implementation detail · filed under optimization
Moonlight-style learning-rate scaling
Applying Moonlight-style learning-rate scaling to Muon.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
we adopt Moonlight-style learning-rate scaling [49].
usedoptimizationin LAGUNA M.1 and LAGUNA XS.2Poolside
Filed alongside
Other methods under optimization :: optimizer.
MuonPer-Head MuonAdamWCategory-specific assignment of Muon and AdamWSplit fused gradients before orthogonalizationNesterov momentumSinkhorn-balanced updateAdam without weight decay for the N-gram embedding tableAdamW for attention and GDN output gatesAdamW for gated-residual low-rank projectionsAdamW for input embeddings and output headAdamW for the MoE routerAsynchronous Micro-Group pipelineCanzonaCUDA graph capture of the optimizer stepEight-step Newton–Schulz iterationHybrid Muon and AdamW parameter-group optimizer assignmentHybrid Newton–Schulz iterationsHybrid optimization with Muon and AdamMuon orthogonalization accuracy refinementMuon restricted to two-dimensional linear-map weightsMuon SplitMuownPeer-to-peer shard retrieval for Muon orthogonalization