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
Unanswered
Hi, Can I Ask How I Can Make Clearml-Datasets In Comparison With Pytorch Datasets/Dataloader? In Particular, Pytorch Dataloaders Would Be Able To Batch Pull And Then Preprocess Data Using Multi-Cpus, Feed It Into The Training Loop And Achieve As High Util


SubstantialElk6 , I think this is what you're looking for:
https://clear.ml/docs/latest/docs/references/sdk/dataset#get_local_copy
Dataset.get_local_copy(..., part=X)

  
  
Posted 2 years ago
195 Views
0 Answers
2 years ago
one year ago