Skip to content

Latest commit

 

History

History
59 lines (41 loc) · 5.08 KB

README.md

File metadata and controls

59 lines (41 loc) · 5.08 KB
Raw

YOLOv3: An Incremental Improvement

Abstract

We present some updates to YOLO! We made a bunch of little design changes to make it better. We also trained this new network that's pretty swell. It's a little bigger than last time but more accurate. It's still fast though, don't worry. At 320x320 YOLOv3 runs in 22 ms at 28.2 mAP, as accurate as SSD but three times faster. When we look at the old .5 IOU mAP detection metric YOLOv3 is quite good. It achieves 57.9 mAP@50 in 51 ms on a Titan X, compared to 57.5 mAP@50 in 198 ms by RetinaNet, similar performance but 3.8x faster.

Citation

@misc{redmon2018yolov3,
    title={YOLOv3: An Incremental Improvement},
    author={Joseph Redmon and Ali Farhadi},
    year={2018},
    eprint={1804.02767},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}

Results and Models

Backbone Scale Lr schd Mem (GB) Inf time (fps) box AP Config Download
DarkNet-53 320 273e 2.7 63.9 27.9 config model | log
DarkNet-53 416 273e 3.8 61.2 30.9 config model | log
DarkNet-53 608 273e 7.4 48.1 33.7 config model | log

Mixed Precision Training

We also train YOLOv3 with mixed precision training.

Backbone Scale Lr schd Mem (GB) Inf time (fps) box AP Config Download
DarkNet-53 608 273e 4.7 48.1 33.8 config model | log

Lightweight models

Backbone Scale Lr schd Mem (GB) Inf time (fps) box AP Config Download
MobileNetV2 416 300e 5.3 23.9 config model | log
MobileNetV2 320 300e 3.2 22.2 config model | log

Notice: We reduce the number of channels to 96 in both head and neck. It can reduce the flops and parameters, which makes these models more suitable for edge devices.

Credit

This implementation originates from the project of Haoyu Wu(@wuhy08) at Western Digital.