Research map / Architectures and efficiency / Pruning and efficient tuning
LoRA meets Riemannion: Muon Optimizer for Parametrization-independent Low-Rank Adapters
Vladimir Bogachev, Vladimir Aletov, Alexander Molozhavenko, Denis Bobkov, Vera Soboleva, Aibek Alanov, Maxim Rakhuba
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace Adam-LoRA with Riemannion optimizer | LoRA rank 16 · Commonsense reasoning (8 tasks avg), Llama 3-8B | accuracy (%): 87.1 -> 88.1 | Improved |
| use Riemannion instead of Muon applied per LoRA factor | LoRA rank 16 · Commonsense reasoning (8 tasks avg), Llama 3-8B | accuracy (%): 83.0 -> 88.1 | Improved |
| use Riemannion instead of RPrecAdamW (Riemannian-aware Adam variant) | LoRA rank 16 · Commonsense reasoning (8 tasks avg), Llama 3-8B | accuracy (%): 86.8 -> 88.1 | Improved |
| apply Muon directly to LoRA factors | LoRA Adam baseline · Commonsense reasoning (8 tasks avg), Llama 3-8B | accuracy (%): 87.1 -> 83.0 | Hurt |
| add locally optimal initialization (LOI) to LoRA (ablation) | LoRA · Commonsense reasoning (8 tasks avg), Llama 3.2-1B | accuracy (%): 67.9 -> 70.1 | Improved |
| add LOI to Riemannian SGD-type LoRA (RiemannLoRA) (ablation) | RiemannLoRA · Commonsense reasoning (8 tasks avg), Llama 3.2-1B | accuracy (%): 69.6 -> 73.4 | Improved |
| use Riemannian optimization instead of Euclidean LoRA (no LOI) (ablation) | LoRA · Commonsense reasoning (8 tasks avg), Llama 3.2-1B | accuracy (%): 67.9 -> 69.6 | Improved |
| Riemannion vs Adam LoRA on individual tasks | LoRA rank 16 · BoolQ, Llama 3-8B | accuracy (%): 74.8 -> 75.7 | No clear effect |
| solve tangent-space LMO exactly or by alternating projection instead of one projection step (ablation) | Riemannion · convergence | convergence | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DoRA — DoRA: Weight-Decomposed Low-Rank Adaptation (ICML 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Riemannian Preconditioned LoRA for Fine-Tuning Foundation Models (ICML 2024 · similarity 0.88)
- Intrinsic Muon: Spectral Optimization on Riemannian Matrix Manifolds (NeurIPS 2026 · similarity 0.88)
- A Stronger Mixture of Low-Rank Experts for Fine-Tuning Foundation Models (ICML 2025 · similarity 0.87)
- AdaPreLoRA: Adafactor Preconditioned Low-Rank Adaptation (NeurIPS 2026 · similarity 0.87)
- LoRA Done RITE: Robust Invariant Transformation Equilibration for LoRA Optimization (ICLR 2025 · similarity 0.86)
- StelLA: Subspace Learning in Low-rank Adaptation using Stiefel Manifold (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
