๋ณธ๋ฌธ ๋ฐ”๋กœ๊ฐ€๊ธฐ

๋ถ„๋ฅ˜ ์ „์ฒด๋ณด๊ธฐ74

๊ฐ•ํ™”ํ•™์Šต ๋ณต์Šต ์ž๋ฃŒ 1: Concept of RL ๊ฐ•ํ™”ํ•™์Šต์€ ๊ธฐ๋ณธ์ ์œผ๋กœ, ์ฅ(Actor ํ˜น์€ Agent)๊ฐ€ action์„ ์ทจํ•˜๋ฉด ๊ทธ์— ๋”ฐ๋ฅธ reward๋ฅผ ๋ฐ›๊ณ , ๋ณ€ํ™”๋œ state๋ฅผ ๊ด€์ฐฐํ•˜์—ฌ ๋‹ค์‹œ action์„ ์ทจํ•˜๋Š” ๋ฐฉ์‹์ด๋‹ค. Q ํ•จ์ˆ˜์— state์™€ action์„ ์ฃผ๋ฉด, ๊ทธ์— ๋”ฐ๋ฅธ reward๋ฅผ ๋ฆฌํ„ดํ•œ๋‹ค๊ณ  ์ƒ๊ฐํ•˜์ž. (๊ทธ๋Ÿฌํ•œ Q ํ˜•๋‹˜์ด ์žˆ๋‹ค๊ณ  ๊ฐ€์ •ํ•˜์ž.) Agent๊ฐ€ ์•Œ๊ณ  ์‹ถ์€ ๊ฒƒ์€ ์ตœ๋Œ€ reward๋ฅผ ๋งŒ๋“œ๋Š” action์ด๋‹ค. ์ด ๋‚ด์šฉ์„ ๋‹ค์Œ ๋‘ ์ˆ˜ํ•™์  ํ‘œ๊ธฐ๋กœ ๋‚˜ํƒ€๋‚ผ ์ˆ˜ ์žˆ๋‹ค. max(a') Q(s, a') : s ๋ผ๋Š” state์— a'์„ ๋ฐ”๊พธ์–ด ์คŒ์œผ๋กœ์จ ์–ป์„ ์ˆ˜ ์žˆ๋Š” ์ตœ๋Œ€ reward๊ฐ’ (Q๊ฐ’) argmax(a') Q(s, a'): (์œ„์™€ ์ด์–ด์ง€๋Š” ์ƒํ™ฉ์—์„œ) ์ตœ๋Œ€ Q๊ฐ’์„ ๊ฐ–๊ฒŒํ•˜๋Š” argument a' Reference: Sung Kim ๋ชจ๋‘๋ฅผ ์œ„ํ•œ R.. 2022. 3. 12.
django ์‹œ์ž‘ํ•˜๊ธฐ 1. pipenv shell (๋ฒ„๋ธ” ์•ˆ์œผ๋กœ ๋“ค์–ด๊ฐ€๊ธฐ) 2. python3 manage.py runserver 3. Ctrl+C๋กœ ์„œ๋ฒ„ํ‚ฌํ•˜๊ณ  python3 manage.py migrate 4. python3 manage.py createsuperuser 5. python3 manage.py runserver๋กœ ์„œ๋ฒ„ ๊ตฌ๋™ํ•˜๋ฉด 127.0.0.1:8000/localhost ์—์„œ ๋กœ๊ทธ์ธ ๊ฐ€๋Šฅ => ๊ด€๋ฆฌ์ž ํŒจ๋„ ์‚ฌ์šฉ ๊ฐ€๋Šฅ 2022. 1. 4.
wsl๋กœ django ํ”„๋กœ์ ํŠธ ์‹œ์ž‘ํ•˜๊ธฐ (2) #wsl #python #pipenv #django #git Before Step: wsl๋กœ django ํ”„๋กœ์ ํŠธ ์‹œ์ž‘ํ•˜๊ธฐ 1. django-admin startproject ์ปค๋งจ๋“œ๋กœ ์žฅ๊ณ  ํ”„๋กœ์ ํŠธ๋ฅผ ์‹œ์ž‘ - config ํด๋”๊ฐ€ ์ƒ๊ธฐ๊ณ  ๊ทธ ๋‚ด๋ถ€์˜ config ํด๋”, manage.py ํŒŒ์ผ์ด ๋ณด์ž„ - ๋‚ด๋ถ€์˜ config ํด๋”, manage.py ํŒŒ์ผ์„ ๋ฐ”๊นฅ์œผ๋กœ ๊บผ๋‚ด์คŒ โ€ป ์•„๋ž˜ ์บก์ฒ˜์˜ ๊ฒฝ๋กœ ๋Œ€์‹  /home/(username)/airbnb-clone 2. VSCode Extensions์—์„œ Python์„ ๋‹ค์šด๋กœ๋“œ (Remote-WSL์—์„œ Install) 3. VSCode ํ™”๋ฉด ํ•˜๋‹จ์—์„œ Python ๋ฒ„์ „์„ pipenv๋ผ๊ณ  ์จ์žˆ๋Š” ๋ฒ„์ „์œผ๋กœ ์„ ํƒ 4. Shift+Ctrl+P์—์„œ "Python: Select Lint.. 2022. 1. 2.
wsl๋กœ django ํ”„๋กœ์ ํŠธ ์‹œ์ž‘ํ•˜๊ธฐ #wsl #python #pipenv #django #git Next Step: wsl๋กœ django ํ”„๋กœ์ ํŠธ ์‹œ์ž‘ํ•˜๊ธฐ (2) 1. PowerShell์—์„œ wsl ์„ค์น˜ํ•˜๊ณ  ์‹œ์Šคํ…œ ๋‹ค์‹œ ์‹œ์ž‘ 2. ์ž๋™์œผ๋กœ ์—ด๋ฆฌ๋Š” ubuntu shell์—์„œ username/password ์„ค์ • 3. cmd๋‚˜ PowerShell์—์„œ bash ์ปค๋งจ๋“œ๋ฅผ ์ž…๋ ฅํ•˜๋ฉด ubuntu shell์„ ์ด์šฉํ•  ์ˆ˜ ์žˆ์Œ 4. python ์ปค๋งจ๋“œ ํ™•์ธ - python3 ์ปค๋งจ๋“œ๋กœ python ์‹คํ–‰ ๊ฐ€๋Šฅ (python 3.8.10์ด ๊ธฐ๋ณธ ์„ค์น˜๋จ) - python ์ปค๋งจ๋“œ๋กœ๋Š” python ์‹คํ–‰ ๋ถˆ๊ฐ€: PATH ๋˜๋Š” Alias ์„ค์ •์ด ํ•„์š” 5. python3๊ฐ€ ์„ค์น˜๋œ ๊ฒฝ๋กœ ํ™•์ธ - ๋Œ€์ฒด๋กœ /usr/bin/python ๋˜๋Š” /usr/bin/python3 ๋˜๋Š”.. 2021. 12. 26.