Hi! I'M A Devops Engineer. My Company Is Self-Hosting Clearml On Kubernetes. I'M A Clearml Newbie, So Pardon My Ignorance. I'M A Little Confused By What Clearml Artifacts (See Screenshot Below) And Custom Models Are. Are They One And The Same? Where Are

Answered

Hi! I'm a DevOps engineer. My company is self-hosting ClearML on Kubernetes. I'm a ClearML newbie, so pardon my ignorance.

I'm a little confused by what ClearML artifacts (see screenshot below) and custom models are. Are they one and the same? Where are they both stored? On the file server?

This doc makes me think that custom models are stored on the file server (same as artifacts), which makes me think they are one and the same.

These docs talk about how to configure a google storage bucket, but it's not clear what will be stored there:

None
None

Regardless, how do I configure a self-hosted ClearML to store custom models on GCS? Would really appreciate the help, thanks!

  				
Posted 
	4 months ago

					More
				  		
  Report
		
					BeefyDolphin18
				
					0
					 × 1

Votes Newest

Answers

Artifact can be anything, that you can use clearml SDK to upload to storage. Which storage is used is defined by your clearml.conf (with its credentials) ClearML web and api server do not store those files

Model is a special artifact: None
Example you have the lineage feature where if you train model B using model A as starting point (aka pre-trained) , and model C from model B, ... The lineage will track modelC was built on top of model B, and A.

What is considered as artifact, what as model ? Clearml have some premade code that detect that based on the framework that you use: via the autologging system: None

  				
Posted 
	4 months ago

					More
				  		
  Report
		
					ManiacalLizard2
				
					0
					 × 1

Write your answer

418 Views

1 Answer

4 months ago