Turn Waste into Worth: Rectifying Top-$k$ Router of MoE
원문 초록
Sparse Mixture of Experts (MoE) models are popular for training large language models due to their computational efficiency. However, the commonly used top-$k$ routing mechanism suffers from redundancy computation and memory costs due to the unbalanced routing. Some experts are overflow, where the exceeding tokens are dropped. While some experts are vacant, which are padded with zeros, negatively impacting model performance. To address the dropped tokens and padding, we propose the Rectify-Router, comprising the Intra-GPU Rectification and the Fill-in Rectification. The Intra-GPU Rectification handles dropped tokens, efficiently routing them to experts within the GPU where they are located to avoid inter-GPU communication. The Fill-in Rectification addresses padding by replacing padding tokens with the tokens that have high routing scores. Our experimental results demonstrate that the Intra-GPU Rectification and the Fill-in Rectification effectively handle dropped tokens and padding, respectively. Furthermore, the combination of them achieves superior performance, surpassing the accuracy of the vanilla top-1 router by 4.7%.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
A Survey on Prompt Tuning
FADRM: Fast and Accurate Data Residual Matching for Dataset Distillation
VolumetricSMPL: A Neural Volumetric Body Model for Efficient Interactions, Contacts, and Collisions
Fast ground penetrating radar dual-parameter full waveform inversion method accelerated by hybrid compilation of CUDA kernel function and PyTorch
FourCastNet 3: A geometric approach to probabilistic machine-learning weather forecasting at scale
Lightweight Model for Poultry Disease Detection from Fecal Images Using Multi-Color Space Feature Optimization and Machine Learning
토론
이 논문과 연결된 포럼 글이 아직 없습니다.