joaopauloschuler

On your hardware (with CPU only), I would expect from 6 to 8 epochs per minute. With GPU, it will be even faster. There is a bottleneck somewhere.

Just thought about something: you can try to increase the batch size to 512 (you'll get some warning messages but you can disconsider): ``` NeuralFit.Fit(NN, ImgTrainingVolumes, ImgValidationVolumes, ImgTestVolumes, {NumClasses=}10, {batchsize=}512,...

It's better! I still think that you should try to change the batch size to 512.

With a bigger batch size, the step time will be bigger. In the plus side, the epoch time is likely to be smaller as there is less threading coordination. About...

I have never tried MNIST with a GPU. Will give a go at my end and let you know.

Thank you for the feed back! This request made me realize that nowhere in the documentation we have explanation about how to build multi-path architectures. I'll add documentation explaining how...

Finished coding.

TNNetGroupedConvolutionReLU failed testing.

@magicse, thank you for the report. I can reproduce it. I'll fix in the next version.

The main idea is good. It's against my religion to have global variables. So, I'll implement your solution passing parameters.