What possibilities exist to export the developed net and in which formats is this possible? (C #, C ++, TensorFlow, IC 61131 ...) o C # -taping and implementation is the focus of our company!
• Which method is the most meaningful for the variation of experiment parameters and how would you implement them? o Layer variation, Layer order, HiddenUnits, OutputMode ....
• What is the option already trained networks with new data to Trainierne?

답변 (2개)

Bernhard Suhm
Bernhard Suhm 2018년 8월 4일

1 개 추천

If by "variation of experiment parameters" you mean tuning the hyperparameters, you best perform that within the neural network toolbox, as described on this page.
You can export trained networks into Caffe, Keras, or the ONNX interchange format.
Christian Möllers
Christian Möllers 2018년 8월 8일

0 개 추천

Thank you for your answer Bernhard, but my problem is that I work with times series data. So I cannot use the validation functions in the training... Moreover, I look for an optimization tool or technic for time series training

카테고리

도움말 센터File Exchange에서 Deep Learning Toolbox에 대해 자세히 알아보기

질문:

2018년 7월 19일

답변:

2018년 8월 8일

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by