8-3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.3"
    },
    "colab": {
      "name": "8-3.ipynb",
      "provenance": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Khb4cafb5-It"
      },
      "source": [
        "# 합성곱 신경망의 시각화"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gtBpMgcc5-I0"
      },
      "source": [
        "<table align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/rickiepark/hg-mldl/blob/master/8-3.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />구글 코랩에서 실행하기</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "10YURv8s5-I0"
      },
      "source": [
        "## 가중치 시각화"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q7Uqs_t65-I0"
      },
      "source": [
        "from tensorflow import keras"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FdM7Nyis5-I1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6306b925-0861-40bb-f3ee-d906faa91f7d"
      },
      "source": [
        "# 코랩에서 실행하는 경우에는 다음 명령을 실행하여 best-cnn-model.h5 파일을 다운로드받아 사용하세요.\n",
        "!wget https://github.com/rickiepark/hg-mldl/raw/master/best-cnn-model.h5"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2021-08-18 14:51:39--  https://github.com/rickiepark/hg-mldl/raw/master/best-cnn-model.h5\n",
            "Resolving github.com (github.com)... 140.82.113.3\n",
            "Connecting to github.com (github.com)|140.82.113.3|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://raw.githubusercontent.com/rickiepark/hg-mldl/master/best-cnn-model.h5 [following]\n",
            "--2021-08-18 14:51:40--  https://raw.githubusercontent.com/rickiepark/hg-mldl/master/best-cnn-model.h5\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.110.133, 185.199.109.133, 185.199.108.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.110.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4044200 (3.9M) [application/octet-stream]\n",
            "Saving to: ‘best-cnn-model.h5’\n",
            "\n",
            "best-cnn-model.h5   100%[===================>]   3.86M  --.-KB/s    in 0.05s   \n",
            "\n",
            "2021-08-18 14:51:41 (72.7 MB/s) - ‘best-cnn-model.h5’ saved [4044200/4044200]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y6Fx0cai5-I1"
      },
      "source": [
        "model = keras.models.load_model('best-cnn-model.h5')"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "tOhJnNvW5-I1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bc5fa0a3-acdc-4428-c9ea-544acd6e16ef"
      },
      "source": [
        "model.layers"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<keras.layers.convolutional.Conv2D at 0x7fe6081ac5d0>,\n",
              " <keras.layers.pooling.MaxPooling2D at 0x7fe6081aca50>,\n",
              " <keras.layers.convolutional.Conv2D at 0x7fe6081ac590>,\n",
              " <keras.layers.pooling.MaxPooling2D at 0x7fe6080fdc50>,\n",
              " <keras.layers.core.Flatten at 0x7fe608102dd0>,\n",
              " <keras.layers.core.Dense at 0x7fe6081026d0>,\n",
              " <keras.layers.core.Dropout at 0x7fe608102c50>,\n",
              " <keras.layers.core.Dense at 0x7fe608102d50>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "7e-EMS_h5-I2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b8169942-9df2-4136-bb4e-f70005ab06c0"
      },
      "source": [
        "conv = model.layers[0]\n",
        "\n",
        "print(conv.weights[0].shape, conv.weights[1].shape)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(3, 3, 1, 32) (32,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D5WNlXS55-I3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4e058ee3-517e-4734-8f72-a4cb4f8c7cfa"
      },
      "source": [
        "conv_weights = conv.weights[0].numpy()\n",
        "\n",
        "print(conv_weights.mean(), conv_weights.std())"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "-0.03802128 0.27455312\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mm_xATS95-I3"
      },
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lm37R-hD5-I3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "68810f67-0486-4c02-a71d-d44bc6c0760d"
      },
      "source": [
        "plt.hist(conv_weights.reshape(-1, 1))\n",
        "plt.xlabel('weight')\n",
        "plt.ylabel('count')\n",
        "plt.show()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "FBxjPV045-I4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        },
        "outputId": "5e83f4fb-69f6-4e5b-cb3f-81e3086dc079"
      },
      "source": [
        "fig, axs = plt.subplots(2, 16, figsize=(15,2))\n",
        "\n",
        "for i in range(2):\n",
        "    for j in range(16):\n",
        "        axs[i, j].imshow(conv_weights[:,:,0,i*16 + j], vmin=-0.5, vmax=0.5)\n",
        "        axs[i, j].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x144 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P-r4Dx2N5-I4"
      },
      "source": [
        "no_training_model = keras.Sequential()\n",
        "\n",
        "no_training_model.add(keras.layers.Conv2D(32, kernel_size=3, activation='relu', \n",
        "                                          padding='same', input_shape=(28,28,1)))"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SRz0Z0z_5-I4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cae417af-a18f-4463-9574-94123cc931f5"
      },
      "source": [
        "no_training_conv = no_training_model.layers[0]\n",
        "\n",
        "print(no_training_conv.weights[0].shape)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(3, 3, 1, 32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7FUqZL695-I4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e98160a3-4849-497f-d29c-041f17fcb6fa"
      },
      "source": [
        "no_training_weights = no_training_conv.weights[0].numpy()\n",
        "\n",
        "print(no_training_weights.mean(), no_training_weights.std())"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.0014592835 0.08036579\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DktjNm4Z5-I5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "outputId": "d25aa974-915f-4dc1-f51f-6e0bfd835614"
      },
      "source": [
        "plt.hist(no_training_weights.reshape(-1, 1))\n",
        "plt.xlabel('weight')\n",
        "plt.ylabel('count')\n",
        "plt.show()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b7QmwoUe5-I5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        },
        "outputId": "c0df75ae-5501-4dc2-b567-8664ad0cf2c5"
      },
      "source": [
        "fig, axs = plt.subplots(2, 16, figsize=(15,2))\n",
        "\n",
        "for i in range(2):\n",
        "    for j in range(16):\n",
        "        axs[i, j].imshow(no_training_weights[:,:,0,i*16 + j], vmin=-0.5, vmax=0.5)\n",
        "        axs[i, j].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x144 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GEc9KyP35-I6"
      },
      "source": [
        "## 함수형 API"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RZ27RyZJ5-I6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "71b66fba-066f-4e98-db5a-ae3c7563ad09"
      },
      "source": [
        "print(model.input)"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "KerasTensor(type_spec=TensorSpec(shape=(None, 28, 28, 1), dtype=tf.float32, name='conv2d_input'), name='conv2d_input', description=\"created by layer 'conv2d_input'\")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FfHfv9w_5-I6"
      },
      "source": [
        "conv_acti = keras.Model(model.input, model.layers[0].output)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CF6JzHj-5-I6"
      },
      "source": [
        "## 특성 맵 시각화"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TU5ENBPB5-I7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "57189673-d0ff-4d38-ffc2-086cc4954fde"
      },
      "source": [
        "(train_input, train_target), (test_input, test_target) = keras.datasets.fashion_mnist.load_data()"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "40960/29515 [=========================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "26435584/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "16384/5148 [===============================================================================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n",
            "4431872/4422102 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8-GBYsqm5-I7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "65a4d48f-367a-4f54-8176-76775b32441d"
      },
      "source": [
        "plt.imshow(train_input[0], cmap='gray_r')\n",
        "plt.show()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aJDrgyzo5-I7"
      },
      "source": [
        "inputs = train_input[0:1].reshape(-1, 28, 28, 1)/255.0\n",
        "\n",
        "feature_maps = conv_acti.predict(inputs)"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wYMHCQ0C5-I7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "98ca5d17-8590-4dfc-b4c1-98b06731cef6"
      },
      "source": [
        "print(feature_maps.shape)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1, 28, 28, 32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n9hhqasw5-I7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 460
        },
        "outputId": "2bf80584-39aa-4967-a64b-621b4e261533"
      },
      "source": [
        "fig, axs = plt.subplots(4, 8, figsize=(15,8))\n",
        "\n",
        "for i in range(4):\n",
        "    for j in range(8):\n",
        "        axs[i, j].imshow(feature_maps[0,:,:,i*8 + j])\n",
        "        axs[i, j].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x576 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RBCPiHm_5-I8"
      },
      "source": [
        "conv2_acti = keras.Model(model.input, model.layers[2].output)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n5U3pxZc5-I8"
      },
      "source": [
        "feature_maps = conv2_acti.predict(train_input[0:1].reshape(-1, 28, 28, 1)/255.0)"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9QcCxjOf5-I8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "72e976df-3572-476d-b136-07279a68d9b7"
      },
      "source": [
        "print(feature_maps.shape)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1, 14, 14, 64)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6xc2oG6D5-I8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "outputId": "964e99a6-334d-4f6d-abac-7f177c1b13a0"
      },
      "source": [
        "fig, axs = plt.subplots(8, 8, figsize=(12,12))\n",
        "\n",
        "for i in range(8):\n",
        "    for j in range(8):\n",
        "        axs[i, j].imshow(feature_maps[0,:,:,i*8 + j])\n",
        "        axs[i, j].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x864 with 64 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}