An Empirical Study of $\mu$P Learning Rate Transfer
Deep learning models have become a cornerstone of modern AI research, yet their initializations and learning rates may at times be set in an opaque or ad-hoc fashion due to the high cost of hyperparameter sweeps. The $\mu$-Parameterization ($\mu$P) offers a possible solution to this challenge, yielding scaling rules for model initialization and learning rates while reportedly enabling zero-shot hyperparameter transfer from small to large models. Despite its evident promise, the $\mu$P method is not yet widely adopted, perhaps due to higher implementation complexity, many variations, or complex theoretical background. This work considers $\mu$P empirically, focusing on the popular transformer architecture, and aims to answer a simple question: does $\mu$-Transfer yield near-optimal learning rates in practice? Studying over a dozen ablations with up to 1.2B parameters and 33B tokens and a large-scale experiment with up to 10B parameters and 190B tokens, we observe a positive answer for most settings, and discuss improvements otherwise.