nntool: after training output matrix size is less than training matrix size

1 次查看（过去 30 天）

LRBLFAST 2013-4-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/71858-nntool-after-training-output-matrix-size-is-less-than-training-matrix-size

Hi, when I am training any network using nntool, the generated output matrix size is not equal to the target matrix size, so plotconfusion is not working.

For e.g. if I trained an input 30x227) alongwith a target (3x227), I am getting the network1_outputs (3x137) which is less than the target (3x227), so plotconfusion giving errors.

I tried for several networks like (Cascade, Feed-Forward, Layer Recurrent etc.)

Thanks in advance.

3 个评论
显示 1更早的评论隐藏 1更早的评论

LRBLFAST 2013-4-16

Dear Mr. Greg Heath,

Kindly reply to this question. Why I am getting such kind of error? I understand that it may be a very basic problem. But actually I need to use nntool to revise some of our work in which we trained the data manually using nntool. When I am training after writing codes, our results are different than the old one. So thought should we go back to the old procedure and train....but getting this problem.

Please reply