Machine learning package seems to hog all the memory

  • samlinck

    (@samlinck)


    2 months, 2 weeks ago

    I get this error when visiting the site
    ‘PHP message: PHP Fatal error: Allowed memory size of 268435456 bytes exhausted (tried to allocate 20480 bytes) in plugins/no-unsafe-inline/vendor/rubix/ml/src/Datasets/Dataset.php on line 71’

    This table is also very full: wp_nunil_inline_scripts (732.8 MiB)

    Settings:
    Enable tag capture = false
    Enable csp = true

    I see that there is a clean database action, but won’t it delete my clusters that i need for the inline scripts?

    • This topic was modified 2 months, 2 weeks ago by samlinck.
Viewing 7 replies - 1 through 7 (of 7 total)
  • Plugin Author Giuseppe

    (@mociofiletto)

    2 months, 2 weeks ago

    It seems you have collected a lot of data.

    Disable tag capture, run the clusterer (“Trigger Clustering” button), then use the “Prune scripts from the database.” button. It will prune big cluster.

    Don’t use clean database, because it is used to completely erase the data.

    Thread Starter samlinck

    (@samlinck)

    2 months, 2 weeks ago

    Hey Guiseppe,

    I already went trough those steps.

    It’s still that big.

    Plugin Author Giuseppe

    (@mociofiletto)

    2 months, 2 weeks ago

    I will check this issue. It seems that the training of the model requires a lot of memory.

    Can you show the main report of database (the first table, at the bottom of the page in the tools tab)?

    Thread Starter samlinck

    (@samlinck)

    2 months, 2 weeks ago

    View post on imgur.com

    Here you go @mociofiletto .

    Plugin Author Giuseppe

    (@mociofiletto)

    2 months, 2 weeks ago

    HI @samlinck, I am working on a possible solution. It will require some time, but I think that it is needed.

    If you can, store your data, because I will ask you to test some code in the next week.

    Thread Starter samlinck

    (@samlinck)

    2 months, 2 weeks ago

    Hi @mociofiletto, with some delay the pruning did his job. The table size is now only 7.5mb anymore.

    Plugin Author Giuseppe

    (@mociofiletto)

    2 months, 1 week ago

    Hi @samlinck can you test the development version?

    I have added support for model persistence and a step by step training. This should reduce memory used by the plugin.

Viewing 7 replies - 1 through 7 (of 7 total)
  • You must be logged in to reply to this topic.

Tags