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

slice层转换问题 #85

Closed
eeric opened this issue Aug 5, 2017 · 4 comments
Closed

slice层转换问题 #85

eeric opened this issue Aug 5, 2017 · 4 comments

Comments

@eeric
Copy link

eeric commented Aug 5, 2017

@nihui @wangkgege ,lightcnn C模型转换成功,但在etlwise_fc1提取的256维特征与caffe版的平均L1规则误差在实数1以内。
是不是按ced04dd修改,就可以纠正?

另外,C模型转后的param的slice层参数出现负数,这个是不是有问题?如下所示:
Slice slice1 1 2 conv1 slice1_1 slice1_2 2 -233 -233
Eltwise etlwise1 2 1 slice1_1 slice1_2 eltwise1 2 0
Pooling pool1 1 1 eltwise1 pool1 0 2 2 0 0
Convolution conv2a 1 1 pool1 conv2a 96 1 1 1 0 1 4608
Slice slice2a 1 2 conv2a slice2a_1 slice2a_2 2 -233 -233
Eltwise etlwise2a 2 1 slice2a_1 slice2a_2 eltwise2a 2 0
...............

再问问,param里面的参数具体是什么含义?先谢谢啦!

@nihui
Copy link
Member

nihui commented Aug 5, 2017

-233 是个特殊标记,表示运行时断定,Slice 里的 -233 就是均匀切分

@nihui
Copy link
Member

nihui commented Aug 5, 2017

注意下输入图像通道 RGB / BGR 的区别

@eeric
Copy link
Author

eeric commented Aug 7, 2017

@nihui,感谢

@eeric eeric closed this as completed Aug 7, 2017
@xiaomage11
Copy link

@eeric,您好,你的问题已经解决了吗,输出的特征跟caffe的一致了吗?

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

3 participants