MoE-SpAc: Efficient MoE Inference Based on Speculative Activation Utility in Heterogeneous Edge Scenarios
Paper
• 2603.09983 • Published
• 1
Li shuhuai
lsh-algorithm
AI & ML interests
None yet
Recent Activity
authored
a paper
3 days ago
MoE-SpAc: Efficient MoE Inference Based on Speculative Activation Utility in Heterogeneous Edge Scenarios upvoted a paper 9 months ago
Online Speculative Decoding Organizations
None yet
authored a paper 3 days ago