YOLOv7YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors译读
Posted songyuc
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了YOLOv7YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors译读相关的知识,希望对你有一定的参考价值。
YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
摘要
在5FPS到160FPS的范围内,YOLOv7从速度和精度上超越了目前所有的检测器,且在所有一直的实时检测器中达到了最高的精度56.8%AP,在 V100 GPU上的速度超过了30FPS。YOLOv7-E6检测器(56 FPS V100, 55.9% AP)超越了基于transformer的检测器 SWIN-L Cascade-Mask R-CNN(9.2 FPS A100, 53.9% AP),速度提升509%,精度提升2%;同时超越了基于卷积的检测器 ConvNeXt-XL Cascade-Mask R-CNN(8.6 FPS A100, 55.2% AP),速度提升551%,精度上提高0.7%AP值;YOLOv7在速度和精度上还超越了YOLOR, YOLOX, Scaled-YOLOv4, YOLOv5, DETR, Deformable DETR, DINO-5scale-R50, ViT-Adapter-B以及许多其它的目标检测器。此外,本文只在 MS COCO数据集上从头训练了YOLOv7,而没有使用其它数据集或预训练权重。源代码位于https://github.com/WongKinYiu/yolov7。
以上是关于YOLOv7YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors译读的主要内容,如果未能解决你的问题,请参考以下文章
tf.trainable_variables() and tf.all_variables()
YOLOv7快速复现 demo演示YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object
论文阅读(Lukas Neumann——ICCV2017Deep TextSpotter_An End-to-End Trainable Scene Text Localization and R
论文阅读——D2-Net: A Trainable CNN for Joint Description and Detection of Local Features
YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors