Polynomial schedule helper
Usage
poly_decay(alpha = 0.501, t0 = 1, burnin_iter = 0)
Arguments
- alpha
polynomial exponent in \((1/2,1)\).
- t0
non-negative schedule offset.
- burnin_iter
non-negative integer. Initial iterations without
polynomial schedule scaling.
Details
Convenience helper that enables polynomial schedule and disables
checkpoint-based decay.