{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Using TensorFlow backend.\n" ] }, { "data": { "text/plain": [ "'2.2.4'" ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import keras\n", "keras.__version__" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 5.2 - 소규모 데이터셋에서 컨브넷 사용하기\n", "\n", "이 노트북은 [케라스 창시자에게 배우는 딥러닝](https://tensorflow.blog/케라스-창시자에게-배우는-딥러닝/) 책의 5장 2절의 코드 예제입니다. 책에는 더 많은 내용과 그림이 있습니다. 이 노트북에는 소스 코드에 관련된 설명만 포함합니다. 이 노트북의 설명은 케라스 버전 2.2.2에 맞추어져 있습니다. 케라스 최신 버전이 릴리스되면 노트북을 다시 테스트하기 때문에 설명과 코드의 결과가 조금 다를 수 있습니다.\n", "\n", "## 소규모 데이터셋에서 밑바닥부터 컨브넷을 훈련하기\n", "\n", "매우 적은 데이터를 사용해 이미지 분류 모델을 훈련하는 일은 흔한 경우입니다. 여러분이 전문적인 컴퓨터 비전 작업을 한다면 실제로 이런 상황을 마주치게 될 가능성이 높습니다.\n", "\n", "보통 '적은' 샘플이란 수백 개에서 수만 개 사이를 의미합니다. 실용적인 예제로 4,000개의 강아지와 고양이 사진(2,000개는 강아지, 2,000개는 고양이)으로 구성된 데이터셋에서 강아지와 고양이 이미지를 분류해 보겠습니다. 훈련을 위해 2,000개의 사진을 사용하고 검증과 테스트에 각각 1,000개의 사진을 사용하겠습니다.\n", "\n", "이 절에서 문제를 해결하기 위해 기본적인 전략 하나를 살펴볼 것입니다. 보유한 소규모 데이터셋을 사용해 처음부터 새로운 모델을 훈련하는 것입니다. 2,000개의 훈련 샘플에서 작은 컨브넷을 어떤 규제 방법도 사용하지 않고 훈련하여 기준이 되는 기본 성능을 만들겠습니다. 이 방법은 71%의 분류 정확도를 달성할 것입니다. 이 방법의 주요 이슈는 과대적합이 될 것입니다. 그다음 컴퓨터 비전에서 과대적합을 줄이기 위한 강력한 방법인 데이터 증식을 소개하겠습니다. 데이터 증식을 통해 네트워크의 성능을 82% 정확도로 향상시킬 것입니다.\n", "\n", "다음 절에서 작은 데이터셋에 딥러닝을 적용하기 위한 핵심적인 기술 두 가지를 살펴보겠습니다. 사전 훈련된 네트워크로 특성을 추출하는 것(90%에서 96%의 정확도를 얻게 됩니다)과 사전 훈련된 네트워크를 세밀하게 튜닝하는 것입니다(최종 모델은 97% 정확도를 얻을 것입니다). 이런 세 가지 전략(처음부터 작은 모델을 훈련하기, 사전 훈련된 모델을 사용해 특성 추출하기, 사전 훈련된 모델을 세밀하게 튜닝하기)은 작은 데이터셋에서 이미지 분류 문제를 수행할 때 여러분의 도구 상자에 포함되어 있어야 합니다." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 작은 데이터셋 문제에서 딥러닝의 타당성\n", "\n", "딥러닝은 데이터가 풍부할 때만 작동한다는 말을 이따금 듣습니다. 부분적으로는 맞습니다. 딥러닝의 근본적인 특징은 훈련 데이터에서 특성 공학의 수작업 없이 흥미로운 특성을 찾을 수 있는 것입니다. 이는 훈련 샘플이 많아야만 가능합니다. 입력 샘플이 이미지와 같이 매우 고차원인 문제에서는 특히 그렇습니다.\n", "\n", "하지만 많은 샘플이 의미하는 것은 상대적입니다. 우선 훈련하려는 네트워크의 크기와 깊이에 상대적입니다. 복잡한 문제를 푸는 컨브넷을 수십 개의 샘플만을 사용해서 훈련하는 것은 불가능합니다. 하지만 모델이 작고 규제가 잘 되어 있으며 간단한 작업이라면 수백 개의 샘플로도 충분할 수 있습니다. 컨브넷은 지역적이고 평행 이동으로 변하지 않는 특성을 학습하기 때문에 지각에 관한 문제에서 매우 효율적으로 데이터를 사용합니다. 매우 작은 이미지 데이터셋에서 어떤 종류의 특성 공학을 사용하지 않고 컨브넷을 처음부터 훈련해도 납득할 만한 결과를 만들 수 있습니다. 이 절에서 실제로 이런 결과를 보게 될 것입니다.\n", "\n", "거기에 더해 딥러닝 모델은 태생적으로 매우 다목적입니다. 말하자면 대규모 데이터셋에서 훈련시킨 이미지 분류 모델이나 스피치-투-텍스트 모델을 조금만 변경해서 완전히 다른 문제에 재사용할 수 있습니다. 특히 컴퓨터 비전에서는 (보통 ImageNet 데이터셋에서 훈련된) 사전 훈련된 모델들이 다운로드받을 수 있도록 많이 공개되어 있어서 매우 적은 데이터에서 강력한 비전 모델을 만드는데 사용할 수 있습니다. 바로 다음 절에서 우리가 해볼 것입니다.\n", "\n", "먼저 데이터를 구하는 것부터 시작해 보죠." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 데이터 내려받기\n", "\n", "여기서 사용할 강아지 vs. 고양이 데이터셋은 케라스에 포함되어 있지 않습니다. 컨브넷이 주류가 되기 전인 2013년 후반에 캐글에서 컴퓨터 비전 경연 대회의 일환으로 이 데이터셋을 만들었습니다. 원본 데이터셋을 `https://www.kaggle.com/c/dogs-vs-cats/data`에서 내려받을 수 있습니다(캐글 계정이 없다면 하나 만들어야 하지만 계정을 만드는 과정은 간단합니다). 실습의 편의를 위해서 번역서의 깃허브에는 이 데이터셋을 미리 다운로드하여 포함하였습니다.\n", "\n", "이 사진들은 중간 정도의 해상도를 가진 컬러 JPEG 파일입니다. 다음이 몇 개 샘플입니다:\n", "\n", "![cats_vs_dogs_samples](https://s3.amazonaws.com/book.keras.io/img/ch5/cats_vs_dogs_samples.jpg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "당연히 2013년 강아지 vs. 고양이 캐글 경연은 컨브넷을 사용한 참가자가 우승하였습니다. 최고 성능은 95%의 정확도를 달성했습니다. 이 예제를 가지고 (다음 절에서) 참가자들이 사용했던 데이터의 10%보다 적은 양으로 모델을 훈련하고도 이와 아주 근접한 정확도를 달성해 보겠습니다.\n", "\n", "이 데이터셋은 25,000개의 강아지와 고양이 이미지(클래스마다 12,500개)를 담고 있고 (압축해서) 543MB 크기입니다. 다운로드하고 압축을 해제한 후 세 개의 서브셋이 들어 있는 새로운 데이터셋을 만들 것입니다. 클래스마다 1,000개의 샘플로 이루어진 훈련 세트, 클래스마다 500개의 샘플로 이루어진 검증 세트, 클래스마다 500개의 샘플로 이루어진 테스트 세트입니다.\n", "\n", "다음은 이를 처리하는 코드입니다:" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import os, shutil" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "# 원본 데이터셋을 압축 해제한 디렉터리 경로\n", "original_dataset_dir = './datasets/cats_and_dogs/train'\n", "\n", "# 소규모 데이터셋을 저장할 디렉터리\n", "base_dir = './datasets/cats_and_dogs_small'\n", "if os.path.exists(base_dir): # 반복적인 실행을 위해 디렉토리를 삭제합니다.\n", " shutil.rmtree(base_dir) # 이 코드는 책에 포함되어 있지 않습니다.\n", "os.mkdir(base_dir)\n", "\n", "# 훈련, 검증, 테스트 분할을 위한 디렉터리\n", "train_dir = os.path.join(base_dir, 'train')\n", "os.mkdir(train_dir)\n", "validation_dir = os.path.join(base_dir, 'validation')\n", "os.mkdir(validation_dir)\n", "test_dir = os.path.join(base_dir, 'test')\n", "os.mkdir(test_dir)\n", "\n", "# 훈련용 고양이 사진 디렉터리\n", "train_cats_dir = os.path.join(train_dir, 'cats')\n", "os.mkdir(train_cats_dir)\n", "\n", "# 훈련용 강아지 사진 디렉터리\n", "train_dogs_dir = os.path.join(train_dir, 'dogs')\n", "os.mkdir(train_dogs_dir)\n", "\n", "# 검증용 고양이 사진 디렉터리\n", "validation_cats_dir = os.path.join(validation_dir, 'cats')\n", "os.mkdir(validation_cats_dir)\n", "\n", "# 검증용 강아지 사진 디렉터리\n", "validation_dogs_dir = os.path.join(validation_dir, 'dogs')\n", "os.mkdir(validation_dogs_dir)\n", "\n", "# 테스트용 고양이 사진 디렉터리\n", "test_cats_dir = os.path.join(test_dir, 'cats')\n", "os.mkdir(test_cats_dir)\n", "\n", "# 테스트용 강아지 사진 디렉터리\n", "test_dogs_dir = os.path.join(test_dir, 'dogs')\n", "os.mkdir(test_dogs_dir)\n", "\n", "# 처음 1,000개의 고양이 이미지를 train_cats_dir에 복사합니다\n", "fnames = ['cat.{}.jpg'.format(i) for i in range(1000)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(train_cats_dir, fname)\n", " shutil.copyfile(src, dst)\n", "\n", "# 다음 500개 고양이 이미지를 validation_cats_dir에 복사합니다\n", "fnames = ['cat.{}.jpg'.format(i) for i in range(1000, 1500)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(validation_cats_dir, fname)\n", " shutil.copyfile(src, dst)\n", " \n", "# 다음 500개 고양이 이미지를 test_cats_dir에 복사합니다\n", "fnames = ['cat.{}.jpg'.format(i) for i in range(1500, 2000)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(test_cats_dir, fname)\n", " shutil.copyfile(src, dst)\n", " \n", "# 처음 1,000개의 강아지 이미지를 train_dogs_dir에 복사합니다\n", "fnames = ['dog.{}.jpg'.format(i) for i in range(1000)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(train_dogs_dir, fname)\n", " shutil.copyfile(src, dst)\n", " \n", "# 다음 500개 강아지 이미지를 validation_dogs_dir에 복사합니다\n", "fnames = ['dog.{}.jpg'.format(i) for i in range(1000, 1500)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(validation_dogs_dir, fname)\n", " shutil.copyfile(src, dst)\n", " \n", "# 다음 500개 강아지 이미지를 test_dogs_dir에 복사합니다\n", "fnames = ['dog.{}.jpg'.format(i) for i in range(1500, 2000)]\n", "for fname in fnames:\n", " src = os.path.join(original_dataset_dir, fname)\n", " dst = os.path.join(test_dogs_dir, fname)\n", " shutil.copyfile(src, dst)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "복사가 잘 되었는지 확인하기 위해 각 분할(훈련/검증/테스트)에 들어 있는 사진의 개수를 카운트해 보죠:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "훈련용 고양이 이미지 전체 개수: 1000\n" ] } ], "source": [ "print('훈련용 고양이 이미지 전체 개수:', len(os.listdir(train_cats_dir)))" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "훈련용 강아지 이미지 전체 개수: 1000\n" ] } ], "source": [ "print('훈련용 강아지 이미지 전체 개수:', len(os.listdir(train_dogs_dir)))" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "검증용 고양이 이미지 전체 개수: 500\n" ] } ], "source": [ "print('검증용 고양이 이미지 전체 개수:', len(os.listdir(validation_cats_dir)))" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "검증용 강아지 이미지 전체 개수: 500\n" ] } ], "source": [ "print('검증용 강아지 이미지 전체 개수:', len(os.listdir(validation_dogs_dir)))" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "테스트용 고양이 이미지 전체 개수: 500\n" ] } ], "source": [ "print('테스트용 고양이 이미지 전체 개수:', len(os.listdir(test_cats_dir)))" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "테스트용 강아지 이미지 전체 개수: 500\n" ] } ], "source": [ "print('테스트용 강아지 이미지 전체 개수:', len(os.listdir(test_dogs_dir)))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "이제 2,000개의 훈련 이미지, 1,000개의 검증 이미지, 1,000개의 테스트 이미지가 준비되었습니다. 분할된 각 데이터는 클래마다 동일한 개수의 샘플을 포함합니다. 균형잡힌 이진 분류 문제이므로 정확도를 사용해 성공을 측정하겠습니다." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 네트워크 구성하기\n", "\n", "이전 예제에서 MNIST를 위해 간단한 컨브넷을 만들었습니다. 이제 컨브넷에 친숙해졌을 것입니다. 여기서 사용할 구조도 일반적으로 동일합니다. `Conv2D`(`relu` 활성화 함수 사용)와 `MaxPooling2D` 층을 번갈아 쌓은 컨브넷을 만들겠습니다.\n", "\n", "이전보다 이미지가 크고 복잡한 문제이기 때문에 네트워크를 좀 더 크게 만들겠습니다. `Conv2D` + `MaxPooling2D` 단계를 하나 더 추가합니다. 이렇게 하면 네트워크의 용량을 늘리고 `Flatten` 층의 크기가 너무 커지지 않도록 특성 맵의 크기를 줄일 수 있습니다. 150 × 150 크기(임의로 선택한 것입니다)의 입력으로 시작해서 `Flatten` 층 이전에 7 × 7 크기의 특성 맵으로 줄어듭니다.\n", "\n", "특성 맵의 깊이는 네트워크에서 점진적으로 증가하지만(32에서 128까지), 특성 맵의 크기는 감소합니다(150 × 150에서 7 × 7까지). 이는 거의 모든 컨브넷에서 볼 수 있는 전형적인 패턴입니다.\n", "\n", "이진 분류 문제이므로 네트워크는 하나의 유닛(크기가 1인 `Dense` 층)과 `sigmoid` 활성화 함수로 끝납니다. 이 유닛은 한 클래스에 대한 확률을 인코딩할 것입니다." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "from keras import layers\n", "from keras import models\n", "\n", "model = models.Sequential()\n", "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n", " input_shape=(150, 150, 3)))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Flatten())\n", "model.add(layers.Dense(512, activation='relu'))\n", "model.add(layers.Dense(1, activation='sigmoid'))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "층들을 거치면서 특성 맵의 차원이 어떻게 변하는지 살펴보겠습니다:" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "_________________________________________________________________\n", "Layer (type) Output Shape Param # \n", "=================================================================\n", "conv2d_1 (Conv2D) (None, 148, 148, 32) 896 \n", "_________________________________________________________________\n", "max_pooling2d_1 (MaxPooling2 (None, 74, 74, 32) 0 \n", "_________________________________________________________________\n", "conv2d_2 (Conv2D) (None, 72, 72, 64) 18496 \n", "_________________________________________________________________\n", "max_pooling2d_2 (MaxPooling2 (None, 36, 36, 64) 0 \n", "_________________________________________________________________\n", "conv2d_3 (Conv2D) (None, 34, 34, 128) 73856 \n", "_________________________________________________________________\n", "max_pooling2d_3 (MaxPooling2 (None, 17, 17, 128) 0 \n", "_________________________________________________________________\n", "conv2d_4 (Conv2D) (None, 15, 15, 128) 147584 \n", "_________________________________________________________________\n", "max_pooling2d_4 (MaxPooling2 (None, 7, 7, 128) 0 \n", "_________________________________________________________________\n", "flatten_1 (Flatten) (None, 6272) 0 \n", "_________________________________________________________________\n", "dense_1 (Dense) (None, 512) 3211776 \n", "_________________________________________________________________\n", "dense_2 (Dense) (None, 1) 513 \n", "=================================================================\n", "Total params: 3,453,121\n", "Trainable params: 3,453,121\n", "Non-trainable params: 0\n", "_________________________________________________________________\n" ] } ], "source": [ "model.summary()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "컴파일 단계에서 이전과 같이 `RMSprop` 옵티마이저를 선택하겠습니다. 네트워크의 마지막이 하나의 시그모이드 유닛이기 때문에 이진 크로스엔트로피(binary crossentropy)를 손실로 사용합니다(4장 5절에서 다양한 경우에 사용할 수 있는 손실 함수 목록을 볼 수 있습니다)." ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [], "source": [ "from keras import optimizers\n", "\n", "model.compile(loss='binary_crossentropy',\n", " optimizer=optimizers.RMSprop(lr=1e-4),\n", " metrics=['acc'])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 데이터 전처리\n", "\n", "데이터는 네트워크에 주입되기 전에 부동 소수 타입의 텐서로 적절하게 전처리되어 있어야 합니다. 지금은 데이터가 JPEG 파일로 되어 있으므로 네트워크에 주입하려면 대략 다음 과정을 따릅니다.\n", "\n", "1.\t사진 파일을 읽습니다.\n", "2.\tJPEG 콘텐츠를 RGB 픽셀 값으로 디코딩합니다.\n", "3.\t그다음 부동 소수 타입의 텐서로 변환합니다.\n", "4.\t픽셀 값(0에서 255 사이)의 스케일을 [0, 1] 사이로 조정합니다(신경망은 작은 입력 값을 선호합니다).\n", "\n", "좀 복잡하게 보일 수 있지만 다행히 케라스는 이런 단계를 자동으로 처리하는 유틸리티를 가지고 있습니다. 케라스는 `keras.preprocessing.image`에 이미지 처리를 위한 헬퍼 도구들을 가지고 있습니다. 특히 `ImageDataGenerator` 클래스는 디스크에 있는 이미지 파일을 전처리된 배치 텐서로 자동으로 바꾸어주는 파이썬 제너레이터를 만들어 줍니다. 이 클래스를 사용해 보겠습니다." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Found 2000 images belonging to 2 classes.\n", "Found 1000 images belonging to 2 classes.\n" ] } ], "source": [ "from keras.preprocessing.image import ImageDataGenerator\n", "\n", "# 모든 이미지를 1/255로 스케일을 조정합니다\n", "train_datagen = ImageDataGenerator(rescale=1./255)\n", "test_datagen = ImageDataGenerator(rescale=1./255)\n", "\n", "train_generator = train_datagen.flow_from_directory(\n", " # 타깃 디렉터리\n", " train_dir,\n", " # 모든 이미지를 150 × 150 크기로 바꿉니다\n", " target_size=(150, 150),\n", " batch_size=20,\n", " # binary_crossentropy 손실을 사용하기 때문에 이진 레이블이 필요합니다\n", " class_mode='binary')\n", "\n", "validation_generator = test_datagen.flow_from_directory(\n", " validation_dir,\n", " target_size=(150, 150),\n", " batch_size=20,\n", " class_mode='binary')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "이 제너레이터의 출력 하나를 살펴보죠. 이 출력은 150 × 150 RGB 이미지의 배치(`(20, 150, 150, 3)` 크기)와 이진 레이블의 배치(`(20,)` 크기)입니다. 각 배치에는 20개의 샘플(배치 크기)이 있습니다. 제너레이터는 이 배치를 무한정 만들어 냅니다. 타깃 폴더에 있는 이미지를 끝없이 반복합니다. 따라서 반복 루프안의 어디에선가 `break` 문을 사용해야 합니다." ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "배치 데이터 크기: (20, 150, 150, 3)\n", "배치 레이블 크기: (20,)\n" ] } ], "source": [ "for data_batch, labels_batch in train_generator:\n", " print('배치 데이터 크기:', data_batch.shape)\n", " print('배치 레이블 크기:', labels_batch.shape)\n", " break" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "제너레이터를 사용한 데이터에 모델을 훈련시켜 보겠습니다. `fit_generator` 메서드는 `fit` 메서드와 동일하되 데이터 제너레이터를 사용할 수 있습니다. 이 메서드는 첫 번째 매개변수로 입력과 타깃의 배치를 끝없이 반환하는 파이썬 제너레이터를 기대합니다. 데이터가 끝없이 생성되기 때문에 케라스 모델에 하나의 에포크를 정의하기 위해 제너레이터로부터 얼마나 많은 샘플을 뽑을 것인지 알려 주어야 합니다. `steps_per_epoch` 매개변수에서 이를 설정합니다. 제너레이터로부터 `steps_per_epoch` 개의 배치만큼 뽑은 다음, 즉 `steps_per_epoch` 횟수만큼 경사 하강법 단계를 실행한 다음에 훈련 프로세스는 다음 에포크로 넘어갑니다. 여기서는 20개의 샘플이 하나의 배치이므로 2,000개의 샘플을 모두 처리할 때까지 100개의 배치를 뽑을 것입니다.\n", "\n", "`fit_generator`를 사용할 때 `fit` 메서드와 마찬가지로 `validation_data` 매개변수를 전달할 수 있습니다. 이 매개변수에는 데이터 제너레이터도 가능하지만 넘파이 배열의 튜플도 가능합니다. `validation_data`로 제너레이터를 전달하면 검증 데이터의 배치를 끝없이 반환합니다. 따라서 검증 데이터 제너레이터에서 얼마나 많은 배치를 추출하여 평가할지 `validation_steps` 매개변수에 지정해야 합니다." ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/30\n", "100/100 [==============================] - 6s 65ms/step - loss: 0.6896 - acc: 0.5355 - val_loss: 0.6735 - val_acc: 0.6030\n", "Epoch 2/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.6615 - acc: 0.6075 - val_loss: 0.6453 - val_acc: 0.6320\n", "Epoch 3/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.6286 - acc: 0.6385 - val_loss: 0.6287 - val_acc: 0.6390\n", "Epoch 4/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.5835 - acc: 0.6910 - val_loss: 0.7105 - val_acc: 0.5880\n", "Epoch 5/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.5435 - acc: 0.7350 - val_loss: 0.5674 - val_acc: 0.7060\n", "Epoch 6/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.5024 - acc: 0.7525 - val_loss: 0.6188 - val_acc: 0.6580\n", "Epoch 7/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.4699 - acc: 0.7690 - val_loss: 0.5425 - val_acc: 0.7140\n", "Epoch 8/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.4342 - acc: 0.8030 - val_loss: 0.5443 - val_acc: 0.7280\n", "Epoch 9/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.4187 - acc: 0.8090 - val_loss: 0.5560 - val_acc: 0.7130\n", "Epoch 10/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.3819 - acc: 0.8245 - val_loss: 0.5522 - val_acc: 0.7350\n", "Epoch 11/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.3544 - acc: 0.8415 - val_loss: 0.6046 - val_acc: 0.7190\n", "Epoch 12/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.3365 - acc: 0.8530 - val_loss: 0.6053 - val_acc: 0.7260\n", "Epoch 13/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.3155 - acc: 0.8660 - val_loss: 0.5915 - val_acc: 0.7210\n", "Epoch 14/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.2844 - acc: 0.8765 - val_loss: 0.5899 - val_acc: 0.7260\n", "Epoch 15/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.2605 - acc: 0.8995 - val_loss: 0.5816 - val_acc: 0.7400\n", "Epoch 16/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.2480 - acc: 0.9010 - val_loss: 0.5851 - val_acc: 0.7360\n", "Epoch 17/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.2240 - acc: 0.9130 - val_loss: 0.8294 - val_acc: 0.7050\n", "Epoch 18/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.2040 - acc: 0.9195 - val_loss: 0.6895 - val_acc: 0.7260\n", "Epoch 19/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.1753 - acc: 0.9335 - val_loss: 0.6405 - val_acc: 0.7500\n", "Epoch 20/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.1629 - acc: 0.9425 - val_loss: 0.6902 - val_acc: 0.7270\n", "Epoch 21/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.1411 - acc: 0.9535 - val_loss: 0.7392 - val_acc: 0.7360\n", "Epoch 22/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.1354 - acc: 0.9500 - val_loss: 0.7328 - val_acc: 0.7260\n", "Epoch 23/30\n", "100/100 [==============================] - 5s 50ms/step - loss: 0.1082 - acc: 0.9685 - val_loss: 0.7835 - val_acc: 0.7360\n", "Epoch 24/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0949 - acc: 0.9710 - val_loss: 0.7963 - val_acc: 0.7250\n", "Epoch 25/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0811 - acc: 0.9770 - val_loss: 0.9073 - val_acc: 0.7100\n", "Epoch 26/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0747 - acc: 0.9775 - val_loss: 0.8273 - val_acc: 0.7370\n", "Epoch 27/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0618 - acc: 0.9835 - val_loss: 0.9788 - val_acc: 0.7240\n", "Epoch 28/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0535 - acc: 0.9865 - val_loss: 1.0135 - val_acc: 0.7270\n", "Epoch 29/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0475 - acc: 0.9870 - val_loss: 0.9734 - val_acc: 0.7440\n", "Epoch 30/30\n", "100/100 [==============================] - 5s 49ms/step - loss: 0.0396 - acc: 0.9895 - val_loss: 0.9951 - val_acc: 0.7360\n" ] } ], "source": [ "history = model.fit_generator(\n", " train_generator,\n", " steps_per_epoch=100,\n", " epochs=30,\n", " validation_data=validation_generator,\n", " validation_steps=50)" ] }, { "cell_type": "markdown", "metadata": { "collapsed": true }, "source": [ "훈련이 끝나면 항상 모델을 저장하는 것이 좋은 습관입니다:" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": [ "model.save('cats_and_dogs_small_1.h5')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "훈련 데이터와 검증 데이터에 대한 모델의 손실과 정확도를 그래프로 나타내 보겠습니다:" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "import matplotlib.pyplot as plt" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "acc = history.history['acc']\n", "val_acc = history.history['val_acc']\n", "loss = history.history['loss']\n", "val_loss = history.history['val_loss']\n", "\n", "epochs = range(len(acc))\n", "\n", "plt.plot(epochs, acc, 'bo', label='Training acc')\n", "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n", "plt.title('Training and validation accuracy')\n", "plt.legend()\n", "\n", "plt.figure()\n", "\n", "plt.plot(epochs, loss, 'bo', label='Training loss')\n", "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n", "plt.title('Training and validation loss')\n", "plt.legend()\n", "\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "이 그래프는 과대적합의 특성을 보여줍니다. 훈련 정확도가 시간이 지남에 따라 선형적으로 증가해서 거의 100%에 도달합니다. 반면 검증 정확도는 70-72%에서 멈추었습니다. 검증 손실은 다섯 번의 에포크만에 최솟값에 다다른 이후에 더 이상 진전되지 않았습니다. 반면 훈련 손실은 거의 0에 도달할 때까지 선형적으로 계속 감소합니다.\n", "\n", "비교적 훈련 샘플의 수(2,000개)가 적기 때문에 과대적합이 가장 중요한 문제입니다. 드롭아웃이나 가중치 감소(L2 규제)와 같은 과대적합을 감소시킬 수 있는 여러 가지 기법들을 배웠습니다. 여기에서는 컴퓨터 비전에 특화되어 있어서 딥러닝으로 이미지를 다룰 때 매우 일반적으로 사용되는 새로운 방법인 데이터 증식을 시도해 보겠습니다." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 데이터 증식 사용하기\n", "\n", "과대적합은 학습할 샘플이 너무 적어 새로운 데이터에 일반화할 수 있는 모델을 훈련시킬 수 없기 때문에 발생합니다. 무한히 많은 데이터가 주어지면 데이터 분포의 모든 가능한 측면을 모델이 학습할 수 있을 것입니다. 데이터 증식은 기존의 훈련 샘플로부터 더 많은 훈련 데이터를 생성하는 방법입니다. 이 방법은 그럴듯한 이미지를 생성하도록 여러 가지 랜덤한 변환을 적용하여 샘플을 늘립니다. 훈련 시에 모델이 정확히 같은 데이터를 두 번 만나지 않도록 하는 것이 목표입니다. 모델이 데이터의 여러 측면을 학습하면 일반화에 도움이 될 것입니다.\n", "\n", "케라스에서는 `ImageDataGenerator`가 읽은 이미지에 여러 종류의 랜덤 변환을 적용하도록 설정할 수 있습니다. 예제를 먼저 만들어 보죠:" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [], "source": [ "datagen = ImageDataGenerator(\n", " rotation_range=40,\n", " width_shift_range=0.2,\n", " height_shift_range=0.2,\n", " shear_range=0.2,\n", " zoom_range=0.2,\n", " horizontal_flip=True,\n", " fill_mode='nearest')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "추가적인 매개변수가 몇 개 더 있습니다(케라스 문서를 참고하세요). 이 코드를 간단히 살펴보죠.\n", "\n", "* `rotation_range`는 랜덤하게 사진을 회전시킬 각도 범위입니다(0-180 사이).\n", "* `width_shift_range`와 `height_shift_range`는 사진을 수평과 수직으로 랜덤하게 평행 이동시킬 범위입니다(전체 넓이와 높이에 대한 비율).\n", "* `shear_range`는 랜덤하게 전단 변환을 적용할 각도 범위입니다.\n", "* `zoom_range`는 랜덤하게 사진을 확대할 범위입니다.\n", "* `horizontal_flip`은 랜덤하게 이미지를 수평으로 뒤집습니다. 수평 대칭을 가정할 수 있을 때 사용합니다(예를 들어, 풍경/인물 사진).\n", "* `fill_mode`는 회전이나 가로/세로 이동으로 인해 새롭게 생성해야 할 픽셀을 채울 전략입니다.\n", "\n", "증식된 이미지 샘플을 살펴보죠:" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# 이미지 전처리 유틸리티 모듈\n", "from keras.preprocessing import image\n", "\n", "fnames = sorted([os.path.join(train_cats_dir, fname) for fname in os.listdir(train_cats_dir)])\n", "\n", "# 증식할 이미지 선택합니다\n", "img_path = fnames[3]\n", "\n", "# 이미지를 읽고 크기를 변경합니다\n", "img = image.load_img(img_path, target_size=(150, 150))\n", "\n", "# (150, 150, 3) 크기의 넘파이 배열로 변환합니다\n", "x = image.img_to_array(img)\n", "\n", "# (1, 150, 150, 3) 크기로 변환합니다\n", "x = x.reshape((1,) + x.shape)\n", "\n", "# flow() 메서드는 랜덤하게 변환된 이미지의 배치를 생성합니다.\n", "# 무한 반복되기 때문에 어느 지점에서 중지해야 합니다!\n", "i = 0\n", "for batch in datagen.flow(x, batch_size=1):\n", " plt.figure(i)\n", " imgplot = plt.imshow(image.array_to_img(batch[0]))\n", " i += 1\n", " if i % 4 == 0:\n", " break\n", "\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "데이터 증식을 사용하여 새로운 네트워크를 훈련시킬 때 네트워크에 같은 입력 데이터가 두 번 주입되지 않습니다. 하지만 적은 수의 원본 이미지에서 만들어졌기 때문에 여전히 입력 데이터들 사이에 상호 연관성이 큽니다. 즉, 새로운 정보를 만들어낼 수 없고 단지 기존 정보의 재조합만 가능합니다. 그렇기 때문에 완전히 과대적합을 제거하기에 충분하지 않을 수 있습니다. 과대적합을 더 억제하기 위해 완전 연결 분류기 직전에 `Dropout` 층을 추가하겠습니다:" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "model = models.Sequential()\n", "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n", " input_shape=(150, 150, 3)))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n", "model.add(layers.MaxPooling2D((2, 2)))\n", "model.add(layers.Flatten())\n", "model.add(layers.Dropout(0.5))\n", "model.add(layers.Dense(512, activation='relu'))\n", "model.add(layers.Dense(1, activation='sigmoid'))\n", "\n", "model.compile(loss='binary_crossentropy',\n", " optimizer=optimizers.RMSprop(lr=1e-4),\n", " metrics=['acc'])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "데이터 증식과 드롭아웃을 사용하여 이 네트워크를 훈련시켜 봅시다:" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Found 2000 images belonging to 2 classes.\n", "Found 1000 images belonging to 2 classes.\n", "Epoch 1/100\n", "100/100 [==============================] - 18s 178ms/step - loss: 0.6883 - acc: 0.5344 - val_loss: 0.6735 - val_acc: 0.5470\n", "Epoch 2/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.6705 - acc: 0.5850 - val_loss: 0.6654 - val_acc: 0.5683\n", "Epoch 3/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.6566 - acc: 0.6031 - val_loss: 0.6366 - val_acc: 0.6250\n", "Epoch 4/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.6311 - acc: 0.6375 - val_loss: 0.6187 - val_acc: 0.6411\n", "Epoch 5/100\n", "100/100 [==============================] - 16s 162ms/step - loss: 0.6194 - acc: 0.6559 - val_loss: 0.5897 - val_acc: 0.6764\n", "Epoch 6/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.6006 - acc: 0.6659 - val_loss: 0.5903 - val_acc: 0.6727\n", "Epoch 7/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5963 - acc: 0.6794 - val_loss: 0.5603 - val_acc: 0.6923\n", "Epoch 8/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5876 - acc: 0.6838 - val_loss: 0.5624 - val_acc: 0.6939\n", "Epoch 9/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5903 - acc: 0.6828 - val_loss: 0.5478 - val_acc: 0.7320\n", "Epoch 10/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5795 - acc: 0.6994 - val_loss: 0.5505 - val_acc: 0.7195\n", "Epoch 11/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5707 - acc: 0.6941 - val_loss: 0.5598 - val_acc: 0.7049\n", "Epoch 12/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5629 - acc: 0.7056 - val_loss: 0.5625 - val_acc: 0.7037\n", "Epoch 13/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5596 - acc: 0.7109 - val_loss: 0.5366 - val_acc: 0.7204\n", "Epoch 14/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5463 - acc: 0.7247 - val_loss: 0.5738 - val_acc: 0.6904\n", "Epoch 15/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5587 - acc: 0.7094 - val_loss: 0.5268 - val_acc: 0.7326\n", "Epoch 16/100\n", "100/100 [==============================] - 16s 159ms/step - loss: 0.5455 - acc: 0.7181 - val_loss: 0.5201 - val_acc: 0.7326\n", "Epoch 17/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5453 - acc: 0.7191 - val_loss: 0.5151 - val_acc: 0.7468\n", "Epoch 18/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5321 - acc: 0.7331 - val_loss: 0.5355 - val_acc: 0.7410\n", "Epoch 19/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5284 - acc: 0.7381 - val_loss: 0.5107 - val_acc: 0.7525\n", "Epoch 20/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.5345 - acc: 0.7325 - val_loss: 0.5092 - val_acc: 0.7416\n", "Epoch 21/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5165 - acc: 0.7431 - val_loss: 0.5267 - val_acc: 0.7424\n", "Epoch 22/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5135 - acc: 0.7491 - val_loss: 0.4926 - val_acc: 0.7558\n", "Epoch 23/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5108 - acc: 0.7487 - val_loss: 0.5313 - val_acc: 0.7449\n", "Epoch 24/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5068 - acc: 0.7510 - val_loss: 0.4802 - val_acc: 0.7751\n", "Epoch 25/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5168 - acc: 0.7431 - val_loss: 0.4865 - val_acc: 0.7539\n", "Epoch 26/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4987 - acc: 0.7556 - val_loss: 0.4872 - val_acc: 0.7627\n", "Epoch 27/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.5069 - acc: 0.7450 - val_loss: 0.5113 - val_acc: 0.7584\n", "Epoch 28/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4986 - acc: 0.7606 - val_loss: 0.5034 - val_acc: 0.7405\n", "Epoch 29/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4893 - acc: 0.7562 - val_loss: 0.5071 - val_acc: 0.7610\n", "Epoch 30/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4811 - acc: 0.7703 - val_loss: 0.5057 - val_acc: 0.7563\n", "Epoch 31/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4894 - acc: 0.7659 - val_loss: 0.5010 - val_acc: 0.7519\n", "Epoch 32/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4786 - acc: 0.7734 - val_loss: 0.4811 - val_acc: 0.7661\n", "Epoch 33/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4683 - acc: 0.7725 - val_loss: 0.5323 - val_acc: 0.7303\n", "Epoch 34/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4721 - acc: 0.7759 - val_loss: 0.4552 - val_acc: 0.7668\n", "Epoch 35/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4657 - acc: 0.7706 - val_loss: 0.4771 - val_acc: 0.7608\n", "Epoch 36/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4712 - acc: 0.7684 - val_loss: 0.4663 - val_acc: 0.7841\n", "Epoch 37/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4652 - acc: 0.7719 - val_loss: 0.5145 - val_acc: 0.7519\n", "Epoch 38/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4625 - acc: 0.7794 - val_loss: 0.4806 - val_acc: 0.7816\n", "Epoch 39/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4650 - acc: 0.7759 - val_loss: 0.4282 - val_acc: 0.7951\n", "Epoch 40/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4555 - acc: 0.7894 - val_loss: 0.4750 - val_acc: 0.7835\n", "Epoch 41/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4675 - acc: 0.7759 - val_loss: 0.4478 - val_acc: 0.7829\n", "Epoch 42/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4461 - acc: 0.7888 - val_loss: 0.4531 - val_acc: 0.7957\n", "Epoch 43/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4487 - acc: 0.7881 - val_loss: 0.4974 - val_acc: 0.7558\n", "Epoch 44/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4592 - acc: 0.7850 - val_loss: 0.4669 - val_acc: 0.7830\n", "Epoch 45/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4435 - acc: 0.7947 - val_loss: 0.5163 - val_acc: 0.7584\n", "Epoch 46/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4486 - acc: 0.7859 - val_loss: 0.4887 - val_acc: 0.7843\n", "Epoch 47/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4335 - acc: 0.7944 - val_loss: 0.4490 - val_acc: 0.7957\n", "Epoch 48/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4423 - acc: 0.7894 - val_loss: 0.5080 - val_acc: 0.7680\n", "Epoch 49/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4430 - acc: 0.7900 - val_loss: 0.5115 - val_acc: 0.7551\n", "Epoch 50/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4295 - acc: 0.8034 - val_loss: 0.4903 - val_acc: 0.7758\n", "Epoch 51/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4179 - acc: 0.8138 - val_loss: 0.4534 - val_acc: 0.7995\n", "Epoch 52/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4244 - acc: 0.8075 - val_loss: 0.4309 - val_acc: 0.8054\n", "Epoch 53/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4322 - acc: 0.8000 - val_loss: 0.4630 - val_acc: 0.7931\n", "Epoch 54/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4353 - acc: 0.7966 - val_loss: 0.4454 - val_acc: 0.7899\n", "Epoch 55/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4195 - acc: 0.8109 - val_loss: 0.4871 - val_acc: 0.7843\n", "Epoch 56/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4350 - acc: 0.7987 - val_loss: 0.4836 - val_acc: 0.7841\n", "Epoch 57/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4131 - acc: 0.7988 - val_loss: 0.4334 - val_acc: 0.8054\n", "Epoch 58/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4171 - acc: 0.8097 - val_loss: 0.4362 - val_acc: 0.8135\n", "Epoch 59/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4170 - acc: 0.8169 - val_loss: 0.5066 - val_acc: 0.7764\n", "Epoch 60/100\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "100/100 [==============================] - 16s 160ms/step - loss: 0.3986 - acc: 0.8169 - val_loss: 0.4489 - val_acc: 0.7938\n", "Epoch 61/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.4187 - acc: 0.8034 - val_loss: 0.5667 - val_acc: 0.7513\n", "Epoch 62/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4018 - acc: 0.8172 - val_loss: 0.4424 - val_acc: 0.7951\n", "Epoch 63/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4126 - acc: 0.8156 - val_loss: 0.4498 - val_acc: 0.7990\n", "Epoch 64/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3859 - acc: 0.8369 - val_loss: 0.4335 - val_acc: 0.7938\n", "Epoch 65/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4055 - acc: 0.8125 - val_loss: 0.4269 - val_acc: 0.8135\n", "Epoch 66/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3987 - acc: 0.8153 - val_loss: 0.4446 - val_acc: 0.8054\n", "Epoch 67/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.4019 - acc: 0.8100 - val_loss: 0.4815 - val_acc: 0.7906\n", "Epoch 68/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3909 - acc: 0.8225 - val_loss: 0.4768 - val_acc: 0.7816\n", "Epoch 69/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3934 - acc: 0.8216 - val_loss: 0.4726 - val_acc: 0.7957\n", "Epoch 70/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3939 - acc: 0.8281 - val_loss: 0.4308 - val_acc: 0.8106\n", "Epoch 71/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3965 - acc: 0.8172 - val_loss: 0.4436 - val_acc: 0.8090\n", "Epoch 72/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3749 - acc: 0.8244 - val_loss: 0.4478 - val_acc: 0.8003\n", "Epoch 73/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3788 - acc: 0.8281 - val_loss: 0.4577 - val_acc: 0.7996\n", "Epoch 74/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3788 - acc: 0.8269 - val_loss: 0.4192 - val_acc: 0.8141\n", "Epoch 75/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3763 - acc: 0.8397 - val_loss: 0.4381 - val_acc: 0.8073\n", "Epoch 76/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3777 - acc: 0.8297 - val_loss: 0.4630 - val_acc: 0.7963\n", "Epoch 77/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3691 - acc: 0.8341 - val_loss: 0.4566 - val_acc: 0.8003\n", "Epoch 78/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3758 - acc: 0.8291 - val_loss: 0.4973 - val_acc: 0.7868\n", "Epoch 79/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3732 - acc: 0.8353 - val_loss: 0.4994 - val_acc: 0.7848\n", "Epoch 80/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3641 - acc: 0.8369 - val_loss: 0.4273 - val_acc: 0.8080\n", "Epoch 81/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3663 - acc: 0.8391 - val_loss: 0.4262 - val_acc: 0.8147\n", "Epoch 82/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3532 - acc: 0.8422 - val_loss: 0.4869 - val_acc: 0.7841\n", "Epoch 83/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3649 - acc: 0.8400 - val_loss: 0.4610 - val_acc: 0.8160\n", "Epoch 84/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3686 - acc: 0.8409 - val_loss: 0.4116 - val_acc: 0.8286\n", "Epoch 85/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3660 - acc: 0.8353 - val_loss: 0.4980 - val_acc: 0.7912\n", "Epoch 86/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3691 - acc: 0.8384 - val_loss: 0.3991 - val_acc: 0.8260\n", "Epoch 87/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3427 - acc: 0.8462 - val_loss: 0.4288 - val_acc: 0.8204\n", "Epoch 88/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3398 - acc: 0.8491 - val_loss: 0.5238 - val_acc: 0.7906\n", "Epoch 89/100\n", "100/100 [==============================] - 16s 159ms/step - loss: 0.3668 - acc: 0.8341 - val_loss: 0.4572 - val_acc: 0.8067\n", "Epoch 90/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3376 - acc: 0.8541 - val_loss: 0.4442 - val_acc: 0.8084\n", "Epoch 91/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3548 - acc: 0.8484 - val_loss: 0.4849 - val_acc: 0.8073\n", "Epoch 92/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3390 - acc: 0.8512 - val_loss: 0.5115 - val_acc: 0.8046\n", "Epoch 93/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3406 - acc: 0.8512 - val_loss: 0.4539 - val_acc: 0.8009\n", "Epoch 94/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3398 - acc: 0.8519 - val_loss: 0.4250 - val_acc: 0.8198\n", "Epoch 95/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3340 - acc: 0.8556 - val_loss: 0.4160 - val_acc: 0.8267\n", "Epoch 96/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3377 - acc: 0.8478 - val_loss: 0.5114 - val_acc: 0.7867\n", "Epoch 97/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3308 - acc: 0.8534 - val_loss: 0.4164 - val_acc: 0.8242\n", "Epoch 98/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3328 - acc: 0.8553 - val_loss: 0.4895 - val_acc: 0.7996\n", "Epoch 99/100\n", "100/100 [==============================] - 16s 160ms/step - loss: 0.3295 - acc: 0.8572 - val_loss: 0.4285 - val_acc: 0.8249\n", "Epoch 100/100\n", "100/100 [==============================] - 16s 161ms/step - loss: 0.3332 - acc: 0.8525 - val_loss: 0.4596 - val_acc: 0.8267\n" ] } ], "source": [ "train_datagen = ImageDataGenerator(\n", " rescale=1./255,\n", " rotation_range=40,\n", " width_shift_range=0.2,\n", " height_shift_range=0.2,\n", " shear_range=0.2,\n", " zoom_range=0.2,\n", " horizontal_flip=True,)\n", "\n", "# 검증 데이터는 증식되어서는 안 됩니다!\n", "test_datagen = ImageDataGenerator(rescale=1./255)\n", "\n", "train_generator = train_datagen.flow_from_directory(\n", " # 타깃 디렉터리\n", " train_dir,\n", " # 모든 이미지를 150 × 150 크기로 바꿉니다\n", " target_size=(150, 150),\n", " batch_size=32,\n", " # binary_crossentropy 손실을 사용하기 때문에 이진 레이블을 만들어야 합니다\n", " class_mode='binary')\n", "\n", "validation_generator = test_datagen.flow_from_directory(\n", " validation_dir,\n", " target_size=(150, 150),\n", " batch_size=32,\n", " class_mode='binary')\n", "\n", "history = model.fit_generator(\n", " train_generator,\n", " steps_per_epoch=100,\n", " epochs=100,\n", " validation_data=validation_generator,\n", " validation_steps=50)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "다음 절에서 이 모델을 사용하기 위해 모델을 저장합니다." ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [], "source": [ "model.save('cats_and_dogs_small_2.h5')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "결과 그래프를 다시 그려 보죠:" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "acc = history.history['acc']\n", "val_acc = history.history['val_acc']\n", "loss = history.history['loss']\n", "val_loss = history.history['val_loss']\n", "\n", "epochs = range(len(acc))\n", "\n", "plt.plot(epochs, acc, 'bo', label='Training acc')\n", "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n", "plt.title('Training and validation accuracy')\n", "plt.legend()\n", "\n", "plt.figure()\n", "\n", "plt.plot(epochs, loss, 'bo', label='Training loss')\n", "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n", "plt.title('Training and validation loss')\n", "plt.legend()\n", "\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "그림 5-12와 5-13을 참고하세요. 데이터 증식과 드롭아웃 덕택에 더이상 과대적합되지 않습니다. 훈련 곡선이 검증 곡선에 가깝게 따라가고 있습니다. 검증 데이터에서 82% 정확도를 달성하였습니다. 규제하지 않은 모델과 비교했을 때 15% 정도 향상되었습니다.\n", "\n", "다른 규제 기법을 더 사용하고 네트워크의 파라미터를 튜닝하면(합성곱 층의 필터 수나 네트워크의 층의 수 등) 86%나 87% 정도까지 더 높은 정확도를 얻을 수도 있습니다. 하지만 데이터가 적기 때문에 컨브넷을 처음부터 훈련해서 더 높은 정확도를 달성하기는 어렵습니다. 이런 상황에서 정확도를 높이기 위한 다음 단계는 사전 훈련된 모델을 사용하는 것입니다. 다음 두 절에서 이에 대해 집중적으로 살펴보겠습니다." ] } ], "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": "3.6.6" } }, "nbformat": 4, "nbformat_minor": 2 }