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 Am Saving Some Intermediate

Hi! I am saving some intermediate .pt files on the experiments and clearml automatically detects them as models, this makes the clearml.model - INFO message appear quite often and make the logs unreadable. Is there a way to specify which files should be considered models?

  
  
Posted 3 years ago
Votes Newest

Answers 6


Hi CostlyOstrich36 ! The message is the following:
clearml.model - INFO - Selected model id: 27c1a1700b0b4e25a4344dc4ef9868faThey are not models, those are intermediate tensors I am caching to make training faster. I don't need to log them.

  
  
Posted 3 years ago

GrievingTurkey78 , Hi!

Can you please provide the message that is given to you? Also, are you saving those files as some sort of models you want logged in the system or you don't need them?

  
  
Posted 3 years ago

GrievingTurkey78 , it appears to be bad methodology on my side as disconnecting the framework won't log any scalars etc...

I think a better solution would be to set the log level to something else, however you will not see 'INFO' messages from the module.

Try something like this:
logging.getLogger('clearml.model').setLevel(logging.WARNING)
This will only show you the warnings regarding the models module.

  
  
Posted 3 years ago

GrievingTurkey78 , you can try using the following example:
task.init( auto_connect_frameworks={ '<FRAMEWORK>': False }, ... )
Docs: https://clear.ml/docs/latest/docs/references/sdk/task#taskinit
However please note this will not auto save the final model and you will need to add the model upload at specific points in the script when you want to upload the script.

Am I understanding correctly this is what you want?

  
  
Posted 3 years ago

So I would have to disconnect pytorch? And then upload the model at the end

  
  
Posted 3 years ago

Thanks! This should work perfectly 👌

  
  
Posted 3 years ago