Muon Reduces the Training Cost of Regulatory DNA Transformers
Viraj Doshi, Malhar Bhide, Akshat Singh, Yash Rathod, Aditya Sivakumar
This paper compares Adam and Muon optimizers for training regulatory DNA Transformer models, finding Muon reduces compute costs while achieving similar performance.
bioRxiv · Jul 22, 2026 · PreprintGenomics