Appendix_B_generate.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Appendix_B_generate.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ru5DFA0hWB0X",
        "colab_type": "text"
      },
      "source": [
        "# Appendix B - Generating Images From A Probability Distribution\n",
        "\n",
        "Make Your First GAN With PyTorch, 2020"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WjY6jMhhaXgE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import random\n",
        "import numpy\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hVEapqKVag2g",
        "colab_type": "code",
        "outputId": "bb3ef932-a966-409f-b51c-afbdcfbac2a2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "# counts are the observed number of coloured pixels at each position from a dataset of 8 examples\n",
        "\n",
        "counts = numpy.array([[6,0,1],\n",
        "                             [0,7,0],\n",
        "                             [2,0,1]], dtype=float)\n",
        "\n",
        "plt.imshow(counts, cmap='Reds', interpolation=None)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fba12fb0e10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQ8AAAD8CAYAAABpXiE9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ojaVgW2ia3HY",
        "colab_type": "code",
        "outputId": "3b0cd7cf-44e8-4089-c970-6f991f788c78",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "# probabilities are (counts) / (total possible)\n",
        "\n",
        "probabilities = counts/8\n",
        "probabilities"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.75 , 0.   , 0.125],\n",
              "       [0.   , 0.875, 0.   ],\n",
              "       [0.25 , 0.   , 0.125]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YvtdBCE9gd4L",
        "colab_type": "code",
        "outputId": "bec32fdc-e3c6-4ddc-fdf4-ed4864be9f25",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "# make image of 0's and 1's with proabilities\n",
        "\n",
        "r = numpy.random.uniform(size=(3,3))\n",
        "diff = probabilities - r\n",
        "img = numpy.ceil(diff)\n",
        "\n",
        "plt.imshow(img, cmap='Blues', interpolation=None)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fba13091550>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 92
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQ8AAAD8CAYAAABpXiE9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PDj1pA_vbxCJ",
        "colab_type": "code",
        "outputId": "dec6585a-f69a-41e1-9dff-b9a322448fbf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 469
        }
      },
      "source": [
        "# plot several outputs each samples against probabilies\n",
        "\n",
        "# plot a 6 column, 4 row array of images\n",
        "f, axarr = plt.subplots(4, 6, figsize=(12,8))\n",
        "for i in range(4):\n",
        "    for j in range(6):\n",
        "        r = numpy.random.uniform(size=(3,3))\n",
        "        diff = probabilities-r\n",
        "        img = numpy.ceil(diff)\n",
        "        axarr[i,j].set_xticks([])\n",
        "        axarr[i,j].set_yticks([])\n",
        "        axarr[i,j].imshow(img, interpolation='none', cmap='Blues')\n",
        "        pass\n",
        "    pass"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAq4AAAHECAYAAAD4eKinAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
            "text/plain": [
              "<Figure size 864x576 with 24 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qAUxMX90j3yr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}