Updating Anaconda and Tensorflow breaks the execute Python extension.

hughesfleming68hughesfleming68 Member Posts: 323 Unicorn
edited March 2019 in Help
Just updated my Tensorflow env which was working fine in Rapidminer before I did my conda update tensorflow 20 minutes ago. I am surprised that updating Anaconda would do this.

I am running in to problems everyday with 9.2.


Best Answer

Answers

  • hughesfleming68hughesfleming68 Member Posts: 323 Unicorn
    edited February 2019
    I solved the problem by downgrading Python to 3.6.5. The question is.....what is happening? Obviously, not upgrading Anaconda can only last so long as a workaround. Is there anything on the Anaconda side that needs to be done?
  • hughesfleming68hughesfleming68 Member Posts: 323 Unicorn
    Thanks Rodrigo. This looks like an Anaconda problem. I have an older RM version installed as well with older versions of the extensions and it is doing the same thing so this has nothing to do with Rapidminer. Worse, this is Sunday, I should be doing something else!
  • hughesfleming68hughesfleming68 Member Posts: 323 Unicorn
    I have it working now with 3.7.2 so setting up a new environment works while updating an older one breaks. Thanks again Rodrigo.
  • rfuentealbarfuentealba Moderator, RapidMiner Certified Analyst, Member, University Professor Posts: 568 Unicorn
    Hi @hughesfleming68,

    The idea behind anaconda is to have different environments per project or kind of project at least. I do have this separation:
    • base
    • rapidminer
    • servers
    • selenium
    • fileparsing
    • datamining
    • webapps
    • daemons
    • project***
    That way, I can try different things with certain stable libraries, and if I have to fix something because an update wreaked havoc, it's just a project or a small piece of code, not the entire work of a lifetime. (ok, 9 months, 5 projects a week, it's not that big of a tragedy).

    All the best,

    Rodrigo.
  • hughesfleming68hughesfleming68 Member Posts: 323 Unicorn
    Thank you Rodrigo. I understand the benefits of different environments. I already had Rapidminer working in a custom tensorflow env. It is unfortunate that the solution is to create a new env and basically set everything up again because a tensorflow update broke the connection with Rapidminer.

    regards,

    Alex
Sign In or Register to comment.