0

I would like to use Kerastuner distributed tunning and have followed the details outlined in https://keras-team.github.io/keras-tuner/tutorials/distributed-tuning/. However, all my worker nodes finish gracefully while my chief process remains running even when the search space is fully explored. Ideally, I would like to be able to make certain summary reports once all the searches are done by the workers. But, it seems that the chief process remains running even if the search space is fully explored. Is there a way to get around this issue?

Pooria
  • 780
  • 1
  • 8
  • 17

1 Answers1

0

There was a recent pull request that stops the server after the trials finished. You should update your keras-tuner version.

For example: pip install git+https://github.com/keras-team/keras-tuner@master

Tomer23
  • 31
  • 4