lab-07-2-linear_regression_without_min_max.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lab 07-2 linear regression(without min/max)\n",
    "\n",
    "* linear regression에 대한 설명 (Data에 대한 표준화나 정규화 없이 수행)\n",
    "\n",
    "### 기본 Library 선언 및 Tensorflow 버전 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.12.0\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import tensorflow as tf\n",
    "\n",
    "tf.set_random_seed(777)  # for reproducibility\n",
    "\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 강의에 설명할 Data입니다\n",
    "* X Data(feature)의 값은 해당 배열의 첫번째 값부터 4번째 값까지로 정의 되고 Y Data(label)는 해당 배열의 마지막 값을 정의(5번째 값)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xy = np.array([[828.659973, 833.450012, 908100, 828.349976, 831.659973],\n",
    "               [823.02002, 828.070007, 1828100, 821.655029, 828.070007],\n",
    "               [819.929993, 824.400024, 1438100, 818.97998, 824.159973],\n",
    "               [816, 820.958984, 1008100, 815.48999, 819.23999],\n",
    "               [819.359985, 823, 1188100, 818.469971, 818.97998],\n",
    "               [819, 823, 1198100, 816, 820.450012],\n",
    "               [811.700012, 815.25, 1098100, 809.780029, 813.669983],\n",
    "               [809.51001, 816.659973, 1398100, 804.539978, 809.559998]])\n",
    "\n",
    "x_train = xy[:, 0:-1]\n",
    "y_train = xy[:, [-1]]\n",
    "\n",
    "plt.plot(x_train, 'ro')\n",
    "plt.plot(y_train)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Tensorflow Session\n",
    "### 위 Data를 기준으로 Linear Regression 모델을 만들도록 하겠습니다\n",
    "* Tensorflow data API를 통해 학습시킬 값들을 담는다 (Batch Size는 한번에 학습시킬 Size로 정한다)\n",
    "* X(features),Y(labels)는 실재 학습에 쓰일 Data (연산을 위해 Type를 맞춰준다)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = tf.data.Dataset.from_tensor_slices((x_train, y_train)).batch(len(x_train)).repeat()\n",
    "\n",
    "iter = dataset.make_initializable_iterator()\n",
    "features, labels = iter.get_next()\n",
    "\n",
    "features = tf.cast(features, tf.float32)\n",
    "labels = tf.cast(labels, tf.float32)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* W와 b은 학습을 통해 생성되는 모델에 쓰이는 Wegith와 Bias (초기값을 variable : 0이나 Random값으로 가능 tf.random_normal([3, 3],tf.zeros([3,3]) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "W = tf.Variable(tf.random_normal([4, 1]))\n",
    "b = tf.Variable(tf.random_normal([1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Linear Regression의 Hyphthesis를 정의한다 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "hypothesis = tf.matmul(features, W) + b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* 가설을 검증할 Cost 함수를 정의합니다(Mean Square Error를 사용)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "cost = tf.reduce_mean(tf.square(hypothesis - labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* GradientDescentOptimizer로 학습 실행"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = tf.train.GradientDescentOptimizer(learning_rate=1e-5)\n",
    "train = optimizer.minimize(cost)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tensorflow를 통한 실행을 위해 Session을 실행합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter: 0, Loss: 10469654.0000, Prediction: [[-2025.9463]\n",
      " [-3036.1917]\n",
      " [-2600.907 ]\n",
      " [-2121.0547]\n",
      " [-2323.4526]\n",
      " [-2335.566 ]\n",
      " [-2215.1558]\n",
      " [-2548.01  ]]\n",
      "Iter: 1, Loss: 11337537342327954079744.0000, Prediction: [[7.5108778e+10]\n",
      " [1.5120166e+11]\n",
      " [1.1894489e+11]\n",
      " [8.3379732e+10]\n",
      " [9.8267472e+10]\n",
      " [9.9094569e+10]\n",
      " [9.0823606e+10]\n",
      " [1.1563650e+11]]\n",
      "Iter: 2, Loss: 12456329037179253078450471170721447936.0000, Prediction: [[-2.4895813e+18]\n",
      " [-5.0117829e+18]\n",
      " [-3.9425887e+18]\n",
      " [-2.7637335e+18]\n",
      " [-3.2572078e+18]\n",
      " [-3.2846230e+18]\n",
      " [-3.0104708e+18]\n",
      " [-3.8329278e+18]]\n",
      "Iter: 3, Loss: inf, Prediction: [[8.2520532e+25]\n",
      " [1.6612232e+26]\n",
      " [1.3068243e+26]\n",
      " [9.1607684e+25]\n",
      " [1.0796455e+26]\n",
      " [1.0887326e+26]\n",
      " [9.9786122e+25]\n",
      " [1.2704757e+26]]\n",
      "Iter: 4, Loss: inf, Prediction: [[-2.7352545e+33]\n",
      " [-5.5063480e+33]\n",
      " [-4.3316454e+33]\n",
      " [-3.0364605e+33]\n",
      " [-3.5786309e+33]\n",
      " [-3.6087515e+33]\n",
      " [-3.3075458e+33]\n",
      " [-4.2111631e+33]]\n",
      "Iter: 5, Loss: inf, Prediction: [[inf]\n",
      " [inf]\n",
      " [inf]\n",
      " [inf]\n",
      " [inf]\n",
      " [inf]\n",
      " [inf]\n",
      " [inf]]\n",
      "Iter: 6, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 7, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 8, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 9, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 10, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 11, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 12, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 13, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 14, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 15, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 16, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 17, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 18, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 19, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 20, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 21, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 22, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 23, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 24, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 25, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 26, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 27, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 28, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 29, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 30, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 31, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 32, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 33, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 34, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 35, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 36, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 37, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 38, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 39, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 40, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 41, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 42, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 43, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 44, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 45, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 46, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 47, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 48, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 49, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 50, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 51, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 52, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 53, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 54, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 55, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 56, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 57, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 58, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 59, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 60, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 61, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 62, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 63, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 64, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 65, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 66, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 67, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 68, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 69, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 70, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 71, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 72, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 73, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 74, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 75, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 76, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 77, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 78, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 79, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 80, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 81, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 82, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 83, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 84, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 85, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 86, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 87, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 88, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 89, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 90, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 91, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 92, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 93, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 94, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 95, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 96, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 97, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 98, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 99, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n",
      "Iter: 100, Loss: nan, Prediction: [[nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]\n",
      " [nan]]\n"
     ]
    }
   ],
   "source": [
    "EPOCHS = 101\n",
    "\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    for step in range(EPOCHS):\n",
    "        sess.run(iter.initializer)\n",
    "        _, loss_value, hypo_value = sess.run([train, cost, hypothesis])\n",
    "        print(\"Iter: {}, Loss: {:.4f}, Prediction: {}\".format(step, loss_value, hypo_value))  "
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}