-
Notifications
You must be signed in to change notification settings - Fork 394
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Send live updates outside a git repo #5108
Conversation
@@ -125,6 +121,50 @@ The end result of running the pipeline looks like this: | |||
|
|||
 | |||
|
|||
### Live experiment updates in SageMaker jobs |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It would be great to show / mention that is the simplest but a limited approach. Another alternative is to bundle git repo to Sagemaker, bundle results at the end of the job. I know users who are using that approach and it's not terribly complicated. We even have script examples. So we can potentially do a blog or cleanup and share them.
I'm not sure about Databricks though ...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It would be great if we can see an example and we can add it. I hesitate to mention other approaches now if we don't have clear instructions on how to implement them yet.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
good stuff. I think there are some alternatives at least in case of Sagemaker - people just bundle the repo, also it's not exactly clear what is the e2e workflow. Those experiments will be detached forever in Studio? do we expect them to just stay there? How about data - does this workflow assume no DVC-tracked data / models?
Yup, that's pretty much it. We can add more in the future, but the simplicity is the point here. It shows that you can get a very similar experience to other trackers with about the same amount of effort. Showing additional value by copying files up and down from sagemaker can be a next step. |
21c9259
to
f9b8679
Compare
Docs for iterative/dvclive#646. That PR is for DVCLive, but it's really about logging to Studio from platforms where code is not run inside a Git repo.
The docs PR covers: