0

I have a signal from which I'm interested in extracting different features of it. For each feature, I've trained a different neural network model and saved as HDF5 files. Now I'm loading them in order to make predictions, but I'm asking myself if I cannot run them in parallel, displaying accordingly all the features at the same time, since the input is the same?

If so, how can I do it.

Thanks

Marlon Teixeira
  • 334
  • 1
  • 14

1 Answers1

1

You can simply run the models simultaneously, using multithreading, multiprocessing or by running multiple Python scripts (one for each model).

Another option is to use keras and connect all models by feeding the same input layer into all models. This will result in a single supermodel and you can let keras worry about parallelization. This setup even allows you to train all model together and use shared layers.

See the keras documentation for details.

Feodoran
  • 1,752
  • 1
  • 14
  • 31
  • By the way, those links can add up to your answer: https://stackoverflow.com/questions/7207309/python-how-can-i-run-python-functions-in-parallel https://stackoverflow.com/questions/45643375/python-how-do-we-parallelize-a-python-program-to-take-advantage-of-a-gpu-server – Marlon Teixeira Mar 24 '20 at 13:49