Troubleshooting

KNIME Extensions
1

ERROR DL4J Feedforward Learner (Classification) 0:267 Execute failed: ConvolutionLayer (index=0, name=layer0) nIn=0, nOut=96; nIn and nOut must be > 0

So when i train my deep learning algorithm using the Alexnet DL4J Node, i get the above error.
I have a dataset of 1319 images
My target column in binary - 0 or 1 converted to string
The image size is 150, 150, 0 all have been reset to the same size.

Im not a coder, a physician infact- so Im just trying things out.

Cheers and regards
Adarsh

2

Hi AdarshGhost,

the DL4J Feedforward Learner has an option to set the image size. Maybe, you forgot to it set there? If this doesn’t solve the problem you could upload an example workflow with e.g. one of your images (or a dummy image if that’s not possible) and I’ll have a look.

Cheers,
David

3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.