Good info! If you could write a blog post or article about it that would be great. How did you trained Neuroph in multiple threads (there were several experiments with that, which approach did you used?). I have some new idea that could really make even bigger difference.
Thanks for the update, we'll make it more clear.
No support for cuda, opencl at the moment. There is a limited support for convolutional networks.
We tried something with arrays beffore, but it was not big difference in speed. If you have idea how to do it go ahead fork it from https://github.com/neuroph/neuroph Multithreading should be improved, something has been done in https://github.com/neuroph/neuroph/tree/multithreaded but I think it needs fixing You can measure performance using JMH
If you have target outputs use supervised, otherwise use unsupervised Use backprop with momentum Yes there is option, right click data set then 'create training and test subset'
Hi John, sorry I missed this. The error reporting for import should be improved, I know about that and you're welcome to help. Let me know if you need help with anything else. The support forum will be moved to github, there are requests at both places at the moment. Zoran
Use the -Xmx switch to give it more memory, and use profiler to figure out why it gets stuck
We'll fix that, thanks for the note. Not sure whats the problem at the moment