PyTorch_42_MultipleIOChannels.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "PyTorch_42_MultipleIOChannels.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPEc2KbK7gPtt2YST4Fcjrg",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/codeplinth/pytorchbootcamp/blob/master/PyTorch_42_MultipleIOChannels.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YVCuGXkiPVIt"
      },
      "source": [
        "import torch\n",
        "from torch import nn\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eZNQBya9PgsI"
      },
      "source": [
        "conv1 = nn.Conv2d(in_channels=1,out_channels=3,kernel_size=3)"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T9R7sfUJQc3C",
        "outputId": "c7ad8c96-1774-44b4-9317-1e8f01dd77f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 260
        }
      },
      "source": [
        "#kernels are randomly initialized\n",
        "conv1.state_dict()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "OrderedDict([('weight', tensor([[[[ 0.0854, -0.2625,  0.3196],\n",
              "                        [-0.1633,  0.2931,  0.2618],\n",
              "                        [-0.1302,  0.0692,  0.1122]]],\n",
              "              \n",
              "              \n",
              "                      [[[-0.2505,  0.2641, -0.2499],\n",
              "                        [-0.0703,  0.0897, -0.2353],\n",
              "                        [ 0.1047,  0.0650, -0.1454]]],\n",
              "              \n",
              "              \n",
              "                      [[[-0.2333,  0.3042, -0.1587],\n",
              "                        [-0.0123, -0.1965, -0.0152],\n",
              "                        [ 0.0145,  0.3129, -0.0544]]]])),\n",
              "             ('bias', tensor([ 0.0449,  0.0748, -0.2507]))])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q4vl9S2gQiJ2",
        "outputId": "11a37d90-c784-4c71-ad21-5b0ed5416676",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 243
        }
      },
      "source": [
        "conv1.state_dict()['weight'][0][0] = torch.tensor([[1.0,0.0,-1.0],\n",
        "                                      [2.0,0.0,-2.0],\n",
        "                                      [1.0,0.0,-1.0]])\n",
        "conv1.state_dict()['weight'][1][0] = torch.tensor([[1.0,2.0,1.0],\n",
        "                                      [0.0,0.0,0.0],\n",
        "                                      [-1.0,-2.0,-1.0]])\n",
        "conv1.state_dict()['weight'][2][0] = torch.ones(3,3)\n",
        "conv1.state_dict()['weight']"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 1.,  0., -1.],\n",
              "          [ 2.,  0., -2.],\n",
              "          [ 1.,  0., -1.]]],\n",
              "\n",
              "\n",
              "        [[[ 1.,  2.,  1.],\n",
              "          [ 0.,  0.,  0.],\n",
              "          [-1., -2., -1.]]],\n",
              "\n",
              "\n",
              "        [[[ 1.,  1.,  1.],\n",
              "          [ 1.,  1.,  1.],\n",
              "          [ 1.,  1.,  1.]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-VNea2RHQ2Uj",
        "outputId": "4f2a1e67-4ec7-4ce6-b7e0-bdcf507775b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "conv1.state_dict()['bias'][:] = torch.tensor([0.0,0.0,0.0])\n",
        "conv1.state_dict()['bias']"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0., 0., 0.])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nVkGbw1DSVnw",
        "outputId": "6f5e0b3d-1316-401c-af8f-ae1a35beb351",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "image = torch.zeros(1,1,5,5)\n",
        "image[0,0,:,2] = 1\n",
        "image"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[0., 0., 1., 0., 0.],\n",
              "          [0., 0., 1., 0., 0.],\n",
              "          [0., 0., 1., 0., 0.],\n",
              "          [0., 0., 1., 0., 0.],\n",
              "          [0., 0., 1., 0., 0.]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4DXmWaeIS4iQ",
        "outputId": "8679d6d4-96c1-438e-9694-e59bf2105e6a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.imshow(image[0,0,:,:].numpy(),interpolation='nearest',cmap=plt.cm.gray)\n",
        "plt.colorbar()\n",
        "plt.show()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8cTjL8xxTdVO",
        "outputId": "b9f9705a-222b-45ba-9450-ddb34a0f0078",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "out = conv1(image)\n",
        "out.shape"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1, 3, 3, 3])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hb0VAxvlUDbK",
        "outputId": "8ff9cf1e-955e-4215-b475-0451aa273bee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 965
        }
      },
      "source": [
        "for channel,image in enumerate(out[0]):\n",
        "    plt.imshow(image.detach().numpy(), interpolation='nearest', cmap=plt.cm.gray)\n",
        "    print(image)\n",
        "    plt.title(f'channel : {channel}')\n",
        "    plt.colorbar()\n",
        "    plt.show()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "tensor([[-4.,  0.,  4.],\n",
            "        [-4.,  0.,  4.],\n",
            "        [-4.,  0.,  4.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[0., 0., 0.],\n",
            "        [0., 0., 0.],\n",
            "        [0., 0., 0.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[3., 3., 3.],\n",
            "        [3., 3., 3.],\n",
            "        [3., 3., 3.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Fty8ZF4UUemY",
        "outputId": "e87b0aad-daef-459c-b8fd-69991e4c615e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "image1 = torch.zeros(1,1,5,5)\n",
        "image1[0,0,2,:] = 1\n",
        "image1"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[0., 0., 0., 0., 0.],\n",
              "          [0., 0., 0., 0., 0.],\n",
              "          [1., 1., 1., 1., 1.],\n",
              "          [0., 0., 0., 0., 0.],\n",
              "          [0., 0., 0., 0., 0.]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G9yphX8nVAJK",
        "outputId": "afb6f873-b7c9-4d94-b47f-4a7656aeff14",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.imshow(image1[0,0,:,:].numpy(),interpolation='nearest',cmap=plt.cm.gray)\n",
        "plt.colorbar()\n",
        "plt.show()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qpXUp4bYVL_0",
        "outputId": "81dc71a7-31a5-4e31-c94f-83ab3a05d22a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 965
        }
      },
      "source": [
        "out1 = conv1(image1)\n",
        "for channel,image in enumerate(out1[0]):\n",
        "    plt.imshow(image.detach().numpy(), interpolation='nearest', cmap=plt.cm.gray)\n",
        "    print(image)\n",
        "    plt.title(f'channel : {channel}')\n",
        "    plt.colorbar()\n",
        "    plt.show()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "tensor([[0., 0., 0.],\n",
            "        [0., 0., 0.],\n",
            "        [0., 0., 0.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[-4., -4., -4.],\n",
            "        [ 0.,  0.,  0.],\n",
            "        [ 4.,  4.,  4.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[3., 3., 3.],\n",
            "        [3., 3., 3.],\n",
            "        [3., 3., 3.]], grad_fn=<SelectBackward>)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jWQpAvMjOwGi",
        "outputId": "85fbbcf1-f445-4fd4-ed69-ecd92ae69569",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        }
      },
      "source": [
        "image2 = torch.zeros(1,2,5,5)\n",
        "image2[0,0,2,:] = -2\n",
        "image2[0,1,2,:] = 1 \n",
        "image2"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 0.,  0.,  0.,  0.,  0.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.],\n",
              "          [-2., -2., -2., -2., -2.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.]],\n",
              "\n",
              "         [[ 0.,  0.,  0.,  0.,  0.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.],\n",
              "          [ 1.,  1.,  1.,  1.,  1.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.],\n",
              "          [ 0.,  0.,  0.,  0.,  0.]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dj0u2-C2PHGn",
        "outputId": "01aef121-f527-4f14-9be4-245856fd857c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 719
        }
      },
      "source": [
        "for channel,image in enumerate(image2[0]):\n",
        "    plt.imshow(image.detach().numpy(), interpolation='nearest', cmap=plt.cm.gray)\n",
        "    print(image)\n",
        "    plt.title(\"channel {}\".format(channel))\n",
        "    plt.colorbar()\n",
        "    plt.show()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "tensor([[ 0.,  0.,  0.,  0.,  0.],\n",
            "        [ 0.,  0.,  0.,  0.,  0.],\n",
            "        [-2., -2., -2., -2., -2.],\n",
            "        [ 0.,  0.,  0.,  0.,  0.],\n",
            "        [ 0.,  0.,  0.,  0.,  0.]])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[0., 0., 0., 0., 0.],\n",
            "        [0., 0., 0., 0., 0.],\n",
            "        [1., 1., 1., 1., 1.],\n",
            "        [0., 0., 0., 0., 0.],\n",
            "        [0., 0., 0., 0., 0.]])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RaqVO1zZVc0F"
      },
      "source": [
        "conv3 = nn.Conv2d(in_channels=2,out_channels=1,kernel_size=3)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FriXsNoXWyxI",
        "outputId": "cd52de67-a1cf-44f0-8ab7-1f572fb8cc5f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "conv3.state_dict()['weight']"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 0.0761,  0.1851,  0.1121],\n",
              "          [-0.1181, -0.2051, -0.0533],\n",
              "          [-0.1168,  0.2248,  0.2094]],\n",
              "\n",
              "         [[-0.1130,  0.1770, -0.0152],\n",
              "          [-0.1872,  0.1334,  0.0679],\n",
              "          [ 0.0073,  0.0073, -0.1927]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XLoARO_ENCI1"
      },
      "source": [
        "conv3.state_dict()['weight'][0][0] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                      [0.0,1.0,0.0],\n",
        "                                      [0.0,0.0,0.0]])\n",
        "conv3.state_dict()['weight'][0][1] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                      [0.0,-2.0,0.0],\n",
        "                                      [0.0,0.0,0.0]])\n",
        "conv3.state_dict()['bias'][:] = torch.tensor([0.0])"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k5J9A58yN9jW",
        "outputId": "fae7b281-2137-4e69-dc98-f6e6e2ef3b84",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "conv3.state_dict()['weight']"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 0.,  0.,  0.],\n",
              "          [ 0.,  1.,  0.],\n",
              "          [ 0.,  0.,  0.]],\n",
              "\n",
              "         [[ 0.,  0.,  0.],\n",
              "          [ 0., -2.,  0.],\n",
              "          [ 0.,  0.,  0.]]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xZc_vXYiOjEx",
        "outputId": "275931e7-645e-4c1f-e002-d97b6f85d628",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        }
      },
      "source": [
        "conv3(image2)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 0.,  0.,  0.],\n",
              "          [-4., -4., -4.],\n",
              "          [ 0.,  0.,  0.]]]], grad_fn=<MkldnnConvolutionBackward>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7xxdFIOOPOmy",
        "outputId": "dae026fd-347c-4c08-bd1a-57569dd94d96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 719
        }
      },
      "source": [
        "#multiple input and output channels\n",
        "image4 = torch.zeros(1,2,5,5)\n",
        "image4[0][0] = torch.ones(5,5)\n",
        "image4[0][1][2][2] = 1\n",
        "for channel,image in enumerate(image4[0]):\n",
        "  plt.imshow(image.detach().numpy(), interpolation='nearest', cmap=plt.cm.gray)\n",
        "  print(image)\n",
        "  plt.title(\"channel {}\".format(channel))\n",
        "  plt.colorbar()\n",
        "  plt.show()"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "tensor([[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"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "tensor([[0., 0., 0., 0., 0.],\n",
            "        [0., 0., 0., 0., 0.],\n",
            "        [0., 0., 1., 0., 0.],\n",
            "        [0., 0., 0., 0., 0.],\n",
            "        [0., 0., 0., 0., 0.]])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z5FlC4yGxiKW"
      },
      "source": [
        "conv4 = nn.Conv2d(in_channels=2,out_channels=3,kernel_size=3)\n",
        "conv4.state_dict()['weight'][0][0] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                                  [0.0,0.5,0.0],\n",
        "                                                  [0.0,0.0,0.0]])\n",
        "conv4.state_dict()['weight'][0][1] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                                  [0.0,0.5,0.0],\n",
        "                                                  [0.0,0.0,0.0]])\n",
        "\n",
        "conv4.state_dict()['weight'][1][0] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                                  [0.0,1.0,0.0],\n",
        "                                                  [0.0,0.0,0.0]])\n",
        "conv4.state_dict()['weight'][1][1] = torch.tensor([[0.0,0.0,0.0],\n",
        "                                                  [0.0,-1.0,0.0],\n",
        "                                                  [0.0,0.0,0.0]])\n",
        "\n",
        "conv4.state_dict()['weight'][2][0] = torch.tensor([[1.0,0.0,-1.0],\n",
        "                                                  [2.0,0.0,-2.0],\n",
        "                                                  [1.0,0.0,-1.0]])\n",
        "conv4.state_dict()['weight'][2][1] = torch.tensor([[1.0,2.0,1.0],\n",
        "                                                  [0.0,0.0,0.0],\n",
        "                                                  [-1.0,-2.0,-1.0]])\n",
        "\n",
        "conv4.state_dict()['bias'][:] = torch.tensor([0.0,0.0,0.0]) "
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "orJ6E4Exx-Ze",
        "outputId": "1deeef88-d7f9-45d2-ea5a-96a9b7692185",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        }
      },
      "source": [
        "z = conv4(image4)\n",
        "z"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[[ 0.5000,  0.5000,  0.5000],\n",
              "          [ 0.5000,  1.0000,  0.5000],\n",
              "          [ 0.5000,  0.5000,  0.5000]],\n",
              "\n",
              "         [[ 1.0000,  1.0000,  1.0000],\n",
              "          [ 1.0000,  0.0000,  1.0000],\n",
              "          [ 1.0000,  1.0000,  1.0000]],\n",
              "\n",
              "         [[-1.0000, -2.0000, -1.0000],\n",
              "          [ 0.0000,  0.0000,  0.0000],\n",
              "          [ 1.0000,  2.0000,  1.0000]]]], grad_fn=<MkldnnConvolutionBackward>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eQJ0nN2kzrFL"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}