Record: Curriculum Learning + LeakyReLU(0.9)² + 7-gram Backoff (val_bpb=0.9633)#764
Open
ndokutovich wants to merge 1 commit intoopenai:mainfrom
Open
Record: Curriculum Learning + LeakyReLU(0.9)² + 7-gram Backoff (val_bpb=0.9633)#764ndokutovich wants to merge 1 commit intoopenai:mainfrom
ndokutovich wants to merge 1 commit intoopenai:mainfrom
Conversation
…bpb=0.9633, 1 seed)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
val_bpb = 0.9633 (seed 42, additional seeds pending compute grant) | 15.56 MB | 8xH100 SXM, 600s
Built on PR #753 (Podracing II) with two novel additions:
1. Curriculum Learning (Shard Reordering)
Training shards reordered by model perplexity — hardest shards first. Based on PR #650 (-0.003 BPB). Zero code change, environment variable only.
2. LeakyReLU(0.9)² Slope Optimization
Following @MatoTeziTanka's controlled sweep (issue #140): slope 0.9 gives -0.013 BPB vs standard 0.5. One parameter change.
Results
Artifact: 15,560,351 bytes (< 16MB)
Steps: 6,647 at 90.3ms/step
GPTQ calibration within training budget (issue #677 compliant)
Reproduction
Acknowledgments
@newjordan (PR #753), @abaybektursun (PR #650), @MatoTeziTanka (slope sweep), @Asukabot0 (n-gram backoff)
Status
1 seed submitted. 2 additional seeds pending OpenAI compute grant.
Previously PR #486 (formerly #2 on leaderboard, TrigramHash originator). $339 personal compute spent.
Test plan