{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "##### 교재 페이지 (210-211) #####" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "5) Neural Nets(NN) for XOR" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From C:\\Users\\jsdata00010\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n", "Instructions for updating:\n", "Colocations handled automatically by placer.\n", "WARNING:tensorflow:From C:\\Users\\jsdata00010\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n", "Instructions for updating:\n", "Use tf.cast instead.\n", "step : 0 \n", "cost : 0.7689432 \n", "Weight :\n", " [[-0.73786277]\n", " [-0.98397326]] \n", "bias :\n", " [-1.3740427]\n", "Hypothesis:\n", " [[0.7980325 ]\n", " [0.5962994 ]\n", " [0.6538894 ]\n", " [0.41391763]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 500 \n", "cost : 0.6945399 \n", "Weight :\n", " [[-0.13889387]\n", " [-0.14975941]] \n", "bias :\n", " [-0.1712202]\n", "Hypothesis:\n", " [[0.54270077]\n", " [0.505365 ]\n", " [0.5080809 ]\n", " [0.4706754 ]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 1000 \n", "cost : 0.6931746 \n", "Weight :\n", " [[-0.0200031 ]\n", " [-0.02047592]] \n", "bias :\n", " [-0.02400697]\n", "Hypothesis:\n", " [[0.5060015 ]\n", " [0.50088274]\n", " [0.50100094]\n", " [0.4958821 ]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 1500 \n", "cost : 0.6931477 \n", "Weight :\n", " [[-0.00282511]\n", " [-0.00284567]] \n", "bias :\n", " [-0.00336317]\n", "Hypothesis:\n", " [[0.5008408]\n", " [0.5001294]\n", " [0.5001345]\n", " [0.4994231]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 2000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-0.00039678]\n", " [-0.00039767]] \n", "bias :\n", " [-0.00047118]\n", "Hypothesis:\n", " [[0.5001178 ]\n", " [0.50001836]\n", " [0.5000186 ]\n", " [0.49991918]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 2500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-5.5622913e-05]\n", " [-5.5661869e-05]] \n", "bias :\n", " [-6.600684e-05]\n", "Hypothesis:\n", " [[0.5000165 ]\n", " [0.50000256]\n", " [0.5000026 ]\n", " [0.49998868]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 3000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-7.7797395e-06]\n", " [-7.7829391e-06]] \n", "bias :\n", " [-9.254255e-06]\n", "Hypothesis:\n", " [[0.5000023 ]\n", " [0.50000036]\n", " [0.50000036]\n", " [0.49999842]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 3500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-1.0861320e-06]\n", " [-1.0863521e-06]] \n", "bias :\n", " [-1.295544e-06]\n", "Hypothesis:\n", " [[0.5000003 ]\n", " [0.50000006]\n", " [0.50000006]\n", " [0.4999998 ]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 4000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-1.4884695e-07]\n", " [-1.4906715e-07]] \n", "bias :\n", " [-2.3756289e-07]\n", "Hypothesis:\n", " [[0.50000006]\n", " [0.5 ]\n", " [0.5 ]\n", " [0.49999997]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 4500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 5000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 5500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 6000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 6500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 7000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 7500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 8000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 8500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 9000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 9500 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n", "step : 10000 \n", "cost : 0.6931472 \n", "Weight :\n", " [[-8.8497018e-08]\n", " [-8.8717215e-08]] \n", "bias :\n", " [-1.1909821e-07]\n", "Hypothesis:\n", " [[0.5]\n", " [0.5]\n", " [0.5]\n", " [0.5]] \n", "Correct:\n", " [[1.]\n", " [1.]\n", " [1.]\n", " [1.]] \n", "Accuracy: 0.5\n" ] } ], "source": [ "import tensorflow as tf\n", "import numpy as np\n", "\n", "x_data = np.array([[0, 0], [0, 1], [1, 0], [1, 1]], dtype=np.float32)\n", "y_data = np.array([[0], [1], [1], [0]], dtype=np.float32)\n", "X = tf.placeholder(tf.float32)\n", "Y = tf.placeholder(tf.float32)\n", "W = tf.Variable(tf.random_normal([2, 1]), name='weight')\n", "b = tf.Variable(tf.random_normal([1]), name='bias')\n", "\n", "# Hypothesis using sigmoid: tf.div(1., 1. + tf.exp(tf.matmul(X, W)))\n", "hypothesis = tf.sigmoid(tf.matmul(X, W) - b)\n", "\n", "# cost/loss function\n", "cost = -tf.reduce_mean(Y * tf.log(hypothesis) + (1 - Y) * tf.log(1 - hypothesis))\n", "train = tf.train.GradientDescentOptimizer(learning_rate=0.1).minimize(cost)\n", "\n", "# Accuracy computation\n", "# True if hypothesis>0.2 else False\n", "predicted = tf.cast(hypothesis > 0.2, dtype=tf.float32)\n", "accuracy = tf.reduce_mean(tf.cast(tf.equal(predicted, Y), dtype=tf.float32))\n", "\n", "# Launch graph\n", "with tf.Session() as sess:\n", " sess.run(tf.global_variables_initializer())\n", " for step in range(10001):\n", " sess.run(train, feed_dict={X: x_data, Y: y_data})\n", " if step % 500 == 0:\n", " print(\"step :\",step, \"\\ncost :\", sess.run(cost, feed_dict={X: x_data, Y: y_data}), \n", " \"\\nWeight :\\n\", sess.run(W), \"\\nbias :\\n\", sess.run(b))\n", " h, c, a = sess.run([hypothesis, predicted, accuracy], feed_dict={X: x_data, Y: y_data})\n", " print(\"Hypothesis:\\n\", h, \"\\nCorrect:\\n \", c, \"\\nAccuracy: \", a)\n", " " ] } ], "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": "3.7.3" } }, "nbformat": 4, "nbformat_minor": 2 }