Data Schedules and Learning Rate Schedules Should Be Co-Designed, Paper Argues

A new paper argues that data schedules and learning rate schedules should be designed together, showing that the optimal placement of math data during web-training flips depending on the learning rate schedule used.

2026-10-27 ~ 2026-10-27 · 2 related posts

• Data Curriculum and LR Schedule Should Be Co-Designed: Math Data Placement Flips With Learning Rate — CatAstro_Piyush · 2026-10-27
• Paper: data schedules and learning rate schedules should be designed together — alifcoder · 2026-10-27