{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "kernelspec": { "display_name": "TensorFlow 2.3 on Python 3.6 (CUDA 10.1)", "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": "3.6.9" }, "colab": { "name": "5-2.mnist_digit_pixel_by_pixel.ipynb", "provenance": [] }, "accelerator": "GPU", "gpuClass": "standard" }, "cells": [ { "cell_type": "markdown", "metadata": { "id": "rwSXLcpRxFsm" }, "source": [ "# MNIST 숫자의 픽셀 이미지" ] }, { "cell_type": "markdown", "metadata": { "id": "IgloWZipxFsv" }, "source": [ "MNIST 숫자를 적재하고 자세한 픽셀 수준의 그림을 그립니다. " ] }, { "cell_type": "markdown", "metadata": { "id": "YQRCCtWrxFsv" }, "source": [ "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/rickiepark/dl-illustrated/blob/master/notebooks/5-2.mnist_digit_pixel_by_pixel.ipynb)" ] }, { "cell_type": "markdown", "metadata": { "id": "kGv8yy3KxFsw" }, "source": [ "#### 라이브러리 적재" ] }, { "cell_type": "code", "metadata": { "id": "9AgJO7t6xFsw" }, "source": [ "from matplotlib import pyplot as plt\n", "import numpy as np\n", "from tensorflow.keras.datasets import mnist" ], "execution_count": 1, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "Pn0ZeUzexFsw" }, "source": [ "#### 데이터 적재" ] }, { "cell_type": "code", "metadata": { "id": "GMhrjLQBxFsw", "colab": { "base_uri": "https://localhost:8080/" }, "outputId": "454b9c1d-12a2-4f7c-f924-27e94d3ba26c" }, "source": [ "(X_train, y_train), (X_test, y_valid) = mnist.load_data()" ], "execution_count": 2, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n", "11490434/11490434 [==============================] - 0s 0us/step\n" ] } ] }, { "cell_type": "markdown", "metadata": { "id": "cdb95CVuxFsx" }, "source": [ "#### 이미지 샘플링" ] }, { "cell_type": "code", "metadata": { "id": "oVg0hkGMxFsx" }, "source": [ "# sample = np.random.randint(0, X_train.shape[0])\n", "sample = 39235" ], "execution_count": 3, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "ltKZyg1gxFsx" }, "source": [ "#### 숫자 출력" ] }, { "cell_type": "code", "metadata": { "id": "-I2CX1cqxFsx", "colab": { "base_uri": "https://localhost:8080/", "height": 558 }, "outputId": "9927cdb1-a3f1-48bf-f634-6da3f101a1b0" }, "source": [ "plt.figure(figsize = (10,10))\n", "mnist_img = X_train[sample]\n", "plt.imshow(mnist_img,cmap=\"Greys\")\n", "ax = plt.gca()\n", "\n", "# First turn off the major labels, but not the major ticks\n", "plt.tick_params(\n", " axis='both', # changes apply to the both x and y axes\n", " which='major', # Change the major ticks only\n", " bottom=True, # ticks along the bottom edge are on\n", " left=True, # ticks along the top edge are on\n", " labelbottom=False, # labels along the bottom edge are off\n", " labelleft=False) # labels along the left edge are off\n", "\n", "# Next turn off the minor ticks, but not the minor labels\n", "plt.tick_params(\n", " axis='both', # changes apply to both x and y axes\n", " which='minor', # Change the minor ticks only\n", " bottom=False, # ticks along the bottom edge are off\n", " left=False, # ticks along the left edge are off\n", " labelbottom=True, # labels along the bottom edge are on\n", " labelleft=True) # labels along the left edge are on\n", "\n", "# Set the major ticks, starting at 1 (the -0.5 tick gets hidden off the canvas)\n", "ax.set_xticks(np.arange(-.5, 28, 1))\n", "ax.set_yticks(np.arange(-.5, 28, 1))\n", "\n", "# Set the minor ticks and labels\n", "ax.set_xticks(np.arange(0, 28, 1), minor=True);\n", "ax.set_xticklabels([str(i) for i in np.arange(0, 28, 1)], minor=True);\n", "ax.set_yticks(np.arange(0, 28, 1), minor=True);\n", "ax.set_yticklabels([str(i) for i in np.arange(0, 28, 1)], minor=True);\n", "\n", "ax.grid(color='black', linestyle='-', linewidth=1.5)\n", "_ = plt.colorbar(fraction=0.046, pad=0.04, ticks=[0,32,64,96,128,160,192,224,255])" ], "execution_count": 4, "outputs": [ { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": { "needs_background": "light" } } ] }, { "cell_type": "markdown", "metadata": { "id": "ygWPNms5xFsy" }, "source": [ "#### 이미지 레이블 확인" ] }, { "cell_type": "code", "metadata": { "id": "X7Y6DpNGxFsy", "colab": { "base_uri": "https://localhost:8080/" }, "outputId": "2d8c9378-defe-40a6-8c73-b4ff4ec03116" }, "source": [ "y_train[sample]" ], "execution_count": 5, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "3" ] }, "metadata": {}, "execution_count": 5 } ] } ] }