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

yolov5s fire-smoke weight is not the best! #8

Open
gengyanlei opened this issue Nov 6, 2020 · 5 comments
Open

yolov5s fire-smoke weight is not the best! #8

gengyanlei opened this issue Nov 6, 2020 · 5 comments

Comments

@gengyanlei
Copy link
Owner

注释:
yolov5s 烟火检测模型是我随便训练几个epoch的结果,并不是最好的结果。目前只能开源1个看起来还说的过去的模型。最好的模型不会开源的。

@yutao007
Copy link

yutao007 commented Feb 7, 2021

谢谢您,您应该不是使用的初始版本的yolov5和yolov4训练的吧,是有什么op的改进吗?我现在使用您的数据yolov5训练出来的结果中map=0.91 ,但是发现是看形状检测火焰的,误识别的情况还是挺高的,(其中数据是使用的您的模型预标,然后我修改,测试时候conf=0.7)

@gengyanlei
Copy link
Owner Author

@yutao007 我用的yolov5 v3版本。yolov4无论怎么样 都可以读取测试的

@yutao007
Copy link

yutao007 commented Feb 7, 2021

谢谢。我这边借用您有标签2000多张数据,拿您的模型预标无标签的百度数据,最后总共挑选了7000张数据,6400训练,600测试。我使用的是yolov5s 2版本的416*416,我发现我的火焰检测误检测很严重,用手比个三角形都会误检查。会不会是我的训练尺寸小了,或者是我的标签数据少了?

@gengyanlei
Copy link
Owner Author

@yutao007
不是很清楚啊,特别小的、特别大的都不是很好预测

@yutao007
Copy link

yutao007 commented Feb 8, 2021

恩拉,谢谢。我在看看怎么弄一下

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

2 participants