The Adam-SGD gap in large-batch LLM pre-training arises mainly from SGD's restricted effective learning rates caused by small gradients and output-layer spikes; clipping lets SGD recover nearly all of Adam's performance.
It is evident that the usual learning rate choice of SGD in the large batch setting will leave a significant 14 0 1000 2000 3000 4000 5000 Iterations 101 Avg
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Revisiting the Adam-SGD Gap in LLM Pre-Training: The Role of Large Effective Learning Rates
The Adam-SGD gap in large-batch LLM pre-training arises mainly from SGD's restricted effective learning rates caused by small gradients and output-layer spikes; clipping lets SGD recover nearly all of Adam's performance.