/final-project-level3-cv-02

πŸ‘ΆπŸ» MLOpsλ₯Ό ν™œμš©ν•œ 2μ„Έ μ–Όκ΅΄ 예츑 μ„œλΉ„μŠ€

Primary LanguagePython

πŸ‘ΆπŸ» MLOpsλ₯Ό ν™œμš©ν•œ 2μ„Έ μ–Όκ΅΄ 예츑 μ„œλΉ„μŠ€

All Contributors

🍼 Project Summary

  • StyleGAN 기반의 생성 λͺ¨λΈμΈ BabyGAN을 μ‚¬μš©ν•˜μ—¬ 두 μ‚¬λžŒμ˜ 얼꡴을 μž…λ ₯으둜 λ°›μ•„ 2μ„Έ 얼꡴을 ν•©μ„±ν•˜λŠ” μ„œλΉ„μŠ€μž…λ‹ˆλ‹€.
  • λͺ¨λΈ μ„œλΉ™κ³Ό 배포 κ²½ν—˜μ„ μœ„ν•œ ν”„λ‘œμ νŠΈμ΄λ©°, FastAPI, Airflow, AWS, Docker λ₯Ό μ‚¬μš©ν•˜μ˜€μŠ΅λ‹ˆλ‹€.

🍼 Repository Structure

  • Front-end, Back-end, Airflow, Docker
β”œβ”€β”€ README.md
β”œβ”€β”€ .env
β”œβ”€β”€ airflow
β”‚   └── dag
β”œβ”€β”€ back-end
β”‚   β”œβ”€β”€ api
β”‚   β”‚   β”œβ”€β”€ controller
β”‚   β”‚   β”œβ”€β”€ model
β”‚   β”‚   └── service
β”‚   β”œβ”€β”€ babygan
β”‚   └── main
β”œβ”€β”€ docker
└── front-end

🍼 Architecture Flow Map

final_project


🍼 Demo

데λͺ¨λ§ν¬ (12μ›” 25일 μ΄ν›„λ‘œ babygan의 inference κΈ°λŠ₯이 μ€‘λ‹¨λ©λ‹ˆλ‹€.)

  • Dashboard : airflow의 batchjob을 λΆ„μ„ν•œ κ²°κ³Όλ₯Ό λ³΄μ—¬μ£ΌλŠ” νŽ˜μ΄μ§€ dashboard
  • BabyGAN : μ‚¬μš©μžμ˜ μž…λ ₯ 이미지에 λ”°λ₯Έ 인퍼런슀 결과물을 λ³΄μ—¬μ£ΌλŠ” νŽ˜μ΄μ§€
    • input
    babygan_input * output & share 스크란샷 2021-12-23 α„‹α…©α„Œα…₯ᆫ 12 43 29
  • Gallery : μ‚¬μš©μžκ°€ κ³΅μœ ν•œ 결과물을 λͺ¨μ•„ λ³Ό 수 μžˆλŠ” νŽ˜μ΄μ§€ 스크란샷 2021-12-23 α„‹α…©α„Œα…₯ᆫ 12 45 50

🍼 Contributors

Thanks goes to these wonderful people (emoji key):


Jihun Heo
πŸ“† πŸ’» πŸ“–
Hyundong Jo
πŸ’» πŸ“–
Sejong Jeong
πŸ’» πŸ“–
Yujin Lee
πŸ’» πŸ“–
Hanbin Lee
πŸ’» πŸ“–
Seowon Kim
πŸ’» πŸ“–

This project follows the all-contributors specification. Contributions of any kind welcome!


🍼 Reference