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 All, We Are About To Setup The Ml Infra On One Of The Ongoing Projects Where We Used Clearml Server For Cv Models Training. The Parts Of The Infra Should Be:

Hi all,
We are about to setup the ML infra on one of the ongoing projects where we used ClearML server for CV models training. The parts of the infra should be:

  • Training- Execution of a set of complex pre-processing pipelines that extracts data from the PDF files and prepares it for the model training. Where each new data batch should be added to previously split train/test/val datasets only once.
  • Run model training (we are about to use AWS GPU instances for that).
  • Run the validation module that inference a newly trained model on some cases verifying it provides accurate predictions + plus metrics check.
  • If the validation passed successfully deploys the model into production.- Inference- The inference steps would also be pretty the same as for training with complex pre-processing of large PDFs, so it's not merely a model inference (we used lambdas + sagemaker on other projects for that)
    Now, I'm wondering, if it's possible to do this all nicely via ClearML?
    Previously we were doing all of that via GitLab + AWS.
    The main concern in switching to ClearML is probably sophisticated preprocessing step for the inference. I just don't know how much the model serving is flexible in ClearML.

Would appreciate any guidance and materials you might have for reading. I already passed through the case studies and found this one particularly useful but did get a grasp on why there is a need for two orchestrators at a time (GitLab + ClearML).

Thank you!

  
  
Posted 2 months ago
Votes Newest

Answers


Hi @<1666253626772819968:profile|SmoggyDog77> , this is indeed a very interesting and somewhat complicated use-case! I think ClearML can support all of your current needs and any that might rise in the future.

I would suggest to contact ClearML directly to get a better idea of how ClearML can assist your MLOps efforts.

The suggested method from my experience would be via the contact form - None

  
  
Posted 2 months ago
140 Views
1 Answer
2 months ago
2 months ago
Tags
Similar posts