08. Convolutional Neural Networks/03. Convolution with Padding - MXNet.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Convolution with Padding - MXNet.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": "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": "a6d0c055-db0a-460f-cdd0-a4ef81039278",
        "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": "markdown",
      "metadata": {
        "id": "9b90a1cOfK_9",
        "colab_type": "text"
      },
      "source": [
        "##Understanding how to pad tensors"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PtN8tpXFcq1b",
        "colab_type": "code",
        "outputId": "517a22f8-7213-48b1-b808-b166f6810dd5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "x = nd.ones((5, 5))\n",
        "print(x)\n",
        "plt.imshow(x.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "ht, hb, vl, vr, = nd.zeros((1, 7)), nd.zeros((1, 7)), nd.zeros((5, 1)), nd.zeros((5, 1))\n",
        "# (1, 7) because after adding 2 columns at left and right we are gonna have total 5 + 1 + 1 columns \n",
        "\n",
        "y = nd.concat(vl, x, vr, dim = 1)\n",
        "print(y)\n",
        "plt.imshow(y.asnumpy(), cmap = 'gray')\n",
        "plt.show()\n",
        "\n",
        "y = nd.concat(ht, y, hb, dim = 0)\n",
        "print(y)\n",
        "plt.imshow(y.asnumpy(), cmap = 'gray')\n",
        "plt.show()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "[[1. 1. 1. 1. 1.]\n",
            " [1. 1. 1. 1. 1.]\n",
            " [1. 1. 1. 1. 1.]\n",
            " [1. 1. 1. 1. 1.]\n",
            " [1. 1. 1. 1. 1.]]\n",
            "<NDArray 5x5 @cpu(0)>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPgAAAD8CAYAAABaQGkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "[[0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]]\n",
            "<NDArray 5x7 @cpu(0)>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU8AAAD8CAYAAAD35CadAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQDwBBsQTYEA8AQbEE2BAPAEG/h/g44rJGZvb8gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "[[0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 1. 1. 1. 1. 1. 0.]\n",
            " [0. 0. 0. 0. 0. 0. 0.]]\n",
            "<NDArray 7x7 @cpu(0)>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPgAAAD8CAYAAABaQGkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBi41ZuBSYwYuNWbgUmMGLjVm4FJj/wETHwEXRSg9zgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hRcvH1t93GwG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def convolution_with_padding(inputs, kernal, padding = (1, 1)):\n",
        "    filter_hight, filter_width = kernal.shape\n",
        "    horizental_pad, vertical_pad = padding\n",
        "    hight, width = (inputs.shape[0]  + (2 * vertical_pad) - filter_hight) + 1, (inputs.shape[1] + (2 * horizental_pad) - filter_width) + 1\n",
        "    \n",
        "    horizental_paddings = nd.zeros((horizental_pad, inputs.shape[1])).as_in_context(ctx)\n",
        "    vertical_paddings = nd.zeros((inputs.shape[0] + 2 * horizental_pad, vertical_pad)).as_in_context(ctx)\n",
        "    \n",
        "    # + (2 * horizental_pad) because we are gonna add of 2 horizontal paddings \n",
        "    \n",
        "    plt.imshow(inputs.asnumpy(), cmap = 'gray')\n",
        "    plt.title(f'shape: {inputs.shape}')\n",
        "    plt.show()\n",
        "    \n",
        "    inputs = nd.concat(horizental_paddings, inputs, horizental_paddings, dim = 0) # dim = 0 for concat alng rows\n",
        "    inputs = nd.concat(vertical_paddings, inputs, vertical_paddings, dim = 1) # dim = 0 for concat alng columns\n",
        "    \n",
        "    plt.imshow(inputs.asnumpy(), cmap = 'gray')\n",
        "    plt.title(f'shape: {inputs.shape}')\n",
        "    plt.show()\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",
        "        \n",
        "    return layer"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SGlxVqpXucgd",
        "colab_type": "text"
      },
      "source": [
        "##Padding with shape (2, 2)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SgESxOJ_sSsB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 826
        },
        "outputId": "fdada3d0-380c-42f5-eb6d-08f60aa0b7ad"
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (3, 3)).as_in_context(ctx)\n",
        "\n",
        "output = convolution_with_padding(image, kernal, padding = (2, 2))\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.title('After convolution')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAEOlJREFUeJzt3X2sVHV+x/H3B5HVuILgww2CCrVo\nareIW7TVssoWl7psumgarcQHbF2vicu2ppt2jW0ijd1EN2qzyUaz+BBw8aFuEDVWq2i22hqlXCyL\nPCggBeUWYY2Liyvqot/+cc61I945M8zTmcvv80omd+Z8z8P3DvfD75w5M3MUEZhZeoaV3YCZlcPh\nN0uUw2+WKIffLFEOv1miHH6zRDn8XUDSfEmLy+5jMJJOkdQnSWX3UkTSdyTdXHYfQ4nDb7XcCNwS\nESHpC5LulrRV0m5JqyR9vXJmSRdJWp/X10k6v56NtGDddwKXSDqm6d84EQ6/VSVpLPBV4JF80nDg\nTeAcYBTwD8BDkibk848DFgN/A4wE/ha4v85ANrXuiPgAeBK4vPHfOC0OfwdJ+p6k/nzkek3SjIry\nCEn35rW1kqZWLHedpNcrRrwLKmpXSHpB0o8kvSvp1cr1ShqVj6jb823/k6SD6mz5a8DLebCIiF9H\nxPyI2BIRn0TE48D/AL+fzz8e2BURT0bmX4FfAyfW2lCL1v3vwDfq/N2S5/B3iKSTgXnA6RFxOPAn\nwJaKWb4JPAgcATwG/Kii9jrwFbIR8R+BxfmoPOAP8nmOAm4AHpY0Jq8tBPYCvw2cBswEvpX3dLyk\nXZKOr9L27wGvFfxOPcBJwNp8Uh+wXtI3JR2U75Z/CKyuto4Wr3s9cOr+bitZEeFbB25k4dsJnAsc\nvE9tPvBMxeNTgD0F61oFzM7vXwH8L6CK+n8BlwE9ZAE5tKI2B/hZnT3fCdxUpXYw8Azw432mXwm8\nR/YfzvvANxp4rhpaNzAJ+Ljsf+uhcvPI3yERsQm4lizoOyU9KOnYilneqrj/PnCIpOEAki7PXwDb\nJWkX8CWyUX5Af+R//bmtwLHACWRB2l6x7I+Bel8U+yVw+L4TJQ0DfgJ8RLY3MzD9XOAHwHRgBNnx\n+12SptS5vWbXfTjwbr3bSp3D30ERcX9ETCMLZQA1T01JOoFsBJ4HHBkRRwBrgMpTb+P2ORV3PNne\nwJtkI/9REXFEfhsZEb9bZ8uryXa9K/sRcDfZXsWfRcRvKspTgOcjoi+y4/YVwHKyvZ2aWrDu3wF+\nXufvljyHv0MknSzpjyV9AfgA2AN8Useih5H9R/GLfD1/QTbyVzoG+CtJB0u6kCwET0TEduBp4FZJ\nIyUNk3SipHPqbHsZ8GVJh1RMuyNf/59GxJ595l8BfGVgNJZ0GtlrFavzx9MlFX2GvOF1584he8Xf\n6lH2cUcqN2Ay2bH4buAd4HHg2Lw2H1hcMe8EssAPzx9/P1/mbeA24DngW3ntCuAFshcI3wU2ADMr\n1jWKLFTb8vp/AxfntePJjqGPL+j7p8Cf5/cH9lg+yJcbuF1SMf88YFP+e24GvltRuwx4ocp2ml33\nIfnv2FP2v/VQuSl/4myIknQF2X8E09q0/lOARcAZ0eQfi6S7gJ9GxFMtae6z6/4OcFxE/F2r132g\ncviHuHaH3w5cPuY3S5RHfrNEeeQ3S9TwTm6sxmkeM2uBiKjr49dNjfySzss/oLJJ0nXNrMvMOqvh\nY/78k2EbyD75tY3sTRhzImJdwTIe+c3arBMj/xnApojYHBEfkX0ibXYT6zOzDmom/OPI3js+YFs+\n7TMk9Sr7Gqi+JrZlZi3W9hf8ImIBsAC822/WTZoZ+fuB4yoej8+nmdkQ0Ez4VwCTJE2UNAK4mOwb\naMxsCGh4tz8i9kqaBzwFHATcExFrayxmZl2io2/v9TG/Wft15E0+ZjZ0OfxmiXL4zRLl8JslyuE3\nS5TDb5Yoh98sUQ6/WaIcfrNEOfxmiXL4zRLl8JslyuE3S5TDb5Yoh98sUQ6/WaIcfrNEOfxmiXL4\nzRLl8JslyuE3S5TDb5Yoh98sUQ6/WaIcfrNEOfxmiXL4zRLl8JslyuE3S1TDl+g263YzZsyoWrvv\nvvsKlz3nnHMK66+99lpDPXWTpsIvaQuwG/gY2BsRU1vRlJm1XytG/q9GxNstWI+ZdZCP+c0S1Wz4\nA3ha0kpJvYPNIKlXUp+kvia3ZWYt1Oxu/7SI6Jd0DLBM0qsR8XzlDBGxAFgAICma3J6ZtUhTI39E\n9Oc/dwJLgTNa0ZSZtV/D4Zd0mKTDB+4DM4E1rWrMzNqrmd3+HmCppIH13B8R/9aSrtrg7LPPLqwf\neeSRhfWlS5e2sh3rgNNPP71qbcWKFR3spDs1HP6I2Ayc2sJezKyDfKrPLFEOv1miHH6zRDn8Zoly\n+M0SlcxHeqdPn15YnzRpUmHdp/q6z7BhxWPXxIkTq9ZOOOGEwmXzU9gHNI/8Zoly+M0S5fCbJcrh\nN0uUw2+WKIffLFEOv1mikjnPf/nllxfWX3zxxQ51Yq0yduzYwvpVV11VtbZ48eLCZV999dWGehpK\nPPKbJcrhN0uUw2+WKIffLFEOv1miHH6zRDn8ZolK5jx/rc9+29Bz1113Nbzsxo0bW9jJ0OREmCXK\n4TdLlMNvliiH3yxRDr9Zohx+s0Q5/GaJOmDO80+ePLmw3tPT06FOrFNGjRrV8LLLli1rYSdDU82R\nX9I9knZKWlMxbYykZZI25j9Ht7dNM2u1enb7FwLn7TPtOuDZiJgEPJs/NrMhpGb4I+J54J19Js8G\nFuX3FwHnt7gvM2uzRo/5eyJie37/LaDqAbWkXqC3we2YWZs0/YJfRISkKKgvABYAFM1nZp3V6Km+\nHZLGAuQ/d7auJTPrhEbD/xgwN78/F3i0Ne2YWafU3O2X9AAwHThK0jbgBuAm4CFJVwJbgYva2WQ9\nZs2aVVg/9NBDO9SJtUqt92ZMnDix4XX39/c3vOyBomb4I2JOldKMFvdiZh3kt/eaJcrhN0uUw2+W\nKIffLFEOv1miDpiP9J588slNLb927doWdWKtcssttxTWa50K3LBhQ9Xa7t27G+rpQOKR3yxRDr9Z\nohx+s0Q5/GaJcvjNEuXwmyXK4TdL1AFznr9ZK1asKLuFIWnkyJGF9fPO2/e7X//fpZdeWrjszJkz\nG+ppwI033li1tmvXrqbWfSDwyG+WKIffLFEOv1miHH6zRDn8Zoly+M0S5fCbJcrn+XNjxowpbdun\nnnpqYV1SYf3cc8+tWhs/fnzhsiNGjCisX3LJJYX1YcOKx489e/ZUrS1fvrxw2Q8//LCwPnx48Z/v\nypUrC+up88hvliiH3yxRDr9Zohx+s0Q5/GaJcvjNEuXwmyVKEdG5jUlt29jtt99eWL/66qsL67U+\n3/3GG2/sd0/1mjx5cmG91nn+vXv3Vq29//77hcuuW7eusF7rXHxfX19h/bnnnqta27FjR+Gy27Zt\nK6yPHj26sF7rPQwHqogo/oPJ1Rz5Jd0jaaekNRXT5kvql7Qqv81qplkz67x6dvsXAoN9Hcs/R8SU\n/PZEa9sys3arGf6IeB54pwO9mFkHNfOC3zxJq/PDgqoHX5J6JfVJKj44NLOOajT8dwAnAlOA7cCt\n1WaMiAURMTUipja4LTNrg4bCHxE7IuLjiPgEuBM4o7VtmVm7NRR+SWMrHl4ArKk2r5l1p5qf55f0\nADAdOErSNuAGYLqkKUAAW4Dik+gdcM011xTWt27dWlg/66yzWtnOfqn1HoJHHnmksL5+/fqqtZde\neqmhnjqht7e3sH700UcX1jdv3tzKdpJTM/wRMWeQyXe3oRcz6yC/vdcsUQ6/WaIcfrNEOfxmiXL4\nzRKVzFd333zzzWW3YPuYMWNGU8svWbKkRZ2kySO/WaIcfrNEOfxmiXL4zRLl8JslyuE3S5TDb5ao\nZM7z24Fn6dKlZbcwpHnkN0uUw2+WKIffLFEOv1miHH6zRDn8Zoly+M0S5fCbJcrhN0uUw2+WKIff\nLFEOv1miHH6zRDn8Zoly+M0SVc8luo8D7gV6yC7JvSAifihpDPAvwASyy3RfFBG/bF+rlhpJhfWT\nTjqpsN7NlyfvBvWM/HuB70bEKcAfAt+WdApwHfBsREwCns0fm9kQUTP8EbE9Il7O7+8G1gPjgNnA\nony2RcD57WrSzFpvv475JU0ATgOWAz0RsT0vvUV2WGBmQ0Td3+En6YvAEuDaiPhV5fFYRISkqLJc\nL9DbbKNm1lp1jfySDiYL/n0R8XA+eYeksXl9LLBzsGUjYkFETI2Iqa1o2Mxao2b4lQ3xdwPrI+K2\nitJjwNz8/lzg0da3Z2btUs9u/x8BlwGvSFqVT7seuAl4SNKVwFbgova0aKmKGPRI8lPDhvltKs2o\nGf6I+E+g2gnX5i6wbmal8X+dZoly+M0S5fCbJcrhN0uUw2+WKIffLFG+RLcNWWeeeWZhfeHChZ1p\nZIjyyG+WKIffLFEOv1miHH6zRDn8Zoly+M0S5fCbJcrn+a1r1frqbmuOR36zRDn8Zoly+M0S5fCb\nJcrhN0uUw2+WKIffLFE+z2+lefLJJwvrF154YYc6SZNHfrNEOfxmiXL4zRLl8JslyuE3S5TDb5Yo\nh98sUap1DXRJxwH3Aj1AAAsi4oeS5gNXAb/IZ70+Ip6osa7ijZlZ0yKiri9CqCf8Y4GxEfGypMOB\nlcD5wEXAexFxS71NOfxm7Vdv+Gu+wy8itgPb8/u7Ja0HxjXXnpmVbb+O+SVNAE4DlueT5klaLeke\nSaOrLNMrqU9SX1OdmllL1dzt/3RG6YvAc8D3I+JhST3A22SvA9xIdmjwlzXW4d1+szZr2TE/gKSD\ngceBpyLitkHqE4DHI+JLNdbj8Ju1Wb3hr7nbr+wrVO8G1lcGP38hcMAFwJr9bdLMylPPq/3TgP8A\nXgE+ySdfD8wBppDt9m8Brs5fHCxal0d+szZr6W5/qzj8Zu3Xst1+MzswOfxmiXL4zRLl8JslyuE3\nS5TDb5Yoh98sUQ6/WaIcfrNEOfxmiXL4zRLl8JslyuE3S5TDb5aoTl+i+21ga8Xjo/Jp3ahbe+vW\nvsC9NaqVvZ1Q74wd/Tz/5zYu9UXE1NIaKNCtvXVrX+DeGlVWb97tN0uUw2+WqLLDv6Dk7Rfp1t66\ntS9wb40qpbdSj/nNrDxlj/xmVhKH3yxRpYRf0nmSXpO0SdJ1ZfRQjaQtkl6RtKrs6wvm10DcKWlN\nxbQxkpZJ2pj/HPQaiSX1Nl9Sf/7crZI0q6TejpP0M0nrJK2V9Nf59FKfu4K+SnneOn7ML+kgYAPw\nNWAbsAKYExHrOtpIFZK2AFMjovQ3hEg6G3gPuHfgUmiSfgC8ExE35f9xjo6I73VJb/PZz8u2t6m3\napeVv4ISn7tWXu6+FcoY+c8ANkXE5oj4CHgQmF1CH10vIp4H3tln8mxgUX5/EdkfT8dV6a0rRMT2\niHg5v78bGLisfKnPXUFfpSgj/OOANyseb6PEJ2AQATwtaaWk3rKbGURPxWXR3gJ6ymxmEDUv295J\n+1xWvmueu0Yud99qfsHv86ZFxJeBrwPfzndvu1Jkx2zddK72DuBEsms4bgduLbOZ/LLyS4BrI+JX\nlbUyn7tB+irleSsj/P3AcRWPx+fTukJE9Oc/dwJLyQ5TusmOgSsk5z93ltzPpyJiR0R8HBGfAHdS\n4nOXX1Z+CXBfRDycTy79uRusr7KetzLCvwKYJGmipBHAxcBjJfTxOZIOy1+IQdJhwEy679LjjwFz\n8/tzgUdL7OUzuuWy7dUuK0/Jz13XXe4+Ijp+A2aRveL/OvD3ZfRQpa/fAn6e39aW3RvwANlu4G/I\nXhu5EjgSeBbYCDwDjOmi3n5Cdin31WRBG1tSb9PIdulXA6vy26yyn7uCvkp53vz2XrNE+QU/s0Q5\n/GaJcvjNEuXwmyXK4TdLlMNvliiH3yxR/wd/5iq6AR88ewAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3yxT/w/jJJV+R68RBwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30, 30)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "OyF1pJzGuhAm"
      },
      "source": [
        "##Padding with shape (3, 3)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a19A2OEghVtZ",
        "colab_type": "code",
        "outputId": "90df89a2-0951-4e78-b697-005edab9da5d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 826
        }
      },
      "source": [
        "kernal = nd.random.normal(0, 1, (5, 5)).as_in_context(ctx)\n",
        "\n",
        "output = convolution_with_padding(image, kernal, padding = (3, 3))\n",
        "\n",
        "plt.imshow(output.asnumpy(), cmap = 'gray')\n",
        "plt.title('After convolution')\n",
        "plt.show()\n",
        "\n",
        "output.shape"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nw2+WKYffLFP/BxpHbLjtPFs1AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(30, 30)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    }
  ]
}