Using NN with parallel give me bad result

I have three classes to classify as my test case. I will have more classes and data to use so I want to use parallel processing. However, I found that even using the same data sets and classes using parallel give me a ~33% which is like random guessing while not using it give me ~100%. Are the parameters need to change for parallel NN?
Not using parallel
Using parallel

Antworten (0)

Diese Frage ist geschlossen.

Gefragt:

am 26 Mai 2016

Geschlossen:

am 20 Aug. 2021

Community Treasure Hunt

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

Start Hunting!

Translated by