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
Sdk’S Dataset.Get And Dataset.Create Are Appending

SDK’s Dataset.get and Dataset.create are appending .datasets/{dataset-name} to the actual project name and failing because my datasets are from an older version of clearml.
How do I prevent it from appending it?

  
  
Posted 6 months ago
Votes Newest

Answers 4


Hi John,

I am able to get the dataset using the dataset-id but not using dataset-name and project

  • I upgraded ClearML from 1.0 to 1.5. I’ve followed the guide to setFeatureCompatibilityVersion during the upgrade
  • The difference between 1.0 and 1.5 seems to be that datasets, when they’re created go into a special project-1/.datasets/{dataset-name} project. In 1.5, we were just storing it directly in the project-1 project.
  • Now, because of this when I try to read a dataset using the SDK and provide project name as project-1 and dataset name as xyz it’s trying to get that from project-1/.datasets/xyz instead of project-1 where the dataset actually is.
  
  
Posted 5 months ago

I’ve resorted to calling list_datasets and then getting the id and calling get by id

  
  
Posted 5 months ago

Hi @<1691258549901987840:profile|PoisedDove36> , not sure I understand. Can you please elaborate with screenshots maybe?

  
  
Posted 6 months ago

I see. In this case you might need to re-register your datasets

  
  
Posted 5 months ago
415 Views
4 Answers
6 months ago
5 months ago
Tags