7 papers
Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training
Zijian Zhang, Rizhen Hu, Athanasios Glentis +4
Reinforcement learning (RL) has become a central component of post-training large language models (LLMs), yet little is understood about how RL adaptation is distributed across tra…
CUDAHercules: Benchmarking Hardware-Aware Expert-level CUDA Optimization for LLMs
Shiyang Li, Zijian Zhang, Guangyan Sun +5
Large language models show promise for automated CUDA programming, however even the strongest coding models (e.g., Claude-Opus-4.6) may still fall short of expert-level, architectu…
InfantAgent-Next: A Multimodal Generalist Agent for Automated Computer Interaction
Bin Lei, Weitai Kang, Zijian Zhang +8
This paper introduces \textsc{InfantAgent-Next}, a generalist agent capable of interacting with computers in a multimodal manner, encompassing text, images, audio, and video. Unlik…
StitchCUDA: An Automated Multi-Agents End-to-End GPU Programing Framework with Rubric-based Agentic Reinforcement Learning
Shiyang Li, Zijian Zhang, Winson Chen +3
Modern machine learning (ML) workloads increasingly rely on GPUs, yet achieving high end-to-end performance remains challenging due to dependencies on both GPU kernel efficiency an…
CudaForge: An Agent Framework with Hardware Feedback for CUDA Kernel Optimization
Zijian Zhang, Rong Wang, Shiyang Li +3
Developing efficient CUDA kernels is increasingly critical for AI applications such as large-scale LLM training. However, manual kernel design is both costly and time-consuming, mo…
Anyprefer: An Agentic Framework for Preference Data Synthesis
Yiyang Zhou, Zhaoyang Wang, Tianle Wang +13
High-quality preference data is essential for aligning foundation models with human values through preference learning. However, manual annotation of such data is often time-consum…