Complete-muE: Optimal Hyperparameter Transfer and Scaling for MoE Models
cs.LG
Submitted: 2026-05-22
Updated: 2026-09-27
Terminology
Sources
- ST-MoE: Designing Stable and Transferable Sparse Expert Models
- DeepSeek-V3 Technical Report
- Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity
- Pangu Ultra MoE: How to Train Your Big MoE on Ascend NPUs
- LongCat-Flash Technical Report
- Completed Hyperparameter Transfer across Modules, Width, Depth, Batch and Duration
- GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding
- EC-DIT: Scaling Diffusion Transformers with Adaptive Expert-Choice Routing
- Nucleus-Image: Sparse MoE for Image Generation
- Hyperparameter Transfer with Mixture-of-Expert Layers
- Rethinking Language Model Scaling under Transferable Hypersphere Optimization
- Qwen3-VL Technical Report
- Flow Matching for Generative Modeling
- GLU Variants Improve Transformer
- Why beta 1 = beta 2 Is Dynamically Special in Adam
- SonicMoE: Accelerating MoE with IO and Tile-aware Optimizations
- KnapFormer: An Online Load Balancer for Efficient Diffusion Transformers Training
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks