08. Convolutional Neural Networks/01. Understanding Convolutional - MXNet.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Understanding Convolutional - MXNet (From Scratch).ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1YN4PT_Idstj",
        "colab_type": "text"
      },
      "source": [
        "#MXNet (From Scratch)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CLGHa8i6WEaA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import mxnet as mx\n",
        "from mxnet import nd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "ctx = mx.gpu()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5HUXvMyHYKoM",
        "colab_type": "text"
      },
      "source": [
        "#Understanding convolution"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u5Lekr1xNl3o",
        "colab_type": "text"
      },
      "source": [
        "##Getting the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SOMvJE7t2j-E",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "mnist = mx.gluon.data.vision.MNIST()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z_t0Blm529jz",
        "colab_type": "code",
        "outputId": "217d72f2-0d56-4f4a-d913-2306992a4f1e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "image, _ = mnist[2]\n",
        "\n",
        "image = image.flatten().astype('float32').as_in_context(ctx)\n",
        "\n",
        "print(image.shape)\n",
        "\n",
        "plt.imshow(image.asnumpy(), cmap = 'gray')\n",
        "plt.show()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(28, 28)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u7Vdh54QWESq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def convolution(inputs, kernal):\n",
        "    filter_hight, filter_width = kernal.shape\n",
        "    hight, width = (inputs.shape[0] - filter_hight) + 1, (inputs.shape[1] - filter_width) + 1\n",
        "    \n",
        "    layer = nd.zeros((int(hight), int(width)))\n",
        "    \n",
        "    row = 0\n",
        "    for i in range(0, inputs.shape[0]):\n",
        "        col = 0\n",
        "        for j in range(0, inputs.shape[1]):\n",
        "            if (i + filter_hight <= inputs.shape[0]) and (j + filter_width <= inputs.shape[1]):\n",
        "                layer[row, col] = (inputs[i:i + filter_hight, j:j + filter_width] * kernal).sum()\n",
        "            col += 1\n",
        "        row += 1\n",
        "    return layer"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xMCXBgIdbxwY",
        "colab_type": "text"
      },
      "source": [
        "## This filter we are using below can detect vertical edges in the image"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P-N08yRpxRRH",
        "colab_type": "code",
        "outputId": "71847184-3339-46ef-b34e-db45e66ba760",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.array([[1.0, 0.0, -1.0],\n",
        "                   [1.0, 0.0, -1.0],\n",
        "                   [1.0, 0.0, -1.0]], ctx = ctx)\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(26, 26)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i7C-WJGMbs1d",
        "colab_type": "text"
      },
      "source": [
        "## This filter we are using below can detect horizontal edges in the image"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uoTV2Wka41M0",
        "colab_type": "code",
        "outputId": "2d3264e3-97c8-4862-e774-1ea190c9023d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.array([[1, 1, 1],\n",
        "                   [0, 0, 0],\n",
        "                   [-1, -1, -1]], ctx = ctx)\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhB8I6i8CwpWycSTXeAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(26, 26)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v9mWwcq76iBO",
        "colab_type": "text"
      },
      "source": [
        "##Lets see some more examples with different types of filters."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JLap3b-76hjs",
        "colab_type": "code",
        "outputId": "6eb9cf42-75f4-42ff-b0ce-65e484b03d4e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (3, 3)).as_in_context(ctx)\n",
        "# wondering what this random filter would generate?? Me too!! lets see.\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(26, 26)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bImJi_y_bou5",
        "colab_type": "text"
      },
      "source": [
        "## Time to check out a 5 by 5 random filter"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7VD_H7Mu6yLg",
        "colab_type": "code",
        "outputId": "a91b4a84-cded-4a4d-d730-5e68ab81f6ac",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (5, 5)).as_in_context(ctx)\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFOUHkqL8QFKUH0iK8gNJUX4gqf8HoZZEwAix2FUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(24, 24)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VpAt4zZAblY3",
        "colab_type": "text"
      },
      "source": [
        "## Lets try one more time"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WE8O2o-G9d2j",
        "colab_type": "code",
        "outputId": "0c5cafe3-ed15-49c5-8938-5e029508b9f3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (5, 5)).as_in_context(ctx)\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(24, 24)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BbkQAYeHbbMn",
        "colab_type": "text"
      },
      "source": [
        "## Lets try 7 by 7 this time"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NpEovLkB6yFm",
        "colab_type": "code",
        "outputId": "6d54735d-7f4c-442c-e529-e0f168594659",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (7, 7)).as_in_context(ctx)\n",
        "\n",
        "output = convolution(image, kernal)\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQgAAAD8CAYAAACLgjpEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(22, 22)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    }
  ]
}