WARNING batched routines are designed for small sizes. It might be better to use the Native/Hybrid classical routines if you want good performance.

[Ericchen328]

When I train this network with a different dataset, the following error occurs：

WARNING batched routines are designed for small sizes. It might be better to use the
Native/Hybrid classical routines if you want good performance.

===============================================================================
do you know how to deal with it

[cnexah]

I didn't meet this problem before.
I guess the reason is that the resolution of input image is too large.
Could you please try to down-sample the input image before feeding into the network?

[Ericchen328]

ok,i will try it thanks you so much