SpecServe: Efficient and SLO-Aware Large Language Model Serving with Adaptive Speculative Decoding
Seminar
Coming Soon
현재 확정된 세미나 일정이 없습니다. 보통 매주 금요일 오후 7시 애기능 생활관 302호에서 진행됩니다.
Recent Seminars
-
SpecServe: Efficient and SLO-Aware Large Language Model Serving with Adaptive Speculative Decoding
-
Preble: Efficient Distributed Prompt Scheduling for LLM Serving
-
HotSwap: Enabling Live Dependency Sharing in Serverless Computing
-
Telemetry-Driven Microservices Orchestration in Cloud-Edge Environments
-
ModServe Modality- and Stage-Aware Resource Disaggregation for Scalable Multimodal Model Serving
-
WindServe: Efficient Phase-Disaggregated LLM Serving with Stream-based Dynamic Scheduling