SinkGD is a stateless optimizer that balances gradients via square-root Sinkhorn iterations, matching or beating Adam and memory-efficient baselines on LLaMA pretraining with SGD-level memory.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Gradient Multi-Normalization for Stateless and Scalable LLM Training
SinkGD is a stateless optimizer that balances gradients via square-root Sinkhorn iterations, matching or beating Adam and memory-efficient baselines on LLaMA pretraining with SGD-level memory.