4
$\begingroup$

I got an error: OSError: [Errno 12] Cannot allocate memory I deleted some files. And I have free memory. I don't want start learning again from first epoch. (I got an error in last epoch).

I use jupyter notebook. I have 19 .h5 file in my output file.
Can I click something to save my last .h5 file?

Print Screen

Edit: Media

Improve this question
$\endgroup$
12
  • $\begingroup$ Are you using tensorflow? $\endgroup$
    – Green Falcon
    Apr 3, 2019 at 15:58
  • $\begingroup$ Yes. I use tensorflow. $\endgroup$
    – Badum
    Apr 3, 2019 at 15:59
  • 1
    $\begingroup$ If you are using tensorflow and your kernel is still alive you can just resume the training because the variables hold what they've learnt until now. The current snapshot of your graph, you can be stored it. $\endgroup$
    – Green Falcon
    Apr 3, 2019 at 16:00
  • $\begingroup$ Yes. My kernel is alive. How can I resume the training? $\endgroup$
    – Badum
    Apr 3, 2019 at 16:01
  • $\begingroup$ just run the sess.run cell. Your current values are still available. $\endgroup$
    – Green Falcon
    Apr 3, 2019 at 16:01

1 Answer 1

Reset to default
1
$\begingroup$

One option is to add TensorFlow Checkpoint which allows for the saving and reloading of training.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.