lab-11-0-cnn_basics.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 2, 2, 1)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x10e242da0>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARUAAAD8CAYAAABZ0jAcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10467a748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sess = tf.InteractiveSession()\n",
    "image = np.array([[[[1],[2]],\n",
    "                    [[3],[4]]]], dtype=np.float32)\n",
    "print(image.shape)\n",
    "plt.imshow(image.reshape(2,2), cmap='Greys')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conv Layer\n",
    "weight.shape = 2 filters (1 , 1 , 1)\n",
    "![image](https://cloud.githubusercontent.com/assets/901975/23337561/56236b68-fc2d-11e6-956e-bc24325a824d.png)\n",
    "\n",
    "![image](https://cloud.githubusercontent.com/assets/901975/23340105/a7ba8040-fc6b-11e6-8bc7-aba9f4cf7b78.png)\n",
    "(TensorFlow For Machine Intelligence: A hands-on introduction to learning algorithms \n",
    "by Sam Abrahams et al.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "imag:\n",
      " [[[[ 1.]\n",
      "   [ 2.]]\n",
      "\n",
      "  [[ 3.]\n",
      "   [ 4.]]]]\n",
      "weight.shape (1, 1, 1, 2)\n",
      "[[ 2.  4.]\n",
      " [ 6.  8.]]\n",
      "[[ 0.5  1. ]\n",
      " [ 1.5  2. ]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAC7CAYAAABl5ukfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAP8DK2tfX5an+ZwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10c0a5be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"imag:\\n\", image)\n",
    "weight = tf.constant([[[[2., 0.5]]]])\n",
    "print(\"weight.shape\", weight.shape)\n",
    "conv2d = tf.nn.conv2d(image, weight, strides=[1, 1, 1, 1], padding='SAME')\n",
    "conv2d_img = conv2d.eval()\n",
    "conv2d_img = np.swapaxes(conv2d_img, 0, 3)\n",
    "for i, one_img in enumerate(conv2d_img):\n",
    "    print(one_img.reshape(2,2))\n",
    "    plt.subplot(1,2,i+1), plt.imshow(one_img.reshape(2,2), cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MAX POOLING\n",
    "![image](https://cloud.githubusercontent.com/assets/901975/23337676/bd154da2-fc30-11e6-888c-d86bc2206066.png)\n",
    "\n",
    "![image](https://cloud.githubusercontent.com/assets/901975/23340355/a4bd3c08-fc6f-11e6-8a99-1e3bbbe86733.png)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 1, 1, 1)\n",
      "[[[[ 4.]]]]\n"
     ]
    }
   ],
   "source": [
    "image = np.array([[[[4],[3]],\n",
    "                    [[2],[1]]]], dtype=np.float32)\n",
    "pool = tf.nn.max_pool(image, ksize=[1, 2, 2, 1],\n",
    "                    strides=[1, 1, 1, 1], padding='VALID')\n",
    "print(pool.shape)\n",
    "print(pool.eval())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SAME: Zero paddings\n",
    "\n",
    "![image](https://cloud.githubusercontent.com/assets/901975/23340337/71b27652-fc6f-11e6-96ef-760998755f77.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 2, 2, 1)\n",
      "[[[[ 4.]\n",
      "   [ 3.]]\n",
      "\n",
      "  [[ 2.]\n",
      "   [ 1.]]]]\n"
     ]
    }
   ],
   "source": [
    "image = np.array([[[[4],[3]],\n",
    "                    [[2],[1]]]], dtype=np.float32)\n",
    "pool = tf.nn.max_pool(image, ksize=[1, 2, 2, 1],\n",
    "                    strides=[1, 1, 1, 1], padding='SAME')\n",
    "print(pool.shape)\n",
    "print(pool.eval())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting MNIST_data/train-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data/train-labels-idx1-ubyte.gz\n",
      "Extracting MNIST_data/t10k-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data/t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "mnist = input_data.read_data_sets(\"MNIST_data/\", one_hot=True)\n",
    "# Check out https://www.tensorflow.org/get_started/mnist/beginners for\n",
    "# more information about the mnist dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x115029ac8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIvxAUIQfCOq/esVX4lsZQ0YAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x104722080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img = mnist.train.images[0].reshape(28,28)\n",
    "plt.imshow(img, cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor(\"Conv2D_1:0\", shape=(1, 14, 14, 5), dtype=float32)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAABcCAYAAAB+6068AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10e310be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sess = tf.InteractiveSession()\n",
    "\n",
    "img = img.reshape(-1,28,28,1)\n",
    "W1 = tf.Variable(tf.random_normal([3, 3, 1, 5], stddev=0.01))\n",
    "conv2d = tf.nn.conv2d(img, W1, strides=[1, 2, 2, 1], padding='SAME')\n",
    "print(conv2d)\n",
    "sess.run(tf.global_variables_initializer())\n",
    "conv2d_img = conv2d.eval()\n",
    "conv2d_img = np.swapaxes(conv2d_img, 0, 3)\n",
    "for i, one_img in enumerate(conv2d_img):\n",
    "    plt.subplot(1,5,i+1), plt.imshow(one_img.reshape(14,14), cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor(\"MaxPool_2:0\", shape=(1, 7, 7, 5), dtype=float32)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAABcCAYAAABOZ1+dAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wLVfUXV6+EDAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b9b9a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pool = tf.nn.max_pool(conv2d, ksize=[1, 2, 2, 1], strides=[\n",
    "                        1, 2, 2, 1], padding='SAME')\n",
    "print(pool)\n",
    "sess.run(tf.global_variables_initializer())\n",
    "pool_img = pool.eval()\n",
    "pool_img = np.swapaxes(pool_img, 0, 3)\n",
    "for i, one_img in enumerate(pool_img):\n",
    "    plt.subplot(1,5,i+1), plt.imshow(one_img.reshape(7, 7), cmap='gray')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3.0
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}