what do you mean by draft? do you want to actually run your pipeline even for testing purposes?

Also, for some reason I don’t have the ability to copy pipelines. Tell me, is this normal?

Every time you click run a pipeline ran is cloned, so you can pass new parameters right from the UI

I also though clearML writes this mapping ( state.json ) into one of its databases: Mongo, Redis, Elasticsearch.

I can add a little piece of context.

• I want to give my users a way to pic a specific batch to get a file they need. Right now there is no way to download just one specific file from an entire dataset.
• I need a way to check whether a file has already been uploaded to some other dataset or not.

Thank you @<1523701070390366208:profile|CostlyOstrich36> 🤓

Hi @<1523701070390366208:profile|CostlyOstrich36> . Thank you for your advise, it definitely makes sense. Regarding to the first point, each dataset has a file state.json . In this file there os a key artifact_name e.g., data , data_001 , etc, and relative path of a file. I thought I can map this key with the chunk number. So, if I pull this file from s3 bucket, I can conclude which chunk I should download to get a specific file. Am I wrong?

It is deterministic. When you do Dataset.get(), clearML downloads file state.json, where you can see all relative file paths and chunks number

You can get a chunk number that contains your file and download that chunk

thank you 🤓