Implement dynamic blending#238
Open
voegtlel wants to merge 1 commit into
Open
Conversation
…ize function to determine
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The task encoder can now define a
blend_sample_sizefunction to determine the amount each sample contributes.Balancing is deficit-based on historic distribution. We do not look-ahead for now, also we do not keep running averages of sizes (to simulate look-ahead).
There are two configs for this:
blend_sample_size_alpha: this is the coefficient (i.e. strength of deficit). For =0, it would fall back to ignoring the size. For <1, softer correction, >1: more aggressive correction.blend_sample_size_epsilon: This is added to all deficits (so it depends on the unit/size ofblend_sample_size). I.e. if 0, only consider undersampled sources in sampling. If >0, consider oversampled sources as well. Higher epsilon will down-weight deficits.