object classes classification layer must be equal in the input trainingData plus 1 for the "Background" class
1 次查看(过去 30 天)
显示 更早的评论
Hi All Matlab folk,
I am trying to manually code the classes for my network can someone assist me please?
I am trying to gain a better understanding of this error! Can some one help me understand this error in an easier way please?
I changed the input layer details in the DAG network and still this it is arguing!
My Input layer:
imageInputLayer([32 32 3],"Mean",[],"Normalization","zerocenter", "Name","imageinput")
My ERROR:
The number object classes in the network classification layer must be equal to the number of classes defined in the input
trainingData plus 1 for the "Background" class
0 个评论
回答(1 个)
Raunak Gupta
2020-2-21
Hi,
As per the example mentioned in trainRCNNObjectDetector the number of classes to be mentioned for training must be (objectClasses + 1) .The objectClasses should also be mentioned in a cell array which can represent the name of those classes. The fullyConnectedLayer will be having outputSize as (objectClasses + 1).You may look into the above-mentioned example for clarity about implementing the same.
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Image Data Workflows 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!