Stepped MoE: Segment-Level Routing with Configurable Inference Complexity
Training large language models (LLMs) is resource-intensive, and adapting them for diverse deployment scenarios with varying computational constraints remains challenging. While elastic architectures enable flexible model deployment and sparsely activated models allow input-adaptive computation, existing approaches tre...
Arnav Kundu, Zhao-Yang Xu, Bai-Ru Hou et al.
· 0 citations