@inproceedings{li2019gradient,
title={Gradient Harmonized Single-stage Detector},
author={Li, Buyu and Liu, Yu and Wang, Xiaogang},
booktitle={AAAI Conference on Artificial Intelligence},
year={2019}
}
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|
R-50-FPN | pytorch | 1x | 3.9 | 0.500 | 9.4 | 36.9 | model |
R-101-FPN | pytorch | 1x | 5.8 | 0.625 | 8.5 | 39.0 | model |
X-101-32x4d-FPN | pytorch | 1x | 7.0 | 0.818 | 7.6 | 40.5 | model |
X-101-64x4d-FPN | pytorch | 1x | 9.9 | 1.191 | 6.1 | 41.6 | model |