{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "import numpy as np\n", "\n", "def sigmoid(x):\n", " return 1 / (1 + np.exp(-x))\n", "\n", "def relu(x):\n", " return np.maximum(x, 0)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "x = np.arange(-10, 10, 0.1)\n", "\n", "plt.plot(x, sigmoid(relu(relu(relu(x)))), label='triple relu with single sigmoid')\n", "plt.plot(x, sigmoid(sigmoid(sigmoid(sigmoid(x)))), label='quadruple sigmoids')\n", "plt.legend()\n", "\n", "plt.show()" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.7" } }, "nbformat": 4, "nbformat_minor": 2 }