MusaCoder: Native GPU Kernel Generation with Full-Stack Training on Moore Threads GPU

Kun Cheng, Songshuo Lu, Sicong Liao +7

A Full-duplex Speech Dialogue Scheme Based On Large Language Models

Peng Wang, Songshuo Lu, Yaohua Tang +3

URPO: A Unified Reward & Policy Optimization Framework for Large Language Models

Songshuo Lu, Hua Wang, Zhi Chen +1

TurboRAG: Accelerating Retrieval-Augmented Generation with Precomputed KV Caches for Chunked Text

Songshuo Lu, Hua Wang, Yutian Rong +2

Beyond the Best Teacher: Expanding and Compressing the Reasoning Solution Manifold

Songshuo Lu, Zhi Chen, Yaohua Tang

The paper proposes an expand‑then‑compress framework that builds a diverse set of RL‑trained teacher models and then distills them into a single student model, improving reasoning,…