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

extract_feat函数加载问题 #54

Open
developWmark opened this issue Jul 25, 2023 · 7 comments
Open

extract_feat函数加载问题 #54

developWmark opened this issue Jul 25, 2023 · 7 comments

Comments

@developWmark
Copy link

https://github.com/cuiziteng/Illumination-Adaptive-Transformer/blob/main/IAT_high/IAT_mmdetection/mmdet/models/detectors/IAT_detector/IAT_yolo.py
extract_feat函数和forward_train只初始化在init函数里面,实际上是不能够被反向传播的,我在extract_feat加了一行打印函数,训练过程中并没有打印。将extract_feat函数和forward_train移出init函数,作为IAT_yolo的成员函数,实际上是重写SingleStageDetecto的函数,这样extract_feat才会参与反向传播,打印日志。或者是我理解错了,请作者耐心解答一下。

@WWJ0720
Copy link

WWJ0720 commented Jul 25, 2023

我也遇到了这个问题,并且也是相同的解决方法。之前初始化在init里面无法反向传播

@developWmark
Copy link
Author

我也遇到了这个问题,并且也是相同的解决方法。之前初始化在init里面无法反向传播

但是这样训练后保存的中间增强的图像,可视化实际上和作者论文呈现的差别很大

@cuiziteng
Copy link
Owner

十分感谢耐心纠正,因为代码失误对您的结论造成困扰,这部分的实验我后面重新跑一遍,并且更新一下结果参数。

@developWmark
Copy link
Author

好的,这个问题也引发了这个问题打断点在init才有效
#52

@mrwrui
Copy link

mrwrui commented Aug 28, 2023

可是为什么无法反向传播,依然能够复现出较高的mAP呢?

@liuwei0066
Copy link

可是为什么无法反向传播,依然能够复现出较高的mAP呢?

YOLOV3的基础性能就挺好的,可能由于随机性,作者跑出来的性能比较好。

@liuwei0066
Copy link

十分感谢耐心纠正,因为代码失误对您的结论造成困扰,这部分的实验我后面重新跑一遍,并且更新一下结果参数。

作者您好,请问您更新新的代码和实验结果了吗?

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

5 participants