Per-component relative learning-rate schedules speed up Transformer pretraining by up to 23% for MoE models, and the schedules tuned on a 34M model transfer to models 27x larger.
Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity
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
-
Decoupled Relative Learning Rate Schedules
Per-component relative learning-rate schedules speed up Transformer pretraining by up to 23% for MoE models, and the schedules tuned on a 34M model transfer to models 27x larger.