2-1-VGG16.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyNbHYSeUIOMWZGDxAUHzZxo",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/JSJeong-me/Prompt-Engineering/blob/main/2-1-VGG16.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JX6exFWy9e-0"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers, models\n",
        "\n",
        "def VGG16(input_shape=(224, 224, 3), num_classes=1000):\n",
        "    model = models.Sequential()\n",
        "\n",
        "    # Conv Block 1\n",
        "    model.add(layers.Conv2D(64, (3, 3), padding='same', activation='relu', input_shape=input_shape))\n",
        "    model.add(layers.Conv2D(64, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.MaxPooling2D((2, 2), strides=(2, 2)))\n",
        "\n",
        "    # Conv Block 2\n",
        "    model.add(layers.Conv2D(128, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(128, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.MaxPooling2D((2, 2), strides=(2, 2)))\n",
        "\n",
        "    # Conv Block 3\n",
        "    model.add(layers.Conv2D(256, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(256, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(256, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.MaxPooling2D((2, 2), strides=(2, 2)))\n",
        "\n",
        "    # Conv Block 4\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.MaxPooling2D((2, 2), strides=(2, 2)))\n",
        "\n",
        "    # Conv Block 5\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.Conv2D(512, (3, 3), padding='same', activation='relu'))\n",
        "    model.add(layers.MaxPooling2D((2, 2), strides=(2, 2)))\n",
        "\n",
        "    # Fully Connected Layers\n",
        "    model.add(layers.Flatten())\n",
        "    model.add(layers.Dense(4096, activation='relu'))\n",
        "    model.add(layers.Dense(4096, activation='relu'))\n",
        "    model.add(layers.Dense(num_classes, activation='softmax'))\n",
        "\n",
        "    return model\n",
        "\n",
        "# 모델 생성 및 컴파일\n",
        "input_shape = (224, 224, 3)\n",
        "num_classes = 1000  # 예: ImageNet 데이터셋의 경우\n",
        "model = VGG16(input_shape=input_shape, num_classes=num_classes)\n",
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "# 모델 요약 출력\n",
        "model.summary()\n"
      ]
    }
  ]
}