ํ‹ฐ์Šคํ† ๋ฆฌ

๐Ÿ’ฃ
๊ฒ€์ƒ‰ํ•˜๊ธฐ

๋ธ”๋กœ๊ทธ ํ™ˆ

๐Ÿ’ฃ

leebomb.tistory.com/m

์ผ๋‹จ ํ•ด๋ผ

๊ตฌ๋…์ž
0
๋ฐฉ๋ช…๋ก ๋ฐฉ๋ฌธํ•˜๊ธฐ

์ฃผ์š” ๊ธ€ ๋ชฉ๋ก

  • DAY73. Tensorflow Text Vectorizing RNN (2) RNN LSTM basic RNN model - ์ˆœํ™˜์‹ ๊ฒฝ๋ง ๋ชจ๋ธ * NotImplementedError ์˜ค๋ฅ˜ํ•ด๊ฒฐ๋ฒ• : array_ops.py ํŒŒ์ผ ๊ต์ฒด(data ํด๋”) -> restart C:\\UsersITWILLanaconda3envs\tensorflow\Lib\site-packages\tensorflow\python\ops C:\ProgramData\Anaconda3\envs\tensorflow\Lib\site-packages\tensorflow\python\ops array_ops.py ํŒŒ์ผ ๊ต์ฒด import tensorflow as tf #seed value import numpy as np #ndarray from tensorflow.keras import Sequential #model fr.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2022. 1. 4.
  • DAY72. Tensorflow Text Vectorizing RNN ํ…์ŠคํŠธ ๋ฒกํ„ฐํ™”? - ํ…์ŠคํŠธ๋ฅผ ์ˆซ์žํ˜• ๋ฒกํ„ฐ๋กœ ๋ณ€ํ™˜ํ•˜๋Š” ์ „์ฒ˜๋ฆฌ ๊ณผ์ • - ๋”ฅ๋Ÿฌ๋‹ ๋ชจ๋ธ์€ ์ˆซ์žํ˜• ์ž๋ฃŒ๋งŒ ์ฒ˜๋ฆฌ ๋ฐฉ๋ฒ• ํ…์ŠคํŠธ → ๋‹จ์–ด → ๋‹จ์–ด ๋ฒกํ„ฐ ๋ณ€ํ™˜ ํ…์ŠคํŠธ → ๋ฌธ์ž → ๋ฌธ์ž ๋ฒกํ„ฐ ๋ณ€ํ™˜ ํ…์ŠคํŠธ → N-gram(๋‹จ์–ด๋‚˜ ๋ฌธ์ž ๊ทธ๋ฃน) → N-gram ๋ฒกํ„ฐ ๋ณ€ํ™˜ * N-gram : ์—ฐ์†๋œ ๋‹จ์–ด๋‚˜ ๋ฌธ์ž์˜ ๊ทธ๋ฃน ๋‹จ์œ„ (ํ…์ŠคํŠธ ์—์„œ ๋‹จ์–ด๋‚˜ ๋ฌธ์ž๋ฅผ ํ•˜๋‚˜ ์”ฉ ์ด๋™ํ•˜๋ฉด์„œ ์ถ”์ถœ) ๋ฒกํ„ฐ ๋ณ€ํ™˜ ๋ฐฉ๋ฒ• (ํ† ํฐ → ์ˆซ์žํ˜• ๋ฒกํ„ฐ ๋ณ€ํ™˜) ์›-ํ•ซ ์ธ์ฝ”๋”ฉ (ํฌ์†Œํ–‰๋ ฌ) ๋‹จ์–ด ์ž„๋ฒ ๋”ฉ (๋ฐ€์ง‘ํ–‰๋ ฌ) ์ •์˜ ๋ฐ์ดํ„ฐ๋ฅผ ์ˆ˜๋งŽ์€ 0๊ณผ ํ•œ๊ฐœ์˜ 1๋กœ ๊ตฌ๋ณ„ ๋ฒกํ„ฐ ๋˜๋Š” ํ–‰๋ ฌ(matrix)์˜ ๊ฐ’์ด ๋Œ€๋ถ€๋ถ„ 0์œผ๋กœ ํ‘œํ˜„๋˜๋Š” ํฌ์†Œ ํ–‰๋ ฌ(sparse matrix) ๋‹จ์–ด์˜ ์˜๋ฏธ๋ฅผ ๊ณ ๋ คํ•˜์—ฌ ์ข€ ๋” ์กฐ๋ฐ€ํ•œ ์ฐจ์›์— ๋‹จ์–ด๋ฅผ ๋ฒกํ„ฐ๋กœ ํ‘œํ˜„ ์ ์€ ์ฐจ์›์œผ๋กœ ๋” ๋งŽ์€ ์ •๋ณด๋ฅผ ์ €์žฅํ•˜๋Š” ๋ฐ€์ง‘ํ–‰๋ ฌ(D.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 31.
  • DAY71. Tensorflow Face detection (2) celeb image classifier 1. celeb5 ์ด๋ฏธ์ง€ ๋ถ„๋ฅ˜๊ธฐ : CNN model 2. Image Generator : model ๊ณต๊ธ‰ํ•  ์ด๋ฏธ์ง€ ์ƒ์„ฑ * cats dogs imageGenerator ์ฐธ๊ณ  from tensorflow.keras import Sequential #keras model from tensorflow.keras.layers import Conv2D, MaxPool2D #Convolution layer from tensorflow.keras.layers import Dense, Flatten #Affine layer ๊ณต๊ธ‰ image ํฌ๊ธฐ img_h = 150 # height img_w = 150 # width input_shape = (img_h, img_w, 3) 1.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 30.
  • DAY70. Tensorflow Face detection (1)face landmark ๋”ฅ ๋Ÿฌ๋‹ ์ ์šฉ ์ด๋ฏธ์ง€ ๋ถ„์„ ์ ˆ์ฐจ STEP 01 Data Crawling (Selenium) ์ด๋ฏธ์ง€ ์ž๋ฃŒ ์ˆ˜์ง‘ STEP 02 1) Face Detection (dlib) ์–ผ๊ตด ์ธ์‹ HOG (Histogram of Gradient Face Detection) : ์ด๋ฏธ์ง€์—์„œ face ๊ฐ์ง€ ์•Œ๊ณ ๋ฆฌ์ฆ˜ * HOG : ๋Œ€์ƒ ์˜์—ญ์„ ์ผ์ • ํฌ๊ธฐ์˜ ์…€๋กœ ๋ถ„ํ• ํ•˜๊ณ , ๊ฐ ์…€๋งˆ๋‹ค edge ํ”ฝ์…€ (gradient magnitude๊ฐ€ ์ผ์ • ๊ฐ’ ์ด์ƒ์ธ ํ”ฝ์…€)๋“ค์˜ ๋ฐฉํ–ฅ์— ๋Œ€ํ•œ ํžˆ์Šคํ† ๊ทธ๋žจ์„ ๊ตฌํ•œ ํ›„ ์ด๋“ค ํžˆ์Šคํ† ๊ทธ๋žจ bin ๊ฐ’๋“ค์„ ์ผ๋ ฌ๋กœ ์—ฐ๊ฒฐํ•œ ๋ฒกํ„ฐ 2) ์–ผ๊ตด ์ •๋ ฌ(Face Alignment) 3) API Install python3.8์˜ ๊ฐ€์ƒํ™˜๊ฒฝ์—์„œ cmake ์„ค์น˜ (base) > conda activate tensorfow (tens.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 29.
  • DAY69. Tensorflow Selenium Crawling Selenium Selenium ์ด๋ž€? ์›๊ฒฉ์œผ๋กœ ํŠน์ • ์›นํŽ˜์ด์ง€์˜ ๋ฒ„ํŠผ ํด๋ฆญ, ์ž…๋ ฅ์ƒ์ž์—์„œ ์ž๋ฃŒ ์ž…๋ ฅ ๋“ฑ์œผ๋กœ ์–ด๋–ค ๊ฒฐ๊ณผ๊ฐ€ ๋‚˜์˜ค๋Š”์ง€ ๋“ฑ์˜ ๋‹ค์–‘ํ•œ ์›น ํŽ˜์ด์ง€์™€ ์‚ฌ์šฉ์ž ๊ฐ„์˜ ์ƒํ˜ธ์ž‘์šฉ์„ ๋™์ ์œผ ๋กœ ์ œ์–ดํ•˜๋Š” ๊ธฐ์ˆ  ๋˜๋Š” ํ”„๋กœ๊ทธ๋žจ ์šฉ๋„ : ๋™์  ์›นํŽ˜์ด์ง€ ์ž๋ฃŒ ์ˆ˜์ง‘(์˜ํ™” ๋ฆฌ๋ทฐ), ๊ตฌ๊ธ€ ์ด๋ฏธ์ง€ ์ˆ˜์ง‘(์ ˆ์ฐจ) ๋™์  ํŽ˜์ด์ง€ vs ์ •์  ํŽ˜์ด์ง€ 1) ์ •์  ํŽ˜์ด์ง€ - ์ด๋ฏธ ์ค€๋น„๋˜์–ด ์žˆ๋Š” ์›น๋ฌธ์„œ๋ฅผ ์‚ฌ์šฉ์ž(client)์—๊ฒŒ ์ œ๊ณต - ์–ธ์ œ ์ ‘์†ํ•ด๋„ ๋™์ผํ•œ ๋ฆฌ์†Œ์Šค๋ฅผ ์ œ๊ณตํ•˜๋Š” ์›น์‚ฌ์ดํŠธ - BeautifulSoup ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋ฅผ ์ด์šฉํ•˜์—ฌ ์›น๋ฌธ์„œ ์ˆ˜์ง‘ 2) ๋™์  ํŽ˜์ด์ง€ - ์‚ฌ์šฉ์ž(client)์˜ ์š”์ฒญ์„ ๋ฐ›์€ ์‹œ์ ์—์„œ ์›น๋ฌธ์„œ๋ฅผ ์‚ฌ์šฉ์ž์—๊ฒŒ ์ œ๊ณต - ์‚ฌ์šฉ์ž์˜ ์š”์ฒญ์— ๋”ฐ๋ผ์„œ ์„œ๋กœ ๋‹ค๋ฅธ ๋ฆฌ์†Œ์Šค๋ฅผ ์ œ๊ณตํ•˜๋Š” ์›น์‚ฌ์ดํŠธ - Selenium ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๋ฅผ ์ด์šฉํ•˜์—ฌ .. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 28.
  • DAY68. Tensorflow CNN model (2)ImageGenerator Cats vs Dogs image classifier - image data generator ์ด์šฉ : ํ•™์Šต ๋ฐ์ดํ„ฐ์…‹ ๋งŒ๋“ค๊ธฐ from tensorflow.keras import Sequential #keras model from tensorflow.keras.layers import Conv2D, MaxPool2D #Convolution layer from tensorflow.keras.layers import Dense, Flatten #Affine layer ๊ณต๊ธ‰ image ํฌ๊ธฐ img_h = 150 #height img_w = 150 #width input_shape = (img_h, img_w, 3) 1. CNN Model layer print('model create') model = Seque.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 27.
  • DAY67. Tensorflow CNN model mnist cnnbasic MNIST + CNN basic ํ•ฉ์„ฑ๊ณฑ์ธต : ํŠน์ง•๋งต(feature map) - ์ด๋ฏธ์ง€ ํŠน์ง• ์ถ”์ถœ ํด๋ง์ธต : ํ”ฝ์…€ ์ถ•์†Œ(๋‹ค์šด์ƒ˜ํ”Œ๋ง) - ์ด๋ฏธ์ง€ ํŠน์ง• ๊ฐ•์กฐ import tensorflow as tf from tensorflow.keras.datasets import mnist # image dataset import matplotlib.pyplot as plt # image ์‹œ๊ฐํ™” import numpy as np 1. image dataset load (x_train, y_train), (x_test,y_test) = mnist.load_data() x_train.shape #(60000, 28, 28) - (size, h, w) print(x_train[0]) 2. ์ •์ˆ˜ํ˜• ->.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 24.
  • DAY66. Tensorflow Keras model (2)Overfitting solution karas mnist history * keras mnist batch์ฐธ๊ณ  History : ํ›ˆ๋ จ๊ณผ ๊ฒ€์ฆ๊ณผ์ •์—์„œ ๋ฐœ์ƒํ•˜๋Š” ์†์‹ค๊ฐ’/์ •ํ™•๋„ ๊ฒฐ๊ณผ ๊ธฐ์–ต ๊ธฐ๋Šฅ from tensorflow.keras.datasets import mnist #mnist load from tensorflow.keras.utils import to_categorical #Y๋ณ€์ˆ˜ : encoding from tensorflow.keras import Sequential #keras model ์ƒ์„ฑ from tensorflow.keras.layers import Dense #DNN layer ๊ตฌ์ถ• import matplotlib.pyplot as plt #์‹œ๊ฐํ™” ๋„๊ตฌ keras ๋‚ด๋ถ€ w,b๋ณ€์ˆ˜ seed ์ ์šฉ import tensorflo.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 23.
  • DAY65. Tensorflow Keras model (1)dnn model keras binary - Keras model : ์ดํ•ญ๋ถ„๋ฅ˜๊ธฐ X๋ณ€์ˆ˜ : ์ •๊ทœํ™”(0~1) Y๋ณ€์ˆ˜ : one hot encoding(2์ง„์ˆ˜) from sklearn.datasets import load_iris #dataset from sklearn.model_selection import train_test_split #split from sklearn.preprocessing import minmax_scale #X๋ณ€์ˆ˜ : ์ •๊ทœํ™”(0~1) from tensorflow.keras.utils import to_categorical #Y๋ณ€์ˆ˜ : encoding from tensorflow.keras import Sequential #keras model ์ƒ์„ฑ from tensorflow.keras.layer.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 22.
  • DAY64. Tensorflow Classification (Sigmoid, Softmax) index ๋ฐ˜ํ™˜ ํ•จ์ˆ˜ argmin : ์ถ• ๋ณ„ ์ตœ์†Œ ๊ฐ’์˜ index ๋ฐ˜ํ™˜ argmax : ์ถ• ๋ณ„ ์ตœ๋Œ€ ๊ฐ’์˜ index ๋ฐ˜ํ™˜ unique : ์ค‘๋ณต ์ œ๊ฑฐ ๊ฒฐ๊ณผ index ๋ฐ˜ํ™˜ kNN classifier kNN ๋ถ„๋ฅ˜๊ธฐ ํŠน์ง• - ์•Œ๋ ค์ง€์ง€ ์•Š์€ ๋ฒ”์ฃผ๋ฅผ ๋Œ€์ƒ์œผ๋กœ ๊ฐ€์žฅ ์œ ์‚ฌํ•œ ๋ฒ”์ฃผ๋กœ ๋ถ„๋ฅ˜ - ๊ธฐ์กด ๋ฒ”์ฃผ๊ฐ€ ์กด์žฌํ•ด์•ผ ํ•จ - ์‹๋ฃŒํ’ˆ(๊ณผ์ผ, ์ฑ„์†Œ, ๋‹จ๋ฐฑ์งˆ ๋“ฑ) - ํ•™์Šตํ•˜์ง€ ์•Š์Œ : ๊ฒŒ์œผ๋ฅธ ํ•™์Šต kNN classifier ์•Œ๊ณ ๋ฆฌ์ฆ˜ * X : ๊ธฐ์กด ๋ถ„๋ฅ˜์ง‘๋‹จ Y : ์ƒˆ๋กœ์šด ๋ถ„๋ฅ˜๋Œ€์ƒ 1. tensorflow Euclidean ๊ฑฐ๋ž˜ ๊ณ„์‚ฐ์‹ distance = tf.sqrt( tf.reduce_sum( tf.square(X-Y) ), axis=1)) 2.์˜ค์ฐจ๊ฐ€ ์ตœ์†Œ์ธ index ๋ฐ˜ํ™˜ (Euclidean ๊ฑฐ๋ฆฌ๊ณ„์‚ฐ์‹) pred = tf.a.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 21.
  • DAY63. Tensorflow LinearRegression (3)keras dnn Linear Regression ์‹ ๊ฒฝ๋ง Keras DNN model(Linear Regression ์‹ ๊ฒฝ๋ง) Hidden layer : 2๊ฐœ(๋‰ด๋Ÿฐ 2๊ฐœ) model = Sequential() #hidden layer1 : shape = [2, 2] model.add(Dense(2, input_shape=(2,), activation = 'relu')) # 1์ธต(hidden1) #hidden layer2 : shape = [2, 2] model.add(Dense(2, activation = 'relu')) # 2์ธต(hidden2) #output layer : shape = [2, 1] model.add(Dense(units=1)) #์ถœ๋ ฅ์ธต(output) keras model High Level API : .. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 20.
  • DAY62. Tensorflow LinearRegression (2)ํšŒ๊ท€๋ชจ๋ธ ์„ ํ˜•ํšŒ๊ท€(Linear Regression) ๊ฐœ์š” ํšŒ๊ท€๋ถ„์„(Regression Analysis) ํŠน์ • ๋ณ€์ˆ˜(๋…๋ฆฝ๋ณ€์ˆ˜)๊ฐ€ ๋‹ค๋ฅธ ๋ณ€์ˆ˜(์ข…์†๋ณ€์ˆ˜)์— ์–ด๋– ํ•œ ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š”๊ฐ€ (์ธ๊ณผ๊ด€๊ณ„ ๋ถ„์„) ex) ๊ฐ€๊ฒฉ์€ ์ œํ’ˆ ๋งŒ์กฑ๋„์— ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š”๊ฐ€? -> ํ•œ ๋ณ€์ˆ˜์˜ ๊ฐ’์œผ๋กœ ๋‹ค๋ฅธ ๋ณ€์ˆ˜์˜ ๊ฐ’ ์˜ˆ์–ธ ์ƒ๊ด€๊ด€๊ณ„๋ถ„์„ : ๋ณ€์ˆ˜ ๊ฐ„์˜ ๊ด€๋ จ์„ฑ ๋ถ„์„ ํšŒ๊ท€๋ถ„์„ : ๋ณ€์ˆ˜ ๊ฐ„์˜ ์ธ๊ณผ๊ด€๊ณ„ ๋ถ„์„ ‘ํ†ต๊ณ„๋ถ„์„์˜ ๊ฝƒ’ โž” ๊ฐ€์žฅ ๊ฐ•๋ ฅํ•˜๊ณ , ๋งŽ์ด ์ด์šฉ ์ข…์†๋ณ€์ˆ˜์— ์˜ํ–ฅ์„ ๋ฏธ์น˜๋Š” ๋ณ€์ˆ˜๋ฅผ ๊ทœ๋ช…(๋ณ€์ˆ˜ ์„ ํ˜• ๊ด€๊ณ„ ๋ถ„์„) ๋…๋ฆฝ๋ณ€์ˆ˜์™€ ์ข…์†๋ณ€์ˆ˜์˜ ๊ด€๋ จ์„ฑ ๊ฐ•๋„ ๋…๋ฆฝ๋ณ€์ˆ˜์˜ ๋ณ€ํ™”์— ๋”ฐ๋ฅธ ์ข…์†๋ณ€์ˆ˜ ๋ณ€ํ™” ์˜ˆ์ธก ํšŒ๊ท€ ๋ฐฉ์ •์‹(Y=a+βX → Y:์ข…์†๋ณ€์ˆ˜, a:์ƒ์ˆ˜, β:ํšŒ๊ท€๊ณ„์ˆ˜, X:๋…๋ฆฝ๋ณ€์ˆ˜) ์„ ๋„์ถœํ•˜์—ฌ ํšŒ๊ท€์„  ์ถ”์ • ๋…๋ฆฝ๋ณ€์ˆ˜์™€ ์ข…์†๋ณ€์ˆ˜๊ฐ€ ๋ชจ๋‘ ๋“ฑ๊ฐ„์ฒ™๋„ ๋˜๋Š” ๋น„์œจ์ฒ™๋„ ๊ตฌ์„ฑ ํšŒ๊ท€ ๋ฐฉ์ •์‹ .. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 17.
  • DAY61. Tensorflow LinearRegression (1)function basic (๊ธฐ๋ณธํ•จ์ˆ˜) Tensorflow ๊ธฐ๋ณธํ•จ์ˆ˜ ์ƒ์ˆ˜ ์ƒ์„ฑ ํ•จ์ˆ˜ tf.constant(value, dtype, shape) : ์ง€์ •ํ•œ ๊ฐ’(value)์œผ๋กœ ์ƒ์ˆ˜ ์ƒ์„ฑ tf.zeros(shape, dtype) : ๋ชจ์–‘๊ณผ ํƒ€์ž…์œผ๋กœ ๋ชจ๋“  ์›์†Œ๊ฐ€ 0์œผ๋กœ ์ƒ์„ฑ tf.ones(shape, dtype) : ๋ชจ์–‘๊ณผ ํƒ€์ž…์œผ๋กœ ๋ชจ๋“  ์›์†Œ๊ฐ€ 1๋กœ ์ƒ์„ฑ tf.identity(input) : ๋‚ด์šฉ๊ณผ ๋ชจ์–‘์ด ๋™์ผํ•œ ์ƒ์ˆ˜ ์ƒ์„ฑ tf.fill(shape, value) : ์ฃผ์–ด์ง„ scalar๊ฐ’์œผ๋กœ ์ดˆ๊ธฐํ™”๋œ ์ƒ์ˆ˜ ์ƒ์„ฑ tf.linspace(start, stop, num) : start~stop ๋ฒ”์œ„์—์„œ num์ˆ˜ ๋งŒํผ ์ƒ์„ฑ tf.range(start, limit, delta) : ์‹œ์ž‘, ์ข…๋ฃŒ, ์ฆ๊ฐ ์œผ๋กœ ์ƒ์ˆ˜ ์ƒ์„ฑ ์ˆ˜ํ•™๊ด€๋ จ ํ•จ์ˆ˜ : tf.math.ํ•จ์ˆ˜() t.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 16.
  • DAY60. Tensorflow Basic (1)์„ค์น˜, ๊ธฐ๋ณธ Tensorflow ๊ฐœ์š” Google ๋จธ์‹  ๋Ÿฌ๋‹ ์ธํ…”๋ฆฌ์ „์Šค ์—ฐ๊ตฌ์†Œ์˜ ๋ธŒ๋ ˆ์ธ ํŒ€ ๊ฐœ๋ฐœ ๋จธ์‹  ๋Ÿฌ๋‹๊ณผ ๋”ฅ ๋‰ด๋Ÿด ๋„คํŠธ์›Œํฌ ์—ฐ๊ตฌ ๋ชฉ์  ์ฃผ์š” ๊ธฐ๋Šฅ ๋‹ค์ฐจ์› ๋ฐฐ์—ด(Tensor) ์ •์˜, ์ตœ์ ํ™”, ํšจ์œจ์  ์‚ฐ์ˆ  ์—ฐ์‚ฐ ๋”ฅ ๋‰ด๋Ÿด ๋„คํŠธ์›Œํฌ์™€ ๋จธ์‹  ๋Ÿฌ๋‹ ํ”„๋กœ๊ทธ๋ž˜๋ฐ ์ง€์› ๋ฉ”๋ชจ๋ฆฌ์™€ ๋ฐ์ดํ„ฐ ์ž๋™ ๊ด€๋ฆฌ, GPU ๊ฐ€์† ๊ธฐ๋Šฅ ์ œ๊ณต (ํ…์„œํ”Œ๋กœ ์ž๋™ CPU, GPU ์ž์› ํ• ๋‹น) ๋น…๋ฐ์ดํ„ฐ ์ฒ˜๋ฆฌ๋ฅผ ์œ„ํ•œ ๋Œ€๊ทœ๋ชจ ๋ณ‘๋ ฌ ์ปดํ“จํŒ… ์ง€์› CPU vs GPU CPU : ์ง๋ ฌ ์ฒ˜๋ฆฌ์— ์ตœ์ ํ™”๋œ ๋ช‡ ๊ฐœ์˜ ์ฝ”์–ด๋กœ ๊ตฌ์„ฑ GPU : ๋ณ‘๋ ฌ ์ฒ˜๋ฆฌ์šฉ์œผ๋กœ ์„ค๊ณ„๋œ ์ˆ˜ ์ฒœ ๊ฐœ์˜ ๋ณด๋‹ค ์†Œํ˜•, ํšจ์œจ์ ์ธ ์ฝ”์–ด * ์• ํ”Œ๋ฆฌ์ผ€์ด์…˜์˜ ์—ฐ์‚ฐ์ง‘์•ฝ์ ์ธ ๋ถ€๋ถ„์„ GPU๋กœ ๋„˜๊ธฐ๊ณ  ๋‚˜๋จธ์ง€ ์ฝ”๋“œ๋งŒ CPU ์ฒ˜๋ฆฌ Tensorflow ๊ฐœ๋ฐœ ํ™˜๊ฒฝ ๊ฐ€์ƒํ™˜๊ฒฝ ์‚ฌ์šฉ ์ด์œ  * ๊ฐ€์ƒํ™˜๊ฒฝ : ๊ฐ€์ƒ์˜ Python ๊ฐœ๋ฐœํ™˜๊ฒฝ ์˜๋ฏธ.. ๊ณต๊ฐ์ˆ˜ 0 ๋Œ“๊ธ€์ˆ˜ 0 2021. 12. 15.
    ๋ฌธ์˜์•ˆ๋‚ด
    • ํ‹ฐ์Šคํ† ๋ฆฌ
    • ๋กœ๊ทธ์ธ
    • ๊ณ ๊ฐ์„ผํ„ฐ

    ํ‹ฐ์Šคํ† ๋ฆฌ๋Š” ์นด์นด์˜ค์—์„œ ์‚ฌ๋ž‘์„ ๋‹ด์•„ ๋งŒ๋“ญ๋‹ˆ๋‹ค.

    ยฉ Kakao Corp.