mirrors
/
PyRetri
mirror of https://github.com/PyRetri/PyRetri.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
PyRetri/docs/MODEL_ZOO.md

61 lines
3.6 KiB
Markdown
Raw Permalink Normal View History

upload
2020-04-02 14:00:49 +08:00
# Model Zoo
upload
2020-04-17 22:37:22 +08:00
Here we provide the pre-trained models to help you reproduce our experimental results easily.
upload
2020-04-02 14:00:49 +08:00
## General image retrieval
### pre-trained models
| Training Set | Backbone | for Short | Download |
| :------------------: | :-------: | :-------: | :----------------------------------------------------------: |
| ImageNet | VGG-16 | I-VGG16 | [model](https://download.pytorch.org/models/vgg16-397923af.pth) |
| Places365 | VGG-16 | P-VGG16 | [model](https://drive.google.com/open?id=1U_VWbn_0L9mSDCBGiAIFbXxMvBeOiTG9) |
| ImageNet + Places365 | VGG-16 | H-VGG16 | [model](https://drive.google.com/open?id=11zE5kGNeeAXMhlHNv31Ye4kDcECrlJ1t) |
| ImageNet | ResNet-50 | I-Res50 | [model](https://download.pytorch.org/models/resnet50-19c8e357.pth) |
| Places365 | ResNet-50 | P-Res50 | [model](https://drive.google.com/open?id=1lp_nNw7hh1MQO_kBW86GG8y3_CyugdS2) |
| ImageNet + Places365 | ResNet-50 | H-Res50 | [model](https://drive.google.com/open?id=1_USt_gOxgV4NJ9Zjw_U8Fq-1HEC_H_ki) |
### performance
| Dataset | Data Augmentation | Backbone | Pooling | Dimension Process | mAP |
| :--------: | :------------------------: | :------: | :-----: | :------------------: | :--: |
| Oxford5k | ShorterResize + CenterCrop | H-VGG16 | GAP | l2 +SVD(whiten) + l2 | 62.9 |
| CUB-200 | ShorterResize + CenterCrop | I-Res50 | SCDA | l2 + PCA + l2 | 27.8 |
| Indoor | DirectResize | P-Res50 | CroW | l2 + PCA + l2 | 51.8 |
| Caltech101 | PadResize | I-Res50 | GeM | l2 + PCA + l2 | 77.9 |
Choosing the implementations mentioned above as baselines and adding some tricks, we have:
| Dataset | Implementations | mAP |
| :--------: | :--------------------------------: | :--: |
| Oxford5k | baseline + K-reciprocal | 72.9 |
| CUB-200 | baseline + K-reciprocal | 38.9 |
| Indoor | baseline + DBA + QE | 63.7 |
| Caltech101 | baseline + DBA + QE + K-reciprocal | 86.1 |
## Person re-identification
For person re-identification, we use the model provided by [Person_reID_baseline](https://github.com/layumi/Person_reID_baseline_pytorch) and reproduce its resutls. In addition, we train a model on DukeMTMC-reID through the open source code for further experiments.
upload
2020-04-17 20:31:26 +08:00
### pre-trained models
upload
2020-04-02 14:00:49 +08:00
| Training Set | Backbone | for Short | Download |
| :-----------: | :-------: | :-------: | :------: |
| Market-1501 | ResNet-50 | M-Res50 | [model](https://drive.google.com/open?id=1-6LT_NCgp_0ps3EO-uqERrtlGnbynWD5) |
| DukeMTMC-reID | ResNet-50 | D-Res50 | [model](https://drive.google.com/open?id=1X2Tiv-SQH3FxwClvBUalWkLqflgZHb9m) |
### performance
| Dataset | Data Augmentation | Backbone | Pooling | Dimension Process | mAP | Recall@1 |
| :-----------: | :--------------------: | :------: | :-----: | :---------------: | ---- | :------: |
| Market-1501 | DirectResize + TwoFlip | M-Res50 | GAP | l2 | 71.6 | 88.8 |
| DukeMTMC-reID | DirectResize + TwoFlip | D-Res50 | GAP | l2 | 62.5 | 80.4 |
Choosing the implementations mentioned above as baselines and adding some tricks, we have:
| Dataset | Implementations | mAP | Recall@1 |
| :-----------: | :-------------------------------------: | :--: | :------: |
| Market-1501 | Baseline + l2 + PCA + l2 + K-reciprocal | 84.8 | 90.4 |
| DukeMTMC-reID | Baseline + l2 + PCA + l2 + K-reciprocal | 78.3 | 84.2 |