Computer-Vision-Leaderboard
The goal of this repository is:
- To keep on track of state-of-the-art (SoTA) on each vision task and new CNN architectures
- To see the comparison of famous CNN models at a glance (performance, speed, size, etc.)
- To access their research papers and implementations on different frameworks
Leaderboards
Leaderboards are provided in the below link for better accessibility and usability.
- ImageNet Classification Leaderboard
- Stanford Online Products Retrieval Leaderboard
- CARS196 Retrieval Leaderboard
- CUB-200-2011 Retrieval Leaderboard
- In-shop Clothes Retrieval Leaderboard
Updates
ImageNet Classification Leaderboard
- 181022. Update MnasNet and Big-Little Net
- 181024. Create the ImageNet Classification Leaderboard webpage for better accessibility and usability.
- 181126. Update A^2-Nets, ResNet variants from GLUON-CV, and ChannelNets
- 181130. Update FishNet and osmr/imgclsmob repo
- 190326. Update SKNet paper
- 190505. Update RandWire-WS, Res2Net, and MultiGrain papers
- 190524. Update ACNet, iSQRT-COV-Net, MobileNetV3, OctConv, and AAConv papers
- 190610. Update EfficientNet, and GPipe papers
Stanford Online Products Retrieval Leaderboard
- 181222. Create the Stanford Online Products Retrieval Leaderboard
- 190326. Update CGD paper
- 190415. Update HDML and EPSHN paper
CARS196 Retrieval Leaderboard
- 181222. Create the CARS196 Retrieval Leaderboard
- 190326. Update CGD paper
- 190415. Update HDML and EPSHN paper
CUB-200-2011 Retrieval Leaderboard
- 181222. Create the CUB-200-2011 Retrieval Leaderboard
- 190326. Update CGD paper
- 190415. Update HDML and EPSHN paper
In-shop Clothes Retrieval Leaderboard
- 190326. Create the In-shop Clothes Retrieval Leaderboard
- 190326. Update CGD paper
- 190415. Update HDML and EPSHN paper
Related Resources
Check the other good resources about CNN models
- Caffe-model
- TensorNets
- DeepDetect : Open Source Deep Learning Server & API
- Pretrained models for Pytorch
- Pretrained models for Chainer & Gluon
- Netscope CNN Analyzer
- Memory consumption and FLOP count
Author
Byung Soo Ko / [email protected]