Search results

Learning rate compared to adamw

Status: Open.

#49 In KellerJordan/Muon;

· davnords opened on Sep 1, 2025

Muon optimizer shows no GPU memory reduction with Accelerator mixed precision (bf16)

Status: Open.

#48 In KellerJordan/Muon;

· gitouni opened on Aug 27, 2025

Does <code>Muon</code> save the <code>state["momentum_buffer"]</code> correctly in a distributed setting?

Status: Open.

#46 In KellerJordan/Muon;

· EIFY opened on Aug 21, 2025

Does Muon support FSDP Zero3

Status: Open.

#45 In KellerJordan/Muon;

· lucaswychan opened on Aug 4, 2025

Request: Add torch.compile Support to Muon Optimizer

Status: Open.

#40 In KellerJordan/Muon;

· wertyuilife2 opened on Jul 14, 2025

pip package seems outdated

Status: Open.

#39 In KellerJordan/Muon;

· wendlerc opened on Jul 9, 2025

Guidelines for whether to use muon for parameters

Status: Open.

#31 In KellerJordan/Muon;

· insomniaaac opened on Jun 19, 2025

Finetuning with Muon

Status: Open.

#25 In KellerJordan/Muon;

· dxqb opened on Jun 4, 2025

How to use with FSDP

Status: Open.

#23 In KellerJordan/Muon;

· asahni-sc opened on May 12, 2025

Question about Muon's Different Performance on Dense vs MoE Models

Status: Open.

#19 In KellerJordan/Muon;

· mactavish91 opened on Mar 12, 2025

Use of Muon with Multi Dimensional Parallel training paradigms.

Status: Open.

#18 In KellerJordan/Muon;

· 23Uday opened on Mar 1, 2025

Muon with Llama?

Status: Open.

#9 In KellerJordan/Muon;

· yangsp5 opened on Dec 11, 2024