Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
Hi, I’M Trying To Bypass Task’S “Installed Packages” And Require An Installation Of My Repo’S

Hi,
I’m trying to bypass Task’s “installed packages” and require an installation of my repo’s requirements.txt .

I’ve set agent.package_manager.force_repo_requirements_txt: true in ~/clearml.config and restarted the server (and agent) but it seems as though the agent is still taking installed packages into account over the requirements.txt .

Is there something I’m missing? Why does ClearML require the repo’s requirements to be installed locally in the first place?

Thanks!

  
  
Posted 2 years ago
Votes Newest

Answers 7


StickyCoyote36 , Hi!

Is the agent ignoring the requirements.txt and only installing the "installed packages" in the task or installing them both?

  
  
Posted 2 years ago

It’s installing them both, I just got it to ignore installed packages by adding Task.force_requirements_env_freeze(requirements_file='requirements.txt') but I’d like a more general solution (i.e. not have to add that line before every Task.init(..) )

  
  
Posted 2 years ago

StickyCoyote36 , I think that is the solution. Is there a reason you want to ignore the "installed packages"? After all those are the packages that the task was ran with.

  
  
Posted 2 years ago

StickyCoyote36 , if I understand correctly due to your M1 chip limitation you run the script from a different machine and then you use the agent to run on the M1 chip machine and you want the requirements.txt in the repo to override the "installed packages" when running with agent, correct?

  
  
Posted 2 years ago

It’s because I have an M1 chip and can’t install the latest tensorflow version locally

  
  
Posted 2 years ago

StickyCoyote36 , I'm looking into a solution.
Please hold on 🙂

  
  
Posted 2 years ago

Something like that yes

  
  
Posted 2 years ago