{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "_xGwQdv44OkO"
},
"source": [
"# 머신 러닝 교과서 3판"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ZDRcmK8K4OkU"
},
"source": [
"# 12장 - 다층 인공 신경망을 밑바닥부터 구현"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "5zzcJG584OkV"
},
"source": [
"**아래 링크를 통해 이 노트북을 주피터 노트북 뷰어(nbviewer.jupyter.org)로 보거나 구글 코랩(colab.research.google.com)에서 실행할 수 있습니다.**\n",
"\n",
""
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Rh7AAS9i4OkV"
},
"source": [
"### 목차"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "zIfT2P-D4OkV"
},
"source": [
"- 인공 신경망으로 복잡한 함수 모델링\n",
" - 단일층 신경망 요약\n",
" - 다층 신경망 구조\n",
" - 정방향 계산으로 신경망 활성화 출력 계산\n",
"- 손글씨 숫자 분류\n",
" - MNIST 데이터셋 구하기\n",
" - 다층 퍼셉트론 구현\n",
"- 인공 신경망 훈련\n",
" - 로지스틱 비용 함수 계산\n",
" - 역전파 알고리즘 이해\n",
" - 역전파 알고리즘으로 신경망 훈련\n",
"- 신경망의 수렴\n",
"- 신경망 구현에 관한 몇 가지 첨언\n",
"- 요약"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "v0ItJ7A04OkW"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"id": "zvyRuFKr4OkW"
},
"outputs": [],
"source": [
"from IPython.display import Image"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ORtCZtmt4OkW"
},
"source": [
"# 인공 신경망으로 복잡한 함수 모델링"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "zZ1axgpt4OkX"
},
"source": [
"..."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "agpO4o5J4OkX"
},
"source": [
"## 단일층 신경망 요약"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 304
},
"id": "1k-XISPU4OkX",
"outputId": "d1d5d136-a386-4d79-c63f-c9f7a4d2b682",
"scrolled": true
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdrS\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 2
}
],
"source": [
"Image(url='https://git.io/JLdrS', width=600)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "OWCJE5nw4OkY"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "fhENAYu74OkY"
},
"source": [
"## 다층 신경망 구조"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 367
},
"id": "umM0VuIx4OkZ",
"outputId": "853fd472-941a-40ea-eeea-4694343f1a6b"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdrx\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 3
}
],
"source": [
"Image(url='https://git.io/JLdrx', width=600)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 366
},
"id": "HoN06CDB4OkZ",
"outputId": "3dc510fb-cfc0-4cf0-b528-934e2d3ab22a"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdrp\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 4
}
],
"source": [
"Image(url='https://git.io/JLdrp', width=500)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "o7Tzuk9m4OkZ"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "xDU-ewSp4Oka"
},
"source": [
"## 정방향 계산으로 신경망 활성화 출력 계산"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 355
},
"id": "REJ94FgH4Oka",
"outputId": "168585f2-585d-4e87-bfab-fb71fc2e0dc4"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdoe\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 5
}
],
"source": [
"Image(url='https://git.io/JLdoe', width=500)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "qMOAw-WC4Okb"
},
"source": [
"사이킷런을 사용해 MNIST 데이터를 적재하려면 다음 코드의 주석을 해제하고 실행하세요."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 72
},
"id": "Conf1Omr4Okb",
"outputId": "ba7ba146-2172-4e5b-bb21-b0776cb5d806"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"\"\\nfrom sklearn.datasets import fetch_openml\\nfrom sklearn.model_selection import train_test_split\\n\\n\\nX, y = fetch_openml('mnist_784', version=1, return_X_y=True)\\ny = y.astype(int)\\nX = ((X / 255.) - .5) * 2\\nX_train, X_test, y_train, y_test = train_test_split(\\n X, y, test_size=10000, random_state=123, stratify=y)\\n\""
],
"application/vnd.google.colaboratory.intrinsic+json": {
"type": "string"
}
},
"metadata": {},
"execution_count": 6
}
],
"source": [
"\"\"\"\n",
"from sklearn.datasets import fetch_openml\n",
"from sklearn.model_selection import train_test_split\n",
"\n",
"\n",
"X, y = fetch_openml('mnist_784', version=1, return_X_y=True)\n",
"y = y.astype(int)\n",
"X = ((X / 255.) - .5) * 2\n",
"X_train, X_test, y_train, y_test = train_test_split(\n",
" X, y, test_size=10000, random_state=123, stratify=y)\n",
"\"\"\""
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "j3zsvP8b4Okb"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "srqaPOBL4Okb"
},
"source": [
"# 손글씨 숫자 분류"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "fQ9VEg904Okc"
},
"source": [
"..."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "dZQsVb2G4Okc"
},
"source": [
"## MNIST 데이터셋 구하기"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "vm_CQgZY4Okc"
},
"source": [
"MNIST 데이터셋은 http://yann.lecun.com/exdb/mnist/에 공개되어 있으며 다음 네 부분으로 구성되어 있습니다.\n",
"\n",
"- 훈련 세트 이미지: train-images-idx3-ubyte.gz(9.9MB, 압축 해제 후 47MB, 60,000개 샘플)\n",
"- 훈련 세트 레이블: train-labels-idx1-ubyte.gz(29KB, 압축 해제 후 60KB, 60,000개 레이블)\n",
"- 테스트 세트 이미지: t10k-images-idx3-ubyte.gz(1.6MB, 압축 해제 후 7.8MB, 10,000개 샘플)\n",
"- 테스트 세트 레이블: t10k-labels-idx1-ubyte.gz(5KB, 압축 해제 후 10KB, 10,000개 레이블)\n",
"\n",
"이 절에서는 MNIST 데이터 중 일부만 사용합니다. 따라서 훈련 데이터셋의 이미지와 레이블만 다운로드합니다.\n",
"\n",
"파일을 다운로드한 후에 다음 코드 셀을 실행하면 파일 압축을 풀 수 있습니다."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "953BW0QA4Okd",
"outputId": "b4956db6-efb1-4faa-979c-c4f674ca0167"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"--2023-11-10 05:26:56-- https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/train-images-idx3-ubyte.gz\n",
"Resolving github.com (github.com)... 140.82.113.4\n",
"Connecting to github.com (github.com)|140.82.113.4|:443... connected.\n",
"HTTP request sent, awaiting response... 302 Found\n",
"Location: https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/train-images-idx3-ubyte.gz [following]\n",
"--2023-11-10 05:26:56-- https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/train-images-idx3-ubyte.gz\n",
"Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
"Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
"HTTP request sent, awaiting response... 200 OK\n",
"Length: 9912422 (9.5M) [application/octet-stream]\n",
"Saving to: ‘train-images-idx3-ubyte.gz’\n",
"\n",
"train-images-idx3-u 100%[===================>] 9.45M --.-KB/s in 0.1s \n",
"\n",
"2023-11-10 05:26:56 (83.3 MB/s) - ‘train-images-idx3-ubyte.gz’ saved [9912422/9912422]\n",
"\n",
"--2023-11-10 05:26:56-- https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/train-labels-idx1-ubyte.gz\n",
"Resolving github.com (github.com)... 140.82.112.4\n",
"Connecting to github.com (github.com)|140.82.112.4|:443... connected.\n",
"HTTP request sent, awaiting response... 302 Found\n",
"Location: https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/train-labels-idx1-ubyte.gz [following]\n",
"--2023-11-10 05:26:56-- https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/train-labels-idx1-ubyte.gz\n",
"Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
"Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
"HTTP request sent, awaiting response... 200 OK\n",
"Length: 28881 (28K) [application/octet-stream]\n",
"Saving to: ‘train-labels-idx1-ubyte.gz’\n",
"\n",
"train-labels-idx1-u 100%[===================>] 28.20K --.-KB/s in 0.003s \n",
"\n",
"2023-11-10 05:26:56 (10.9 MB/s) - ‘train-labels-idx1-ubyte.gz’ saved [28881/28881]\n",
"\n",
"--2023-11-10 05:26:57-- https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/t10k-images-idx3-ubyte.gz\n",
"Resolving github.com (github.com)... 140.82.114.3\n",
"Connecting to github.com (github.com)|140.82.114.3|:443... connected.\n",
"HTTP request sent, awaiting response... 302 Found\n",
"Location: https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/t10k-images-idx3-ubyte.gz [following]\n",
"--2023-11-10 05:26:57-- https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/t10k-images-idx3-ubyte.gz\n",
"Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
"Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
"HTTP request sent, awaiting response... 200 OK\n",
"Length: 1648877 (1.6M) [application/octet-stream]\n",
"Saving to: ‘t10k-images-idx3-ubyte.gz’\n",
"\n",
"t10k-images-idx3-ub 100%[===================>] 1.57M --.-KB/s in 0.07s \n",
"\n",
"2023-11-10 05:26:57 (21.0 MB/s) - ‘t10k-images-idx3-ubyte.gz’ saved [1648877/1648877]\n",
"\n",
"--2023-11-10 05:26:57-- https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/t10k-labels-idx1-ubyte.gz\n",
"Resolving github.com (github.com)... 140.82.114.3\n",
"Connecting to github.com (github.com)|140.82.114.3|:443... connected.\n",
"HTTP request sent, awaiting response... 302 Found\n",
"Location: https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/t10k-labels-idx1-ubyte.gz [following]\n",
"--2023-11-10 05:26:57-- https://raw.githubusercontent.com/rickiepark/python-machine-learning-book-3rd-edition/master/ch12/t10k-labels-idx1-ubyte.gz\n",
"Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
"Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
"HTTP request sent, awaiting response... 200 OK\n",
"Length: 4542 (4.4K) [application/octet-stream]\n",
"Saving to: ‘t10k-labels-idx1-ubyte.gz’\n",
"\n",
"t10k-labels-idx1-ub 100%[===================>] 4.44K --.-KB/s in 0s \n",
"\n",
"2023-11-10 05:26:57 (55.3 MB/s) - ‘t10k-labels-idx1-ubyte.gz’ saved [4542/4542]\n",
"\n"
]
}
],
"source": [
"# 코랩을 사용할 때는 다음 코드를 실행하세요.\n",
"!wget https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/train-images-idx3-ubyte.gz\n",
"!wget https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/train-labels-idx1-ubyte.gz\n",
"!wget https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/t10k-images-idx3-ubyte.gz\n",
"!wget https://github.com/rickiepark/python-machine-learning-book-3rd-edition/raw/master/ch12/t10k-labels-idx1-ubyte.gz"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"id": "gAOPi8444Okd"
},
"outputs": [],
"source": [
"# MNIST 데이터 압축을 푸는 코드\n",
"\n",
"import sys\n",
"import gzip\n",
"import shutil\n",
"import os\n",
"\n",
"if (sys.version_info > (3, 0)):\n",
" writemode = 'wb'\n",
"else:\n",
" writemode = 'w'\n",
"\n",
"zipped_mnist = [f for f in os.listdir() if f.endswith('ubyte.gz')]\n",
"for z in zipped_mnist:\n",
" with gzip.GzipFile(z, mode='rb') as decompressed, open(z[:-3], writemode) as outfile:\n",
" outfile.write(decompressed.read())"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ocRxdefS4Okd"
},
"source": [
"----\n",
"\n",
"위 코드 셀을 실행할 때 에러가 발생할 경우:\n",
"\n",
"위 코드 셀을 실행할 때 문제가 있다면 터미널에서 Unix/Linux gzip 명령을 사용해 파일의 압축을 푸는 것이 좋습니다. 예를 들어 MNIST 다운로드 디렉토리에서 다음 명령을 실행합니다.\n",
"\n",
" gzip *ubyte.gz -d\n",
"\n",
"또는 마이크로소프트 윈도우를 사용한다면 선호하는 압축 프로그램을 사용할 수 있습니다. 이미지는 바이트 형태로 저장되어 있으므로 다음에 나오는 함수를 사용해 넘파이 배열로 읽어 MLP 모델을 훈련합니다.\n",
"\n",
"gzip을 사용하지 않는다면 만들어진 파일 이름이 다음과 같은지 확인하세요.\n",
"\n",
"- train-images-idx3-ubyte\n",
"- train-labels-idx1-ubyte\n",
"- t10k-images-idx3-ubyte\n",
"- t10k-labels-idx1-ubyte\n",
"\n",
"만약 압축 해제 후에 (파일 확장자를 예측하는 일부 도구들 때문에) 파일 이름이 `train-images.idx3-ubyte`처럼 된다면 다음 코드를 진행하기 전에 `train-images-idx3-ubyte`로 이름을 바꾸어 주세요.\n",
"\n",
"----"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"id": "2pFzNdM24Okd"
},
"outputs": [],
"source": [
"import os\n",
"import struct\n",
"import numpy as np\n",
"\n",
"def load_mnist(path, kind='train'):\n",
" \"\"\"`path`에서 MNIST 데이터 불러오기\"\"\"\n",
" labels_path = os.path.join(path,\n",
" '%s-labels-idx1-ubyte' % kind)\n",
" images_path = os.path.join(path,\n",
" '%s-images-idx3-ubyte' % kind)\n",
"\n",
" with open(labels_path, 'rb') as lbpath:\n",
" magic, n = struct.unpack('>II',\n",
" lbpath.read(8))\n",
" labels = np.fromfile(lbpath,\n",
" dtype=np.uint8)\n",
"\n",
" with open(images_path, 'rb') as imgpath:\n",
" magic, num, rows, cols = struct.unpack(\">IIII\",\n",
" imgpath.read(16))\n",
" images = np.fromfile(imgpath,\n",
" dtype=np.uint8).reshape(len(labels), 784)\n",
" images = ((images / 255.) - .5) * 2\n",
"\n",
" return images, labels"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "hIdJUukR4Okd",
"outputId": "65af7376-637c-4387-c6b1-c2498de93b1f"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"sample_data\t\t t10k-labels-idx1-ubyte train-images-idx3-ubyte.gz\n",
"t10k-images-idx3-ubyte\t t10k-labels-idx1-ubyte.gz train-labels-idx1-ubyte\n",
"t10k-images-idx3-ubyte.gz train-images-idx3-ubyte train-labels-idx1-ubyte.gz\n"
]
}
],
"source": [
"!ls"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "CG3DAEF_4Oke",
"outputId": "83614961-c854-49d0-c472-d12269a03a19"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"행: 60000, 열: 784\n"
]
}
],
"source": [
"X_train, y_train = load_mnist('', kind='train')\n",
"print('행: %d, 열: %d' % (X_train.shape[0], X_train.shape[1]))"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "0onTIGXH4Oke",
"outputId": "6983ea74-759c-4d4a-dc78-1530eca1df2d"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"행: 10000, 열: 784\n"
]
}
],
"source": [
"X_test, y_test = load_mnist('', kind='t10k')\n",
"print('행: %d, 열: %d' % (X_test.shape[0], X_test.shape[1]))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "2VSBU6TT4Okf"
},
"source": [
"각 클래스의 첫 번째 이미지를 그립니다:"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 338
},
"id": "FbAKUDJG4Okg",
"outputId": "cdc1b54c-0e6a-4d49-8436-9f0e57167992"
},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
""
],
"image/png": "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\n"
},
"metadata": {}
}
],
"source": [
"import matplotlib.pyplot as plt\n",
"\n",
"fig, ax = plt.subplots(nrows=2, ncols=5, sharex=True, sharey=True)\n",
"ax = ax.flatten()\n",
"for i in range(10):\n",
" img = X_train[y_train == i][0].reshape(28, 28)\n",
" ax[i].imshow(img, cmap='Greys')\n",
"\n",
"ax[0].set_xticks([])\n",
"ax[0].set_yticks([])\n",
"plt.tight_layout()\n",
"# plt.savefig('images/12_5.png', dpi=300)\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "HUnoHgFM4Okg"
},
"source": [
"숫자 7 샘플 25개를 그립니다:"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 486
},
"id": "BavCchb_4Okh",
"outputId": "bcc33930-eb5c-40bc-fc38-b1039b959835",
"scrolled": true
},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
""
],
"image/png": "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\n"
},
"metadata": {}
}
],
"source": [
"fig, ax = plt.subplots(nrows=5, ncols=5, sharex=True, sharey=True,)\n",
"ax = ax.flatten()\n",
"for i in range(25):\n",
" img = X_train[y_train == 7][i].reshape(28, 28)\n",
" ax[i].imshow(img, cmap='Greys')\n",
"\n",
"ax[0].set_xticks([])\n",
"ax[0].set_yticks([])\n",
"plt.tight_layout()\n",
"# plt.savefig('images/12_6.png', dpi=300)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"id": "Ia7lqEr64Okh"
},
"outputs": [],
"source": [
"import numpy as np\n",
"\n",
"np.savez_compressed('mnist_scaled.npz',\n",
" X_train=X_train,\n",
" y_train=y_train,\n",
" X_test=X_test,\n",
" y_test=y_test)"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "l0lcuihY4Oki",
"outputId": "ce71c4ca-3e93-4be3-e6d5-32b25cfd30e3"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"['X_train', 'y_train', 'X_test', 'y_test']"
]
},
"metadata": {},
"execution_count": 16
}
],
"source": [
"mnist = np.load('mnist_scaled.npz')\n",
"mnist.files"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "kL_pNeYW4Oki",
"outputId": "4c00b6eb-bb9a-46fe-d419-e7ce7fbc542d"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(60000, 784)"
]
},
"metadata": {},
"execution_count": 17
}
],
"source": [
"X_train, y_train, X_test, y_test = [mnist[f] for f in ['X_train', 'y_train',\n",
" 'X_test', 'y_test']]\n",
"\n",
"del mnist\n",
"\n",
"X_train.shape"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "yo4f_UJ64Oki"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "wvMEd6qy4Oki"
},
"source": [
"## 다층 퍼셉트론 구현"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {
"id": "ps0ZyJjU4Oki"
},
"outputs": [],
"source": [
"import numpy as np\n",
"import sys\n",
"\n",
"\n",
"class NeuralNetMLP(object):\n",
" \"\"\"피드포워드 신경망 / 다층 퍼셉트론 분류기\n",
"\n",
" 매개변수\n",
"\n",
" ------------\n",
" n_hidden : int (기본값: 30)\n",
" 은닉 유닛 개수\n",
" l2 : float (기본값: 0.)\n",
" L2 규제의 람다 값\n",
" l2=0이면 규제 없음. (기본값)\n",
" epochs : int (기본값: 100)\n",
" 훈련 세트를 반복할 횟수\n",
" eta : float (기본값: 0.001)\n",
" 학습률\n",
" shuffle : bool (기본값: True)\n",
" 에포크마다 훈련 세트를 섞을지 여부\n",
" True이면 데이터를 섞어 순서를 바꿉니다\n",
" minibatch_size : int (기본값: 1)\n",
" 미니 배치의 훈련 샘플 개수\n",
" seed : int (기본값: None)\n",
" 가중치와 데이터 셔플링을 위한 난수 초깃값\n",
"\n",
" 속성\n",
" -----------\n",
" eval_ : dict\n",
" 훈련 에포크마다 비용, 훈련 정확도, 검증 정확도를 수집하기 위한 딕셔너리\n",
"\n",
" \"\"\"\n",
" def __init__(self, n_hidden=30,\n",
" l2=0., epochs=100, eta=0.001,\n",
" shuffle=True, minibatch_size=1, seed=None):\n",
"\n",
" self.random = np.random.RandomState(seed)\n",
" self.n_hidden = n_hidden\n",
" self.l2 = l2\n",
" self.epochs = epochs\n",
" self.eta = eta\n",
" self.shuffle = shuffle\n",
" self.minibatch_size = minibatch_size\n",
"\n",
" def _onehot(self, y, n_classes):\n",
" \"\"\"레이블을 원-핫 방식으로 인코딩합니다\n",
"\n",
" 매개변수\n",
" ------------\n",
" y : 배열, 크기 = [n_samples]\n",
" 타깃 값.\n",
" n_classes : int\n",
" 클래스 개수\n",
"\n",
" 반환값\n",
" -----------\n",
" onehot : 배열, 크기 = (n_samples, n_labels)\n",
"\n",
" \"\"\"\n",
" onehot = np.zeros((n_classes, y.shape[0]))\n",
" for idx, val in enumerate(y.astype(int)):\n",
" onehot[val, idx] = 1.\n",
" return onehot.T\n",
"\n",
" def _sigmoid(self, z):\n",
" \"\"\"로지스틱 함수(시그모이드)를 계산합니다\"\"\"\n",
" return 1. / (1. + np.exp(-np.clip(z, -250, 250)))\n",
"\n",
" def _forward(self, X):\n",
" \"\"\"정방향 계산을 수행합니다\"\"\"\n",
"\n",
" # 단계 1: 은닉층의 최종 입력\n",
" # [n_samples, n_features] dot [n_features, n_hidden]\n",
" # -> [n_samples, n_hidden]\n",
" z_h = np.dot(X, self.w_h) + self.b_h\n",
"\n",
" # 단계 2: 은닉층의 활성화 출력\n",
" a_h = self._sigmoid(z_h)\n",
"\n",
" # 단계 3: 출력층의 최종 입력\n",
" # [n_samples, n_hidden] dot [n_hidden, n_classlabels]\n",
" # -> [n_samples, n_classlabels]\n",
" z_out = np.dot(a_h, self.w_out) + self.b_out\n",
"\n",
" # 단계 4: 출력층의 활성화 출력\n",
" a_out = self._sigmoid(z_out)\n",
"\n",
" return z_h, a_h, z_out, a_out\n",
"\n",
" def _compute_cost(self, y_enc, output):\n",
" \"\"\"비용 함수를 계산합니다\n",
"\n",
" 매개변수\n",
" ----------\n",
" y_enc : 배열, 크기 = (n_samples, n_labels)\n",
" 원-핫 인코딩된 클래스 레이블\n",
" output : 배열, 크기 = [n_samples, n_output_units]\n",
" 출력층의 활성화 출력 (정방향 계산)\n",
"\n",
" 반환값\n",
" ---------\n",
" cost : float\n",
" 규제가 포함된 비용\n",
"\n",
" \"\"\"\n",
" L2_term = (self.l2 *\n",
" (np.sum(self.w_h ** 2.) +\n",
" np.sum(self.w_out ** 2.)))\n",
"\n",
" term1 = -y_enc * (np.log(output))\n",
" term2 = (1. - y_enc) * np.log(1. - output)\n",
" cost = np.sum(term1 - term2) + L2_term\n",
"\n",
" # 다른 데이터셋에서는 극단적인 (0 또는 1에 가까운) 활성화 값이 나올 수 있습니다.\n",
" # 파이썬과 넘파이의 수치 연산이 불안정하기 때문에 \"ZeroDivisionError\"가 발생할 수 있습니다.\n",
" # 즉, log(0)을 평가하는 경우입니다.\n",
" # 이 문제를 해결하기 위해 로그 함수에 전달되는 활성화 값에 작은 상수를 더합니다.\n",
" #\n",
" # 예를 들어:\n",
" #\n",
" # term1 = -y_enc * (np.log(output + 1e-5))\n",
" # term2 = (1. - y_enc) * np.log(1. - output + 1e-5)\n",
"\n",
" return cost\n",
"\n",
" def predict(self, X):\n",
" \"\"\"클래스 레이블을 예측합니다\n",
"\n",
" 매개변수\n",
" -----------\n",
" X : 배열, 크기 = [n_samples, n_features]\n",
" 원본 특성의 입력층\n",
"\n",
" 반환값:\n",
" ----------\n",
" y_pred : 배열, 크기 = [n_samples]\n",
" 예측된 클래스 레이블\n",
"\n",
" \"\"\"\n",
" z_h, a_h, z_out, a_out = self._forward(X)\n",
" y_pred = np.argmax(z_out, axis=1)\n",
" return y_pred\n",
"\n",
" def fit(self, X_train, y_train, X_valid, y_valid):\n",
" \"\"\"훈련 데이터에서 가중치를 학습합니다\n",
"\n",
" 매개변수\n",
" -----------\n",
" X_train : 배열, 크기 = [n_samples, n_features]\n",
" 원본 특성의 입력층\n",
" y_train : 배열, 크기 = [n_samples]\n",
" 타깃 클래스 레이블\n",
" X_valid : 배열, 크기 = [n_samples, n_features]\n",
" 훈련하는 동안 검증에 사용할 샘플 특성\n",
" y_valid : 배열, 크기 = [n_samples]\n",
" 훈련하는 동안 검증에 사용할 샘플 레이블\n",
"\n",
" 반환값:\n",
" ----------\n",
" self\n",
"\n",
" \"\"\"\n",
" n_output = np.unique(y_train).shape[0] # number of class labels\n",
" n_features = X_train.shape[1]\n",
"\n",
" ########################\n",
" # 가중치 초기화\n",
" ########################\n",
"\n",
" # 입력층 -> 은닉층 사이의 가중치\n",
" self.b_h = np.zeros(self.n_hidden)\n",
" self.w_h = self.random.normal(loc=0.0, scale=0.1,\n",
" size=(n_features, self.n_hidden))\n",
"\n",
" # 은닉층 -> 출력층 사이의 가중치\n",
" self.b_out = np.zeros(n_output)\n",
" self.w_out = self.random.normal(loc=0.0, scale=0.1,\n",
" size=(self.n_hidden, n_output))\n",
"\n",
" epoch_strlen = len(str(self.epochs)) # 출력 포맷을 위해\n",
" self.eval_ = {'cost': [], 'train_acc': [], 'valid_acc': []}\n",
"\n",
" y_train_enc = self._onehot(y_train, n_output)\n",
"\n",
" # 훈련 에포크를 반복합니다\n",
" for i in range(self.epochs):\n",
"\n",
" # 미니 배치로 반복합니다\n",
" indices = np.arange(X_train.shape[0])\n",
"\n",
" if self.shuffle:\n",
" self.random.shuffle(indices)\n",
"\n",
" for start_idx in range(0, indices.shape[0] - self.minibatch_size +\n",
" 1, self.minibatch_size):\n",
" batch_idx = indices[start_idx:start_idx + self.minibatch_size]\n",
"\n",
" # 정방향 계산\n",
" z_h, a_h, z_out, a_out = self._forward(X_train[batch_idx])\n",
"\n",
" ##################\n",
" # 역전파\n",
" ##################\n",
"\n",
" # [n_examples, n_classlabels]\n",
" delta_out = a_out - y_train_enc[batch_idx]\n",
"\n",
" # [n_examples, n_hidden]\n",
" sigmoid_derivative_h = a_h * (1. - a_h)\n",
"\n",
" # [n_examples, n_classlabels] dot [n_classlabels, n_hidden]\n",
" # -> [n_examples, n_hidden]\n",
" delta_h = (np.dot(delta_out, self.w_out.T) *\n",
" sigmoid_derivative_h)\n",
"\n",
" # [n_features, n_examples] dot [n_examples, n_hidden]\n",
" # -> [n_features, n_hidden]\n",
" grad_w_h = np.dot(X_train[batch_idx].T, delta_h)\n",
" grad_b_h = np.sum(delta_h, axis=0)\n",
"\n",
" # [n_hidden, n_examples] dot [n_examples, n_classlabels]\n",
" # -> [n_hidden, n_classlabels]\n",
" grad_w_out = np.dot(a_h.T, delta_out)\n",
" grad_b_out = np.sum(delta_out, axis=0)\n",
"\n",
" # 규제와 가중치 업데이트\n",
" delta_w_h = (grad_w_h + self.l2*self.w_h)\n",
" delta_b_h = grad_b_h # 편향은 규제하지 않습니다\n",
" self.w_h -= self.eta * delta_w_h\n",
" self.b_h -= self.eta * delta_b_h\n",
"\n",
" delta_w_out = (grad_w_out + self.l2*self.w_out)\n",
" delta_b_out = grad_b_out # 편향은 규제하지 않습니다\n",
" self.w_out -= self.eta * delta_w_out\n",
" self.b_out -= self.eta * delta_b_out\n",
"\n",
" #############\n",
" # 평가\n",
" #############\n",
"\n",
" # 훈련하는 동안 에포크마다 평가합니다\n",
" z_h, a_h, z_out, a_out = self._forward(X_train)\n",
"\n",
" cost = self._compute_cost(y_enc=y_train_enc,\n",
" output=a_out)\n",
"\n",
" y_train_pred = self.predict(X_train)\n",
" y_valid_pred = self.predict(X_valid)\n",
"\n",
" # 넘파이 1.20에서 `np.float`가 deprecated되므로 대신 `float`를 사용합니다.\n",
" train_acc = ((np.sum(y_train == y_train_pred)).astype(float) /\n",
" X_train.shape[0])\n",
" valid_acc = ((np.sum(y_valid == y_valid_pred)).astype(float) /\n",
" X_valid.shape[0])\n",
"\n",
" sys.stderr.write('\\r%0*d/%d | 비용: %.2f '\n",
" '| 훈련/검증 정확도: %.2f%%/%.2f%% ' %\n",
" (epoch_strlen, i+1, self.epochs, cost,\n",
" train_acc*100, valid_acc*100))\n",
" sys.stderr.flush()\n",
"\n",
" self.eval_['cost'].append(cost)\n",
" self.eval_['train_acc'].append(train_acc)\n",
" self.eval_['valid_acc'].append(valid_acc)\n",
"\n",
" return self"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {
"id": "-vdALhZX4Okj"
},
"outputs": [],
"source": [
"n_epochs = 200"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "hNzZGfqF4Okk",
"outputId": "a8565384-aad5-4f91-ef64-b2611f441dd6"
},
"outputs": [
{
"output_type": "stream",
"name": "stderr",
"text": [
"200/200 | 비용: 5065.78 | 훈련/검증 정확도: 99.28%/97.98% "
]
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<__main__.NeuralNetMLP at 0x793aad8f1300>"
]
},
"metadata": {},
"execution_count": 20
}
],
"source": [
"nn = NeuralNetMLP(n_hidden=100,\n",
" l2=0.01,\n",
" epochs=n_epochs,\n",
" eta=0.0005,\n",
" minibatch_size=100,\n",
" shuffle=True,\n",
" seed=1)\n",
"\n",
"nn.fit(X_train=X_train[:55000],\n",
" y_train=y_train[:55000],\n",
" X_valid=X_train[55000:],\n",
" y_valid=y_train[55000:])"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 449
},
"id": "8YRZV-vn4Okk",
"outputId": "55f1e6f2-a62a-4416-eaa7-8965c55c7056"
},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
""
],
"image/png": "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\n"
},
"metadata": {}
}
],
"source": [
"import matplotlib.pyplot as plt\n",
"\n",
"plt.plot(range(nn.epochs), nn.eval_['cost'])\n",
"plt.ylabel('Cost')\n",
"plt.xlabel('Epochs')\n",
"# plt.savefig('images/12_07.png', dpi=300)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 454
},
"id": "z4O1gYie4Okl",
"outputId": "6727ee2e-b888-42c4-90b8-9b52d5925324"
},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
""
],
"image/png": "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\n"
},
"metadata": {}
}
],
"source": [
"plt.plot(range(nn.epochs), nn.eval_['train_acc'],\n",
" label='Training')\n",
"plt.plot(range(nn.epochs), nn.eval_['valid_acc'],\n",
" label='Validation', linestyle='--')\n",
"plt.ylabel('Accuracy')\n",
"plt.xlabel('Epochs')\n",
"plt.legend(loc='lower right')\n",
"# plt.savefig('images/12_08.png', dpi=300)\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "zscbDz7F4Okl",
"outputId": "b71cc9c1-1880-4f17-e576-835a6d7a1a67"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"테스트 정확도: 97.54%\n"
]
}
],
"source": [
"y_test_pred = nn.predict(X_test)\n",
"# 넘파이 1.20에서 `np.float`가 deprecated되므로 대신 `float`를 사용합니다.\n",
"acc = (np.sum(y_test == y_test_pred)\n",
" .astype(float) / X_test.shape[0])\n",
"\n",
"print('테스트 정확도: %.2f%%' % (acc * 100))"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 486
},
"id": "bOSkex4s4Okl",
"outputId": "bb4dbe39-150d-479a-95da-b26a5f0e50c6"
},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
""
],
"image/png": "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\n"
},
"metadata": {}
}
],
"source": [
"miscl_img = X_test[y_test != y_test_pred][:25]\n",
"correct_lab = y_test[y_test != y_test_pred][:25]\n",
"miscl_lab = y_test_pred[y_test != y_test_pred][:25]\n",
"\n",
"fig, ax = plt.subplots(nrows=5, ncols=5, sharex=True, sharey=True)\n",
"ax = ax.flatten()\n",
"for i in range(25):\n",
" img = miscl_img[i].reshape(28, 28)\n",
" ax[i].imshow(img, cmap='Greys', interpolation='nearest')\n",
" ax[i].set_title('%d) t: %d p: %d' % (i+1, correct_lab[i], miscl_lab[i]))\n",
"\n",
"ax[0].set_xticks([])\n",
"ax[0].set_yticks([])\n",
"plt.tight_layout()\n",
"# plt.savefig('images/12_09.png', dpi=300)\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "j3Wd15HC4Okl"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "h8bE0O6b4Okl"
},
"source": [
"# 인공 신경망 훈련"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "2k-FpEok4Okl"
},
"source": [
"..."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "R4si-7664Okl"
},
"source": [
"## 로지스틱 비용 함수 계산"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 312
},
"id": "GAakQltn4Okl",
"outputId": "3079816c-4232-478f-9097-24410ca932e7"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdov\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 25
}
],
"source": [
"Image(url='https://git.io/JLdov', width=300)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "LN03rNNR4Okm"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ETGepTEw4Okm"
},
"source": [
"## 역전파 알고리즘 이해"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "pFPsW5UE4Okm"
},
"source": [
"..."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "iovHPtie4Okm"
},
"source": [
"## 역전파 알고리즘으로 신경망 훈련"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 237
},
"id": "kkqR0QCI4Okm",
"outputId": "2bccb629-0abd-439f-cc5b-03a72f3bb85c"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdoa\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 26
}
],
"source": [
"Image(url='https://git.io/JLdoa', width=400)"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 405
},
"id": "Bj3Colmn4Okm",
"outputId": "62724872-78e1-4ead-8d55-3d9688e518de"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdoz\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 27
}
],
"source": [
"Image(url='https://git.io/JLdoz', width=500)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "yCWIr16M4Okm"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "_t9nD_3j4Okn"
},
"source": [
"# 신경망의 수렴"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 336
},
"id": "X5M8ZRzm4Okn",
"outputId": "2bd38ca6-9cbe-4aa4-a215-ed534420c252"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/html": [
"![](\"https://git.io/JLdoK\")
"
],
"text/plain": [
""
]
},
"metadata": {},
"execution_count": 28
}
],
"source": [
"Image(url='https://git.io/JLdoK', width=500)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "hwvf3FQX4Okn"
},
"source": [
"
\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "8dwCxPtm4Okn"
},
"source": [
"..."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "zKoOueKo4Okn"
},
"source": [
"# 요약"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "o2TWIlZV4Okn"
},
"source": [
"..."
]
}
],
"metadata": {
"anaconda-cloud": {},
"colab": {
"name": "ch12.ipynb",
"provenance": []
},
"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"
},
"gpuClass": "standard"
},
"nbformat": 4,
"nbformat_minor": 0
} ![](\"https://jupyter.org/assets/share.png\")
주피터 노트북 뷰어로 보기\n",
" ![](\"https://www.tensorflow.org/images/colab_logo_32px.png\"){kind=logo}
구글 코랩(Colab)에서 실행하기\n",
"