Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

#11

Closed
Hiram1026 opened this issue Apr 22, 2022 · 10 comments
Closed

#11

Hiram1026 opened this issue Apr 22, 2022 · 10 comments

Comments

@Hiram1026
Copy link

Hiram1026 commented Apr 22, 2022

No description provided.

@SantJay
Copy link
Collaborator

SantJay commented Apr 22, 2022

您好,可以提供一下您看的那片文章吗?我们目前发布的数据集在CVF openaccess版本的基础上做过更新,在更新后数据集上跑出来的指标确实会高一些。

@Hiram1026
Copy link
Author

好的 https://github.com/DocF/multispectral-object-detection
顺便请问一下你们CVF openaccess的文章能提供一下吗?

@SantJay
Copy link
Collaborator

SantJay commented Apr 23, 2022

您提供的这篇文章确实用了CVF openaccess版本LLVIP上的旧数据。其实应该以我们GitHub上的新数据为准,新数据也同步在了Arxiv版本LLVIP上。

@Hiram1026
Copy link
Author

目前从你们homepage取得的应该是最新的吧,顺便请问一下新老数据有没有明显的区分方式

@SantJay
Copy link
Collaborator

SantJay commented Apr 24, 2022

我们从未发布过旧数据集,从homepage上获取的就是最新的:)。

@SantJay
Copy link
Collaborator

SantJay commented May 6, 2022

我们使用的显卡是Tesla T4,测试模型使用的last.pt,imgz=1280,batchsize = 8,模型结构yolov5l。

@XueZ-phd
Copy link

XueZ-phd commented Nov 16, 2022

请问下你们训练时的配置,什么显卡,最后是用的best.pt吗?

我在yolo上跑,自己用的rtx3090 imgz=1280 batchsize=8,我跑出了下面的结果,这个结果较高,我有点怀疑我的实验结果 image 我是用的best.pt基本上在50个epoch内就能取到了。

image

我们使用的显卡是Tesla T4,测试模型使用的last.pt,imgz=1280,batchsize = 8,模型结构yolov5l。

我也有同样的问题。我想知道作者在arxiv v3版本中使用的yolov5l是什么版本的?正如 @Hiram1026 所言,不同版本的yolov5l获得的结果是不同的,规律基本上是越新的版本结果越好。
我将作者arxiv上3个版本中的结果汇总如下,yolov5不同版本的时间也汇总如下。按照作者发表的时间,我猜测使用的是YOLOv5L 5.0 version。为了能够与作者的结果进行公平对比,我想知道作者的YOLOv5L版本号是什么。
image

image

@SantJay
Copy link
Collaborator

SantJay commented Nov 17, 2022

我们Arxiv v3版本中使用的yolov5l版本为version5.0,各位可以参考一下。

@XueZ-phd
Copy link

我们Arxiv v3版本中使用的yolov5l版本为version5.0,各位可以参考一下。

感谢回复!

@Hiram1026 Hiram1026 changed the title 您好,请问下你们yolov5的检测指标是在yolov5 v6.0版本上跑出来的吗? Nov 25, 2022
@Hiram1026 Hiram1026 closed this as not planned Won't fix, can't repro, duplicate, stale Nov 25, 2022
@heitorrapela
Copy link

heitorrapela commented Nov 26, 2022

I saw that here you have test other versions of yolo. I have these results with the yolov5l (version 7.0) (commit: bfa1f23) on llvip infrared trained with the hparams from this repo.

infrared single modality
image

rgb single modality
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants