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

how to operate it #2

Open
Lee-AI-sco opened this issue May 2, 2019 · 8 comments
Open

how to operate it #2

Lee-AI-sco opened this issue May 2, 2019 · 8 comments

Comments

@Lee-AI-sco
Copy link

如果我只是想运行一下你的程序看一下效果,那我把你的代码和模型权重文件下载下来后,运行predict.py就可以了吗?

@wmylxmj
Copy link
Owner

wmylxmj commented May 3, 2019

将h5文件放入models文件夹,运行run.py,预测结果将输出在outputs文件夹

@Lee-AI-sco
Copy link
Author

第一个问题:你的模型是通过迁移学习得到的吗(因为你的代码是通过yolo实现的)?
第二个问题:weigts.h5是权重模型,model.h5是什么模型呢?(我学艺不精,如果可以的话,能否留下你的qq或者微信呢,感激不尽)

@wmylxmj
Copy link
Owner

wmylxmj commented May 4, 2019

我的权重仅仅是训练那320张图片得到的,训练集可能有点少。
model.h5就是模型文件(包涵模型和权重)在keras中通过model.save得到。

@fengduanqiao
Copy link

我想运用你的代码重新training,现在把dataset数据集改了,标签改过来,load_pretrained=False,infos里面的class也都改了,设置完成之后,我还需要修改哪些地方呢?还想问一下prepare.py里面的anchors = [(10, 13), ...这个anchors代表什么意思呢?多谢~

@wmylxmj
Copy link
Owner

wmylxmj commented May 16, 2019

改过数据集和标签后,修改prepare.py里的classes列表,换成你自己的类别。然后运行prepare.py,信息将被写入infos文件夹。
anchors是指先验框,由coco数据集中进行聚类得到,代表边界框最可能的9种大小。网络输出的边界框大小是由先验框比例缩放得来的。详细可以看YOLOv3的论文。

@fengduanqiao
Copy link

fengduanqiao commented May 16, 2019 via email

@zhonganzuzhihu
Copy link

换了别的数据集之后,训练的loss在16左右下不去,有没有什么好的办法来使loss下降的快速一点呢

@wmylxmj
Copy link
Owner

wmylxmj commented May 24, 2019

不同的数据集收敛的loss值不一样的,你可以看看当前的权重是否能出效果

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