Open
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Batch Optimization + MLP4 + RoPE100k
Compared with the baseline 6L-384d run, this version applies a focused set of training and model updates:
TRAIN_BATCH_TOKENSwas reduced from 196,608 to 98,304,MLP_MULTwas increased from 2 to 4, bothMATRIX_LRandSCALAR_LRwere lowered from 0.04 to 0.035,WARMDOWN_ITERSwas shortened from 800 to 600, andROPE_BASEwas raised from 10,000 to 100,000.In practice, these changes improve optimization efficiency and model capacity while keeping the run inside the 10-minute / 16MB track limits on a single GPU. The best result from this configuration reached 1.4784 val_bpb on a small GPU (20 GB VRAM) in 10 min.