| 일 | 월 | 화 | 수 | 목 | 금 | 토 |
|---|---|---|---|---|---|---|
| 1 | ||||||
| 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| 9 | 10 | 11 | 12 | 13 | 14 | 15 |
| 16 | 17 | 18 | 19 | 20 | 21 | 22 |
| 23 | 24 | 25 | 26 | 27 | 28 | 29 |
| 30 |
Tags
- matlab 영상처리
- c언어 정리
- 딥러닝 스터디
- 해리스 코너 검출
- tensorflow 예제
- c언어
- C언어 공부
- 팀프로젝트
- 김성훈 교수님 PyTorch
- 영상처리
- 모두의 딥러닝 예제
- c++공부
- 파이토치 김성훈 교수님 강의 정리
- Pytorch Lecture
- MFC 프로그래밍
- pytorch zero to all
- 파이토치 강의 정리
- pytorch
- TensorFlow
- 딥러닝
- object detection
- 미디언 필터링
- 가우시안 필터링
- c++
- 모두의 딥러닝
- 컴퓨터 비전
- 케라스 정리
- 골빈해커
- 딥러닝 공부
- 파이토치
Archives
- Today
- Total
목록Module Cardinarity (1)
ComputerVision Jack
Aggregated Residual Transformations for Deep Neural Networks Abstract 논문에서 소개하는 Network는 같은 Topology에 대해 변형 시킨 Building Block 반복하여 설계하였다. 이 설계는 적은 Hyper-Parameter 통해서 동종의 Multi-Branch한 구조의 결과로 나타났다. 앞에서 소개한 전략은 "Cardinality" (the size of the set of transformations) 새로운 Dimension 생성한다. 이 Dimension은 Depth와 Width Dimension 외에도 필수적인 요소이다. 게다가 Capacity 위해 Network 깊고, 넓게 만드는 것 보다, Cardinality 증가 시키는 것..
Reading Paper/Classification Networks
2021. 12. 21. 09:55