Shanghai, China

TianyuLiu

Advanced
@smart-lty

Joint PhD. student at USTC&Shanghai AI Lab

smart-lty.github.io/

nano-PEARL. Draft-Target Disaggregation LLM Serving System via Parallel Speculative Decoding.

211

ParallelSpeculativeDecoding. [ICLR 2025] PEARL: Parallel Speculative Decoding with Adaptive Draft Length

170

Claude-Team. Claude, Codex, Gemini 协作驱动的 Claude-Team!

93

REST. The code of paper "Learning Rule-Induced Subgraph Representations for Inductive Relation Prediction" in NeurIPS 2023.

14

LogitSpec. [ACL 26 Findings] LogitSpec: Accelerating Retrieval-based Speculative Decoding via Next Next Token Speculation

11

personal_daily_arxiv. This repo is used to search papers on Arxiv with specific keywords in their title and abstract.

3

smart-lty. about me

1

agent-notebooks. Shell

1
8
Apply