO//1unf541iHq3qKe7zgAgDAaNEh68EHfKYBQoQADUGr/XrBGf//kO53Xt6UuPKxV8U8AAFROeXlSOkdIANEowACUyvIN23X9S9PVtVkd3X12dwXXagcAIAFGQQTiUIABKLGdefkaMS5L+flOY4b2VY1qab4jAQDCjAIMiMM7AkCJ3f3OXM1cvkljhvZV28a1fMcBAIQdBRgQh3cEgBJ5Y9oKjZ28VFf9rL0GdG/mOw4AIBVMnCi1b+87BRAqFGAAirVw9Rbd9tos9W/bUDed2sV3HABAqjjrLN8JgNDhHDAA+7R1Z56GjZ2qWtWr6rHBvZWexscGAKAECgqkSZOk77/3nQQIFb5JAdgr55xueWWmlvy0XY8N7q0mdTN8RwIApIpdu6TTT5deftl3EiBUKMAA7NU//7dY785apZtO7aIj2jfyHQcAkEry8oJbBuEAiqAAA5DQ1CXr9adJ83Ryt6b69c84gRoAUEqFBVgalywBolGAAYizbutOjRw3TS0a1NBfzu/JxZYBAKXHHjAgId4RAIrIL3C69sVp2rB9l14bcZTq1Uj3HQkAkIoowICEeEcAKOLhjxfqf4t+0gPn9tAhzev5jgMASFUNGkgffyx17uw7CRAqFGAAdvv3/DV69NNFuqBfS11wWCvfcQAAqax6denEE32nAEKHc8AASJKWrd+u616aroMPqqs/nt3ddxwAQKrbskV66SVp6VLfSYBQoQADoJzcfI0Yl6UC5zRmaB9lpDNiFQDgAK1YIV10kfTll76TAKHCIYgA9Md35mrWik166pK+atOolu84AICKgEE4gITYAwZUcq9lLdf4r5dq2HEddMohzXzHAQBUFBRgQEIUYEAlNv/Hzbr99Vk6vF1D3XgKo1QBAMoQBRiQEAUYUEltycnV8LFZqpuRrkcH91bVND4OAABliAIMSIh3BFAJOed08ysztXT9dk248gg1qZPhOxIAoKLp3l36+muuAwbEoAADKqFnvvhB783+UbcP7Kr+7Rr6jgMAqIhq15b69/edAggdjjkCKpkpi9frvvfm69RDmurKY9v7jgMAqKhWrJCeeUZavdp3EiBUKMCASmTtlp0aOT5LLRvU0IPn95SZ+Y4EAKioZs+WrrhCys72nQQIFQowoJLIyy/QNROmaeP2XD0xpK/qZqT7jgQAqMgYhANIiHcEUEk89NFCfZX9kx48r4e6Na/rOw4AoKKjAAMSYg8YUAl8Mm+1nvjse110WCud36+V7zgAgMqAAgxIiAIMqOCW/rRd1780XYc0r6vfn3WI7zgAgMqisABL55B3IBoFGFCB5eTma8T4qZKk0UP6KiM9zXMiAEClcdpp0ty5Urt2vpMAocI+YaAC+8PbczR7xWY9fWk/tW5U03ccAEBlUrdu8AOgCPaAARXUy1OWacI3yzTi+A46qVtT33EAAJXNrFnSww9LW7f6TgKESrEFmJk9a2ZrzGx2MgIBOHDzVm3Wb9+YrSPbN9JvTu7sOw6QksxsgJktMLNFZnZrgvYhZjYz8vOlmfX0kRMIrS+/lK6/XtqyxXcSIFRKsgfsOUkDyjkHgDKyOSdXw8dOVb0a6fr7xb1VNY0d3UBpmVmapMclnSapm6SLzaxbzGw/SDrOOddD0t2SnkpuSiDkGAURSKjYb2bOuc8lrU9CFgAHyDmnGyfO0LINO/T4kD7KrFPddyQgVfWXtMg5l+2c2yXpRUlnR8/gnPvSObchcneypJZJzgiEGwUYkBCbxoEK5B//zdaHc1frttO66rC2DX3HAVJZC0nLou4vjzy2N7+S9F65JgJSDQUYkFCZvSPM7CpJV0lS69aty2qxAEromx/W6/73F+i07s30q2MY8hc4QJbgMZdwRrOfKyjAjtlLO/0jKicKMCChMtsD5px7yjnXzznXLzMzs6wWC6AE1mzJ0ajxWWrdsKYeOK+HzBJ9dwRQCssltYq631LSytiZzKyHpKclne2c+ynRgugfUWmNGCEtWyZlZPhOAoQKmySAFJeXX6Crx0/T5pxcPf+r/qqTke47ElARfCupk5m1k7RC0kWSBkfPYGatJb0m6RLn3MLkRwRCrk6d4AdAESUZhn6CpK8kdTGz5Wb2q/KPBaCk/vLhQn39w3rdO+hQdW3GBS+BsuCcy5M0StIHkuZJmuicm2Nmw8xsWGS230lqJOkJM5tuZlM8xQXC6dNPpXvv9Z0CCJ1i94A55y5ORhAApffR3NUa85/vdXH/1jq3LwOwAWXJOTdJ0qSYx8ZETV8h6Ypk5wJSxkcfSQ89JN1xh+8kQKgwCiKQopb8tE2/mThd3VvU1V1nxl6eCAAAz/LyGIADSIACDEhBObn5Gj42SyZp9JC+ykhP8x0JAICiKMCAhHhXACnorjfnaO6qzXr2sn5q1bCm7zgAAMSjAAMSYg8YkGImTlmml6Ys06ifd9QJXZv6jgMAQGIUYEBCFGBACpmzcpPufGO2ju7YSNef3Nl3HAAA9u5vf5MWcoUGIBabJYAUsWlHrkaMy1L9mul65KLeSqvCxZYBACGWkcFFmIEE2AMGpADnnG56eYZWbNihJ4b0UePa1X1HAgBg38aPl+6/33cKIHQowIAU8NTn2fpw7mrdNvBg9W3T0HccAACK9/bb0rPP+k4BhA4FGBByk7N/0gMfLNDphx6kXx7d1nccAABKhkE4gIQowIAQW7M5R6PGT1ObhjV137mHyozzvgAAKSIvT0rjOpVALDZLACGVl1+gUROmadvOPI274nDVyUj3HQkAgJJjDxiQEO8KIKQe/GCBvvlhvf52YU91aVbHdxwAAEqHAgxIiHcFEEIfzPlRT36erSGHt9Y5vVv6jgMAQOlNmiQVFPhOAYQO54ABIbN43TbdOHGGerSsp9+d2c13HAAA9o8Z54ABCVCAASGSk5uv4eOyVKWK6fHBfVS9Kh0XACBFPfig9MgjvlMAoUMBBoTInW/M1vwfN+vhi3qpVcOavuMAALD/XnstOAwRQBEUYEBIvPTtUr08dbmu/nlH/bxLE99xAAA4MAzCASREAQaEwOwVm3Tnm3N0bKfGuvakzr7jAABw4CjAgIQowADPNu3I1YhxWWpUq5oevrCX0qpwsWUAQAVAAQYkxLsC8KigwOmGiTO0cuMOvfTrI9WodnXfkQAAKBvVqkk1avhOAYQOBRjg0ZOfZ+vjeat115nd1LdNA99xAAAoO1On+k4AhBKHIAKefPn9Oj34wXyd3uMgXXZUW99xAAAAkAQUYIAHqzfn6JoJ09SucS3df24PmXHeFwCgghk1Snr6ad8pgNChAAOSLDe/QKPGZ2nbznyNHtpXtatzJDAAoAJ69VXp2299pwBCh29+QJI98P58fbt4gx65qJc6N63jOw4AAOWDURCBhNgDBiTR+7NX6R///UGXHtlGZ/dq4TsOAADlhwIMSIgCDEiS7LVbdePLM9WzVX3dcfrBvuMAAFC+KMCAhCjAgCTYsStfI8ZlKT3N9MSQPqpeNc13JAAAyldmplSvnu8UQOiwWQIoZ8453fHGLC1YvUXPXd5fLepzUUoAQCWQne07ARBK7AEDytmEb5bptawVuvbETjquc6bvOAAAAPCIAgwoR7OWb9Lv35qjn3XO1DUndPIdBwCA5CgokM46S5o40XcSIHQowIBysml7roaPm6rGtavp4Qt7qUoVLrYMAKgk8vKkt9+WFi3ynQQIHc4BA8pBQYHTbyZO1+rNOZr46yPVsFY135EAAEievLzgllEQgTjsAQPKwej/fK9P5q/Rb0/vpt6tG/iOAwBAclGAAXtFAQaUsf8tWqe/frhAZ/ZsrkuPbOM7DgAAyUcBBuwVBRhQhn7clKNrJkxT+8zauu8Xh8qM874AAJWQc1KnTlLDhr6TAKFTogLMzAaY2QIzW2Rmt5Z3KCAVbduZp1Hjs7QjN19jhvZRreps9QNSWXF9nwX+HmmfaWZ9fOQEQqlRI2nhQmnoUN9JgNAptgAzszRJj0s6TVI3SRebWbfyDgakkv9+t1anPvy5pizZoPvP7aGOTer4jgTgAJSw7ztNUqfIz1WSRic1JAAgJZVkE31/SYucc9mSZGYvSjpb0tzyCrVozRbl5rvyWjxQZvILnP715WK9PHW52mfW0svDjtRhbTncAqgAStL3nS3peeeckzTZzOqb2UHOuVXlmmzFCumnn4o+ZiYdemgwvXSptHFj0fa0NOmQQ4LpxYulzZuLtlerJnXtGkxnZ0tbtxZtz8iQOncOpr/7Ttqxo2h7zZpSx47B9IIF0s6dRdvr1JHatQum582TcnOLtterJ7WJnDM7e3ZwDaloDRpIrVoF0zNnKk7jxlLz5sHzZs+Ob2/SRGrWLDgvaW6Cry/NmgXz7NwZ5I/VvHnwO3bsCF5/rFatgozbtknffx/f3qZN8Bo3bw7Wf6x27YJ1tHFj8PeL1bFjsI7Xr5eWL49v79w5+ButXSutSvDv17Vr8DdevTr4iXXIIcH/yKpVwTJiHXpo8D9Wmv+9deukP/5Ruv126ZRT4pcJVGIlKcBaSFoWdX+5pMPLJ07g0me+0cpNOeX5K4Ayk1bFNOL4DrrmxE7KSE/zHQdA2ShJ35donhaSyrcAu/deaXTMzraMjD1F0R13SGPHFm3PzJTWrAmmr7tOevPNou3t2+8pHK66Svrkk6LtPXtK06cH00OHSt98U7T96KOlL74Ips85Jyiyog0YIL33XjB9yinxRcT55++5YO8xx0ibNhVt/+UvpWeeCab79JHy84u2X3ut9PDDQQHVs6fi3HGHdM89QQGTqP2++6RbbglyJWp//HFpxIigOOvdO779+eelSy6Rpk2Tjj02vv3116VBg6T//U8aODC+/aOPpJNOkj78ULrwwvj2yZOlww8PlnPFFfHtc+ZI3bpJ48ZJ118f3750aVAk/uMf0p13xrdv2CDVrx+swwceiG/PzQ0G09if/71rr41fHlDJlaQASzSKQNzuKTO7SsEhGGrduvUBhbr3F4dqZ25+8TMCIdCxSR11bFLbdwwAZaskfV/S+0dJ0pVXBl/Wo1WJOqPgmmuCIiha9ep7pm+5Rbr00qLttWrtmb7rrqDYiFav3p7p++4LvrBHa9Roz/Qjj0hbthRtb9p0z/RTT8XvQWvRYs/0Cy/E7yFrEzWi7MsvBwM8ROvQIbitVk169VXF6dIluK1bN3F74d7Bpk0Tt/fqFdy2bZu4vV+/4LZr18Tt/fsHt717J27v0SO4PeqoxO2dOgW3J56YuL1ly+D2jDOkRP9jhX+f884LCrVYNWsGt5dcEhR6sQr/v0r7v5eRIZ18cvzygErOXOyHWOwMZkdK+r1z7tTI/dskyTn35709p1+/fm7KlCllmRMAEEJmNtU51893jrJWkr7PzJ6U9JlzbkLk/gJJx+/rEET6RwCoPPbWR5ZkFMRvJXUys3ZmVk3SRZLeKuuAAACESEn6vrckXRoZDfEISZvK/fwvAEDKK/YQROdcnpmNkvSBpDRJzzrn5pR7MgAAPNlb32dmwyLtYyRNkjRQ0iJJ2yVd7isvACB1lOhCRc65SQo6GgAAKoVEfV+k8CqcdpJGJjsXACC1lehCzAAAAACAA0cBBgAAAABJQgEGAAAAAElCAQYAAAAASUIBBgAAAABJQgEGAAAAAEliwSi6ZbxQs7WSlpT5gv1oLGmd7xAhxHqJxzqJxzqJV9HWSRvnXKbvEKmigvWPUsX7fy4LrJN4rJPEWC/xKto6SdhHlksBVpGY2RTnXD/fOcKG9RKPdRKPdRKPdYKKhP/neKyTeKyTxFgv8SrLOuEQRAAAAABIEgowAAAAAEgSCrDiPeU7QEixXuKxTuKxTuKxTlCR8P8cj3USj3WSGOslXqVYJ5wDBgAAAABJwh4wAAAAAEgSCrBSMLMbzcyZWWPfWXwzswfNbL6ZzTSz182svu9MvpjZADNbYGaLzOxW33nCwMxamdm/zWyemc0xs2t9ZwoDM0szs2lm9o7vLEBZo4/cgz5yD/rIougf964y9ZEUYCVkZq0knSxpqe8sIfGRpO7OuR6SFkq6zXMeL8wsTdLjkk6T1E3SxWbWzW+qUMiTdINz7mBJR0gayXqRJF0raZ7vEEBZo4+MQx8p+si9oH/cu0rTR1KAldzfJN0siZPmJDnnPnTO5UXuTpbU0mcej/pLWuScy3bO7ZL0oqSzPWfyzjm3yjmXFZneouADtYXfVH6ZWUtJp0t62ncWoBzQR0ahj9yNPjIG/WNila2PpAArATM7S9IK59wM31lC6peS3vMdwpMWkpZF3V8uPkiLMLO2knpL+tpvEu8eVvAFtcB3EKAs0UcWiz5yD/rIKPSPRVSqPrKq7wBhYWYfS2qWoOkOSbdLOiW5ifzb1zpxzr0ZmecOBbvTxyUzW4hYgsfYAhxhZrUlvSrpOufcZt95fDGzMyStcc5NNbPjfecBSos+Mh59ZInQR+4F/eMelbGPpACLcM6dlOhxMztUUjtJM8xMCg4jyDKz/s65H5MYMen2tk4Kmdn/STpD0omu8l7PYLmkVlH3W0pa6SlLqJhZuoLOZZxz7jXfeTw7WtJZZjZQUoakumY21jk31HMuoEToI+PRR5YIfWQC9I9xKl0fyXXASsnMFkvq55xb5zuLT2Y2QNJDko5zzq31nccXM6uq4ATrEyWtkPStpMHOuTleg3lmwTexf0la75y7zneeMIls3bvROXeG7yxAWaOPDNBHBugj49E/7ltl6SM5Bwz76zFJdSR9ZGbTzWyM70A+RE6yHiXpAwUn0k6szB1LlKMlXSLphMj/x/TIli0AqAzoI0UfuRf0j2APGAAAAAAkC3vAAAAAACBJKMAAAAAAIEkowAAAAAAgSSjAAAAAACBJKMAAAAAAIEkowAAAAAAgSSjAAAAAACBJKMAAAAAAIEn+H3Pa+Krz4WH4AAAAAElFTkSuQmCC\n", | |
"text/plain": [ | |
"<Figure size 1080x720 with 2 Axes>" | |
] | |
}, | |
"metadata": { | |
"needs_background": "light" | |
}, | |
"output_type": "display_data" | |
} | |
], | |
"source": [ | |
"#Latex\n", | |
"display(Math(r'relu(x) = \\max(0,x) '))\n", | |
"display(Math(r\"'relu(x) = 1.(x>0) \"))\n", | |
"\n", | |
"v = np.linspace(-5,5,100)\n", | |
"plt.figure(1,figsize=(15,10))\n", | |
"plt.subplot(221)\n", | |
"plt.plot(v,relu[0](v))\n", | |
"plt.title(\"Funcion RELU\")\n", | |
"plt.subplot(222)\n", | |
"plt.plot(v,relu[1](v), \"--\", color=\"red\")\n", | |
"plt.title(\"Derivada función RELU\")\n", | |
"plt.savefig('images\\sigmoide.png')\n", | |
"plt.show()" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": {}, | |
"outputs": [], | |
"source": [] | |
} | |
], | |
"metadata": { | |
"kernelspec": { | |
"display_name": "Python 3", | |
"language": "python", | |
"name": "python3" | |
}, | |
"language_info": { | |
"codemirror_mode": { | |
"name": "ipython", | |
"version": 3 | |
}, | |
"file_extension": ".py", | |
"mimetype": "text/x-python", | |
"name": "python", | |
"nbconvert_exporter": "python", | |
"pygments_lexer": "ipython3", | |
"version": "3.7.7" | |
} | |
}, | |
"nbformat": 4, | |
"nbformat_minor": 4 | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Excelente,
pero en la presentación Latex de las funciones cosh y tanh tiene Vd. un error , debe presentar e elevado a x menos e elevado a -x, en vez de e elevado a -x menos e elevado a -x.
Buen trabajo