| Architecture | Generalized Focal Loss, Deformable Convolution, ResNet |
|---|---|
| lr sched | 2x |
| SHOW MORE |
| Architecture | ResNeXt, Generalized Focal Loss, Deformable Convolution |
|---|---|
| lr sched | 2x |
| SHOW MORE |
[ALGORITHM]
We provide config files to reproduce the object detection results in the paper Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection
@article{li2020generalized,
title={Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection},
author={Li, Xiang and Wang, Wenhai and Wu, Lijun and Chen, Shuo and Hu, Xiaolin and Li, Jun and Tang, Jinhui and Yang, Jian},
journal={arXiv preprint arXiv:2006.04388},
year={2020}
}
| Backbone | Style | Lr schd | Multi-scale Training | Inf time (fps) | box AP | Config | Download |
|---|---|---|---|---|---|---|---|
| R-50 | pytorch | 1x | No | 19.5 | 40.2 | config | model | log |
| R-50 | pytorch | 2x | Yes | 19.5 | 42.9 | config | model | log |
| R-101 | pytorch | 2x | Yes | 14.7 | 44.7 | config | model | log |
| R-101-dcnv2 | pytorch | 2x | Yes | 12.9 | 47.1 | config | model | log |
| X-101-32x4d | pytorch | 2x | Yes | 12.1 | 45.9 | config | model | log |
| X-101-32x4d-dcnv2 | pytorch | 2x | Yes | 10.7 | 48.1 | config | model | log |
[1] 1x and 2x mean the model is trained for 90K and 180K iterations, respectively. \
[2] All results are obtained with a single model and without any test time data augmentation such as multi-scale, flipping and etc.. \
[3] dcnv2 denotes deformable convolutional networks v2. \
[4] FPS is tested with a single GeForce RTX 2080Ti GPU, using a batch size of 1.
Object Detection on COCO minival
| MODEL | BOX AP |
|---|---|
| Generalized Focal Loss (X-101-32x4d-dcnv2, 2x, pytorch) | 48.1 |
| Generalized Focal Loss (R-101-dcnv2, 2x, pytorch) | 47.1 |
| Generalized Focal Loss (X-101-32x4d, 2x, pytorch) | 45.9 |
| Generalized Focal Loss (R-101, 2x, pytorch) | 44.7 |
| Generalized Focal Loss (R-50, 2x, pytorch) | 42.9 |
| Generalized Focal Loss (R-50, 1x, pytorch) | 40.2 |
