4 papers
JustGoal: With Only a Goal, the AI Baseline Is Above the Expert Kernels
Yaojian Chen, Dazhi Peng, Zhui Zhu +13
A small number of GPU kernels often determines the efficiency of large language model training and inference. High-performance kernel development typically relies on expert knowled…
Self-Evolving Agent Harness
Chuyang Wei, Haoxiang Guan, Minghang Zhu +5
Harness self-evolution revises an agent's software scaffolding, including prompts, memory, workflows, and executable components, while keeping model weights fixed. We formalize a f…
ForgeTrain: Forging Production-Grade Training Frameworks via Harness-Driven AI Development
Qingfeng He, Zhui Zhu, Shangzhan Li +15
Training large models still relies on general-purpose frameworks such as Megatron-LM, whose generality tax constrains scenario-specific optimization and adds runtime overhead throu…
ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search
Jiyan He, Guang Liang, Hao Liu +19
In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency. ZGCM-1 is founded on a core prem…