Non-record: MTP Auxiliary Loss Exploration#955
Open
genji0306 wants to merge 2 commits intoopenai:mainfrom
Open
Non-record: MTP Auxiliary Loss Exploration#955genji0306 wants to merge 2 commits intoopenai:mainfrom
genji0306 wants to merge 2 commits intoopenai:mainfrom
Conversation
A systematic orchestration framework that coordinates solution analysis, hypothesis generation, experiment design, and automated testing for Parameter Golf. Includes control plane CLI, JSON schemas, metric helpers, and experiment materializer. Full source: https://github.com/genji0306/parameter-golf-orchestrator
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Exploration of the existing Multi-Token Prediction (MTP) code path in
train_gpt.py, which is fully implemented but disabled (MTP_NUM_HEADS=0) across all current submissions. MTP heads are excluded from the exported artifact, so enabling them as an auxiliary loss has zero impact on the 16MB constraint.Status
Pending H100 validation. Will update with results after compute access.