Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection
Introduction
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}
}
Results and Models
Backbone |
Style |
Lr schd |
Multi-scale Training |
Inf time (fps) |
box AP |
Download |
R-50 |
pytorch |
1x |
No |
19.5 |
40.2 |
model | log |
R-50 |
pytorch |
2x |
Yes |
19.5 |
42.9 |
model | log |
R-101 |
pytorch |
2x |
Yes |
14.7 |
44.7 |
model | log |
R-101-dcnv2 |
pytorch |
2x |
Yes |
12.9 |
47.1 |
model | log |
X-101-32x4d |
pytorch |
2x |
Yes |
12.1 |
45.9 |
model | log |
X-101-32x4d-dcnv2 |
pytorch |
2x |
Yes |
10.7 |
48.1 |
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.