๋ณธ๋ฌธ ๋ฐ”๋กœ๊ฐ€๊ธฐ

728x90

Keras

4/18 ์›” ์›”์š”์ผ! ์˜ค๋Š˜์€ ๋ณต์žกํ•œ ์ด๋ฏธ์ง€ ํ•™์Šต(์บ๊ธ€์˜ ๊ฐœ์™€ ๊ณ ์–‘์ด ์˜ˆ์ œ)๊ณผ Generator๋ฅผ ๋ฐฐ์šด๋‹ค. 1. ์ด๋ฏธ์ง€ ํŒŒ์ผ → csv ํŒŒ์ผ๋กœ ๋ณ€ํ™˜ jpg ํŒŒ์ผ์„ ์ฝ์–ด์„œ RGB pixel ๊ฐ’์„ ์–ป์–ด๋‚ด๊ณ (decoding) ์‹ค์ˆ˜๋กœ ๋ณ€ํ™˜ํ•œ ๋‹ค์Œ ์ •๊ทœํ™” ์ž‘์—…์„ ์œ„ํ•ด ๊ด€๋ฆฌ์ž ๊ถŒํ•œ์œผ๋กœ tqdm(์ƒํƒœ ์ง„ํ–‰๋ฅ  ์•Œ๋ ค์ฃผ๋Š” ํ”„๋กœ๊ทธ๋ ˆ์Šค ๋ฐ” ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ)๊ณผ ipywidgets, ์ด๋ฏธ์ง€ ์ฒ˜๋ฆฌ๋ฅผ ์œ„ํ•œ opencv ์„ค์น˜ @ Anaconda Prompt conda install -c conda-forge tqdm conda install -c conda-forge ipywidgets pip install opencv-python jupyter notebook @ Jupyter Notebook import numpy as np import pandas.. ๋”๋ณด๊ธฐ
4/8 ๊ธˆ ๊ธˆ์š”์ผ! ๐Ÿฑ‍๐Ÿ ์˜ค๋Š˜์€ Regression์„ ๋๋‚ธ๋‹ค~~ 4/11 ์›”์š”์ผ์€ ๋จธ์‹ ๋Ÿฌ๋‹ ํ•„๋‹ต ํ‰๊ฐ€, 4/17 ์ผ์š”์ผ์€ ์ˆ˜ํ–‰ํ‰๊ฐ€ 4๊ฐ€์ง€ ์ œ์ถœ์ด ์žˆ๋‹ค. ๊ฒฐ์ธก์น˜ ์ฒ˜๋ฆฌ๋Š” ์‚ญ์ œํ•˜๊ฑฐ๋‚˜, imputation(๋ณด๊ฐ„, ๋Œ€์ฒด) - ํ‰๊ท ํ™” ๊ธฐ๋ฒ•(๋…๋ฆฝ๋ณ€์ˆ˜๋ฅผ ๋Œ€ํ‘œ๊ฐ’์œผ๋กœ ๋Œ€์ฒด), ๋จธ์‹ ๋Ÿฌ๋‹ ๊ธฐ๋ฒ•(์ข…์†๋ณ€์ˆ˜๊ฐ€ ๋Œ€์ƒ. KNN) KNN(K-Nearest Neighbors, K-์ตœ๊ทผ์ ‘ ์ด์›ƒ) : hyperparameter๋Š” k(=1์ผ ๋•Œ ์–ด๋Š ์ •๋„์˜ ์„ฑ๋Šฅ ๋ณด์žฅ)์™€ ๊ฑฐ๋ฆฌ์ธก์ • ๋ฐฉ์‹(์ฃผ๋กœ ์œ ํด๋ผ๋””์•ˆ ์‚ฌ์šฉ) ๋ฐ˜๋“œ์‹œ ์ •๊ทœํ™”๋ฅผ ์ง„ํ–‰ํ•ด์•ผ ํ•จ. ๋ชจ๋“  ๋ฐ์ดํ„ฐ์— ๋Œ€ํ•ด ๊ฑฐ๋ฆฌ๋ฅผ ๊ณ„์‚ฐํ•ด์•ผ ํ•˜๋ฏ€๋กœ ์‹œ๊ฐ„์ด ์˜ค๋ž˜ ๊ฑธ๋ฆด ์ˆ˜ ์žˆ์Œ 1. Logistic Regression + KNN - BMI data import numpy as np import pandas as pd fro.. ๋”๋ณด๊ธฐ
4/7 ๋ชฉ ๋ชฉ์š”์ผ! ์˜ค๋Š˜๋„ Multinomial Classification๋ฅผ ๋Œ€ํ‘œ์ ์ธ ์˜ˆ์ œ(MNIST)๋ฅผ ํ†ตํ•ด ๋ฐฐ์šด๋‹ค~ ์†์œผ๋กœ ์“ด ์ˆซ์ž๋“ค๋กœ ์ด๋ฃจ์–ด์ง„ ๋Œ€ํ˜• ๋ฐ์ดํ„ฐ๋ฒ ์ด์Šค~ MNIST ์ด๋ฏธ์ง€๋Š” ๊ทธ ์ž์ฒด๊ฐ€ 2์ฐจ์›์ด๊ณ  ๊ทธ๋Ÿฐ ๊ฒŒ ์—ฌ๋Ÿฟ์ด๊ธฐ ๋•Œ๋ฌธ์— 3์ฐจ์›. ์ด๋ฏธ์ง€๋ฅผ 1์ฐจ์›์œผ๋กœ ravel() ํ•ด์•ผ ํ•จ https://www.kaggle.com/competitions/digit-recognizer/data?select=test.csv Digit Recognizer | Kaggle www.kaggle.com Tensorflow Ver. 1.15์€ ๋ฐฐ์šด ์ด๋ก ์„ ์ฝ”๋“œ๋กœ ์ดํ•ดํ•˜๊ธฐ์—๋Š” ์ข‹์ง€๋งŒ ์ฝ”๋“œ๊ฐ€ ๋„ˆ๋ฌด ์–ด๋ ต๋‹ค. 1. Multinomial Classification by Tensorflow Ver. 1.15 - MNIST import nump.. ๋”๋ณด๊ธฐ

728x90