Data splits to default to MLContext seed when not specified #4764
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
harishsk
approved these changes
Feb 2, 2020
justinormont
reviewed
Feb 2, 2020
| if (samplingKeyColumn == null) | ||
| { | ||
| samplingKeyColumn = data.Schema.GetTempColumnName("SamplingKeyColumn"); | ||
| data = new GenerateNumberTransform(env, data, samplingKeyColumn, (uint?)seed); | ||
| data = new GenerateNumberTransform(env, data, samplingKeyColumn, (uint?)(seed ?? host.Rand.Next())); |
There was a problem hiding this comment.
Why fallback to host.Rand.Next() instead of the fixed seed given to the MLContext?
I think this is doing (1), then (4) from: #4752 (comment)
TrainTestSplit etc should listen to MLContext seed if not given by the user for the specific component.
I think the precedence for PRNG seeds should be:
- User seed specified for the component instance (or from parent component if hiding behind another)
- User seed specified in
MLContext- Default seed in component (unsure if we should remove these, might be a component-by-component discussion)
- Random (and more random than current -- see Use a GUID when creating the temp path #4645 (comment))
There was a problem hiding this comment.
MLContext seed isn't exposed, i.e. I cannot directly use mlContext.Seed. The seed for Host.Rand is set from MLContext though.
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #4752