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
What’S The Point Of Tracking Artifacts Dynamically?

What’s the point of tracking artifacts dynamically?
https://allegro.ai/clearml/docs/docs/examples/reporting/artifacts.html#dynamically-tracked-artifacts
In the end, it stores only the latest version and doesn’t include the history of changes. Thus, if I understand correctly, it is identical to uploading the latest version of artifact (e.g. just before task.close() ) without tracking. I think it would be cool to have all the versions tracked - I would then track how columns are added and removed while I’m doing data cleaning and feature engineering

  
  
Posted 2 years ago
Votes Newest

Answers 7


Hi FiercePenguin76
Artifacts are as you mentioned, you can create as many as you like but at the end , there is no "versioning" on top , it can be easily used this way with name+counter.
Contrary to that, Models do offer to create multiple entries with the same name and version is implied by order. Wdyt?

  
  
Posted 2 years ago

not sure - ideally I would like to see these tables (e.g. with series_name, series_dtype, number_of_non_na_values as columns) back to back in the GUI to track the transformations. I think it isn’t possible with Dataset . Anyway, this whole scenario is not a must have, but a nice to have.

  
  
Posted 2 years ago

Registering some metadata as a model doesn’t feel correct to me.

Yes I'm with you 🙂
BTW what kind of meta-data would need versions during the life time of a Task ?

  
  
Posted 2 years ago

Thanks for the answer! Registering some metadata as a model doesn’t feel correct to me. But anyway this is certainly not a show-stopper. Just wanted to clarify.

  
  
Posted 2 years ago

mostly the transformation of the pandas Dataframe - how the columns are added/removed/change types, NAs removed, rows removed etc

  
  
Posted 2 years ago

Make sense. BTW: you can manually add data visualization to a Dataset with dataset.get_logger().report_table(...)

  
  
Posted 2 years ago

Interesting!
Wouldn't Dataset (class) be a good solution ?

  
  
Posted 2 years ago
572 Views
7 Answers
2 years ago
one year ago
Tags