YOLOv8 初体验
简介
YOLOv8模型设计快速,准确,易于使用,使其成为广泛的目标检测和图像分割任务的绝佳选择。
The YOLOv8 model is designed to be fast, accurate, and easy to use, making it an excellent choice for a wide range of object detection and image segmentation tasks.
它可以在大型数据集上进行训练,并能够在各种硬件平台上运行,从cpu到gpu。
It can be trained on large datasets and is capable of running on a variety of hardware platforms, from CPUs to GPUs.
YOLO 简史
YOLO (You Only Look Once)是一个流行的目标检测和图像分割模型,由华盛顿大学的 Joseph Redmon 和 Ali Farhadi开发。
YOLO (You Only Look Once) is a popular object detection and image segmentation model developed by Joseph Redmon and Ali Farhadi at the University of Washington.
YOLO的第一个版本于2015年发布,并因其高速和准确而迅速受到欢迎。
The first version of YOLO was released in 2015 and quickly gained popularity due to its high speed and accuracy.
YOLOv2于2016年发布,并通过组合 batch normalization、anchor boxes 和 dimension clusters 对原始模型进行了改进。
YOLOv2 was released in 2016 and improved upon the original model by incorporating batch normalization, anchor boxes, and dimension clusters.
YOLOv3于2018年发布,通过使用更高效的 backbone 网络、添加特征金字塔和利用 focal loss进一步提高了模型的性能。
YOLOv3 was released in 2018 and further improved the model's performance by using a more efficient backbone network, adding a feature pyramid, and making use of focal loss.
2020年,YOLOv4发布,引入了许多创新,如使用 Mosaic data augmentation 、新的 anchor-free detection head 和新的损失函数。
In 2020, YOLOv4 was released which introduced a number of innovations such as the use of Mosaic data augmentation, a new anchor-free detection head, and a new loss function.
2021年,Ultralytics发布了YOLOv5,进一步提高了模型的性能,并添加了新功能,如支持全景分割和目标跟踪。
In 2021, Ultralytics released YOLOv5, which further improved the model's performance and added new features such as support for panoptic segmentation and object tracking.
YOLO已经广泛应用于各种应用,包括自动驾驶、安全和监控以及医疗成像。
YOLO has been widely used in a variety of applications, including autonomous vehicles, security and surveillance, and medical imaging.
它还被用于赢得一些比赛,如 COCO目标检测挑战赛 和 DOTA 目标检测挑战赛。
It has also been used to win several competitions, such as the COCO Object Detection Challenge and the DOTA Object Detection Challenge.
有关YOLO的历史和发展的更多信息,可以参考以下参考资料:
-
Redmon, J., & Farhadi, A. (2015). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779-788).
-
Redmon, J., & Farhadi, A. (2016). YOLO9000: Better, faster, stronger. In Proceedings
Ultralytics YOLOv8
Ultralytics YOLOv8 是 Ultralytics公司 开发的YOLO目标检测和图像分割模型的最新版本。
Ultralytics YOLOv8 Ultralytics YOLOv8 is the latest version of the YOLO object detection and image segmentation model developed by Ultralytics.
YOLOv8是一种前沿的、最先进的(SOTA)模型,建立在之前YOLO版本的成功基础上,并引入了新功能和改进,以进一步提高性能和灵活性。
YOLOv8 is a cutting-edge, state-of-the-art (SOTA) model that builds upon the success of previous YOLO versions and introduces new features and improvements to further boost performance and flexibility.
YOLOv8的一个关键特性是它的可扩展性。
One key feature of YOLOv8 is its extensibility.
它被设计为一个框架,支持所有以前版本的YOLO,使它可以轻松地在不同版本之间切换并比较它们的性能。
It is designed as a framework that supports all previous versions of YOLO, making it easy to switch between different versions and compare their performance.
对于那些想要利用最新的YOLO技术,同时仍然能够使用现有YOLO模型的用户来说,YOLOv8是一个理想的选择。
This makes YOLOv8 an ideal choice for users who want to take advantage of the latest YOLO technology while still being able to use their existing YOLO models.
除了可扩展性,YOLOv8还包括许多其他创新,使其成为广泛的目标检测和图像分割任务的诱人选择。
In addition to its extensibility, YOLOv8 includes a number of other innovations that make it an appealing choice for a wide range of object detection and image segmentation tasks.
这些包括: 一个新的 backbone 网络,一个新的 anchor-free 检测头和一个新的损失函数
。
These include a new backbone network, a new anchor-free detection head, and a new loss function.
YOLOv8也非常高效,可以在各种硬件平台上运行,从cpu到gpu。
YOLOv8 is also highly efficient and can be run on a variety of hardware platforms, from CPUs to GPUs.
总的来说,YOLOv8 是一个强大而灵活的工具,用于目标检测和图像分割,它提供了两个领域的精华: 最新的SOTA技术以及使用和比较所有以前的YOLO版本的能力
。
Overall, YOLOv8 is a powerful and flexible tool for object detection and image segmentation that offers the best of both worlds: the latest SOTA technology and the ability to use and compare all previous YOLO versions.
参考:
总结
核心我已经 使用反引号 做了标记,即:
- 一个新的 backbone 网络
- 一个新的 anchor-free 检测头
- 一个新的损失函数
看代码仓库,内部包含了:v3,v5,v8 模型。看来都是集成的自己开发的亲儿子啊。
特点
-
SOTA
-
支持任务:物体检测、跟踪、实例分割、图像分类和姿态估计
-
支持安装: pip、conda、 git clone、Docker
-
支持使用方式:命令行、python 脚本
-
集成:
- Roboflow: 标记自定义数据集并导出至 YOLOv8 进行训练
- ClearML:自动跟踪、可视化,甚至远程训练 YOLOv8
- Comet: 保存 YOLOv8 模型、恢复训练,并以交互式方式查看和调试预测
- Neural Magic以及OpenVINO:推理加速
-
Ultralytics HUB: 该公司推出的 无需代码就能解决这些任务的平台。