5

I was trying to apply a deep learning algorithm(CNN) in python but after separating training-testing data and transforming time series to image step my Colab Notebook crashed and restarted itself again.

It gives an error like "Your session crashed after using all RAM" and when I checked app.log I saw something about tcmalloc: large alloc. I didn't find anything to fix this crashed.

Do you have any idea how to prevent this warning and fixed this situation?

amo-ej1
  • 3,279
  • 26
  • 35
BegumAy
  • 53
  • 1
  • 4
  • You could go to a larger machine with more memory available, but you could also think about what you are doing during your preprocessing and training phases. Are you running out of memory because your model is too complex or are you running out of memory because during your preprocessing you keep objects you no longer need in memory (which you can delete). Also consuming your training/test data in chunks could be worthwhile. – amo-ej1 Mar 25 '21 at 08:18
  • I tried to fit my model with higher ram and it works. Also i used garbage collector for release memory, that's work for me. Thank you for your reply. – BegumAy Mar 27 '21 at 15:14

1 Answers1

3

Your session ran out of all available RAM. You can purchase Colab Pro to get extra RAM or you can use a Higher RAM machine and use the Neural Network there

neel g
  • 1,138
  • 1
  • 11
  • 25