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

训练过程中准确率一直没有变化 #12

Open
cc-cv opened this issue Dec 24, 2018 · 5 comments
Open

训练过程中准确率一直没有变化 #12

cc-cv opened this issue Dec 24, 2018 · 5 comments

Comments

@cc-cv
Copy link

cc-cv commented Dec 24, 2018

在每个ct图像上我取了100个反例,跑了1000个epoch,准确率一直是0.593750,loss也一直基本没有什么变化。请问是什么问题呢?您试验的效果怎么样

@shartoo
Copy link
Owner

shartoo commented Dec 24, 2018

嗯,我也遇到这个问题,目前在排查中。。

@shartoo
Copy link
Owner

shartoo commented Jan 2, 2019

我怀疑这个论文是不是有问题,我尝试了很多修改loss都没法下降。

@jieshaoxxiansen
Copy link

我怀疑这个论文是不是有问题,我尝试了很多修改loss都没法下降。

我有一个疑问,在你提取数据的时候,你将annotations.csv提取为正标签,将candidates_V2.csv提取为负标签,您是否注意到candidates_V2.csv文件中已经包含了annotations.csv的全部标签

@liuchuangjun321
Copy link

loss不收敛代码在交叉熵使用的时候有个错误:
cross_entropy=tf.reduce_sum(tf.nn.softmax_cross_entropy_with_logits(logits=net_out,label = real_label))
改为cross_entropy=tf.nn.softmax_cross_entropy_with_logits(logits=net_out,label = real_label)

@shartoo
Copy link
Owner

shartoo commented Apr 17, 2019

@jieshaoxxiansen 你说的我倒真没注意到,我以为candidates_v2.csv全部是负样本标签呢。感谢提醒。
@liuchuangjun321 你说的这个我之前也改过好像也没法纠正问题。

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