Giter Club home page Giter Club logo

yolo9000's Introduction

YOLO9000

yolo9000's People

Contributors

visionnoob avatar

Stargazers

Martin Hwang avatar

Watchers

James Cloos avatar Martin Hwang avatar  avatar Chanhee Jeong avatar

yolo9000's Issues

Repository에 기본적인 test 및 travis 관리

아래 내용 알아보고, 공부하기
그리고 적용하기.

NMS 함수 만들기

  1. Test data로 확인
  2. Test data merge가 원하는 방식으로 되는지 확인

Darknet-19 pretrained weight 에서 input normalization 문제.

pretrained weight를 가져와서 테스트해보니
input을 zero-mean으로 하면 안되고
그냥 그대로(?) to_tensor로 만들어야 정답이 제대로 나옵니다.

가령 아래와 같이 짜면 값이 안 나오고

dataset = dset.ImageFolder(root="samples/",
                           transform=transforms.Compose([
                               transforms.Resize((256,256)),
                               transforms.ToTensor(),       # Tensor로 바꾸고 (0~1로 자동으로 normalize)
                               transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                     std=[0.229, 0.224, 0.225]),
                           ]))

이렇게 해야합니다..

dataset = dset.ImageFolder(root="samples/",
                           transform=transforms.Compose([
                               transforms.Resize((256,256)),
                               transforms.ToTensor()      # Tensor로 바꾸고 (0~1로 자동으로 normalize)
                           ]))

yolo9000 paper 내용

We also shrink the network to operate on 416 input images instead of 448×448. We do this because we want an odd number of locations in our feature map so there is a single center cell. Objects, especially large objects, tend to occupy the center of the image so it’s good to have a single location right at the center to predict these objects instead of four locations that are all nearby. YOLO’s convolutional layers downsample the image by a factor of 32 so by using an input image of 416 we get an output feature map of 13 × 13

여기에서

we want an odd number of locations in our feature map
이 어떤 의미 인가요?

random Seed 관련

이번에 네이버 AI Hackathon에서 멘토님에게 받은 조언 중 하나가 실험 시 random seed 설정
이었습니다! network 랜덤 초기화 등 다양한 random value 생성 시 seed를 정해주면
성능재현의 측면에서 좋다는 조언을 받았습니다!

K-means Cluster 개발

  1. Test data만들기

    • 임의의 가우시안 분포 데이터 자동 generator
  2. Test

    • K-mean cluster 동작시켜서 잘 Clustering하는지 확인

LOSS 함수 구현

  1. Forward output test 작성 후, 확인
  2. Test 확인은 손으로 풀어본 값 맞추기

IOU 계산

  1. Test data 만들기

    • 2개 이미지의 레이블 계산
  2. Test data 통과하는지 확인

Dataloader label tensor block 만들기

YOLO v2 output에 맞춰서 label 만들기
reference는 재원님 명세 확인

  1. Test data

    • 변환된 output tensor block GT visualizer 짜기
  2. Test

    • 변환된 output tensor block과 실제 GT 비교

PEP8 적용

이번 프로젝트에도 PEP8 적용 관련 스크립트가 추가되면 좋을 것 같습니다.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.