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
Hello, I Was Wondering If Anyone Could Help With An Issue Relating To Cloned Tasks. I Have Been Able To Register And Run A Regular Task Which I Then Try To Clone Using The Examples Provided. This Runs As Expected And Gets Registered With Clearml-Server.

Hello, I was wondering if anyone could help with an issue relating to cloned Tasks. I have been able to register and run a regular Task which I then try to clone using the examples provided. This runs as expected and gets registered with Clearml-server.

` Task.force_requirements_env_freeze()
task = Task.init(project_name='hyper param tuning', task_name='HyperParamSearchBase')

config_dict = {'psid': '36c37412-2716-4a67-b5a3-92084cf0b3a2',
'overlay': 'DSR-570_cml'}

config_dict = task.connect(config_dict)

Do processing here

single_psid_hyperparam_search(config_dict) `
However when I try to run a clone of the registered task, I get the following error. No other output is produced except the contents of the clearml.cfg

ImportError: cannot import name 'read_binary'
clearml_agent: ERROR: Command '['python3.6', '-m', 'virtualenv', '/home/annona/.clearml/venvs-builds/3.6']' returned non-zero exit status 1.

This occurs both when I am cloning and running via the UI and also within code. The method used to clone is similar to the tutorials.

` template_task = Task.get_task(project_name='hyper param tuning', task_name="HyperParamSearchBase")

for psid in plist:

cloned_task = Task.clone(source_task=template_task,
                         name=f'{template_task.name} {psid}', parent=template_task.id)

config_dict = {'psid': psid,
               'overlay': 'DSR-570_cml'}

cloned_task.set_parameters(config_dict)
cloned_task.upload_artifact('config_dict', config_dict)

Task.enqueue(cloned_task.id, queue_name='default') `

Any thoughts, greatly appreciated. Thanks for an awesome tool.

  
  
Posted 2 years ago
Votes Newest

Answers 4


Hi FlatStarfish45 , can you attach the clearml agent logs?

  
  
Posted 2 years ago

in your configuration, you have agent.package_manager.type = conda:wq . can you try running it with conda ?

after changing your ~/clearml.conf file, you need to restart the agent .

  
  
Posted 2 years ago

Thanks TimelyPenguin76 this is the output from execution

  
  
Posted 2 years ago

Hey that worked. It is reproducing the same results as the base task now. Interesting, in that I didn't specify the :wq. Perhaps it got added when the agent initialized. Thanks for your help TimelyPenguin76

  
  
Posted 2 years ago
529 Views
4 Answers
2 years ago
one year ago
Tags