Some Date, Some Place, Some Country
Contact Us:
jxwuyi@mail.tsinghua.edu.cn | 𝕏 @OutTwitter
The open-sourcing paradigm for LLMs is largely shifting from only open-sourcing the final models to a more complete training recipe that allows everyone to reproduce the LLMs’ training process, covering dataset preparation, training infrastructures, algorithm innovations, and evaluation. This workshop aims to create a venue for the entire open-source community to get together to share recent achievements and collaborate for future open-source advances. It can also be a great chance for the NeurIPS community to participate and build upon the open-source ecosystem for research advances. The organizers will make their best efforts to link leading contributors from the US, China, and Europe for a global event.
Ion Stoica
UC Berkeley
Junyang Lin
Alibaba Group
Daya Guo
DeepSeek
Noah Smith
University of Washington & Allen Institute for AI
Percy Liang
Stanford University
Leandro von Werra
Hugging Face
Yi Wu
Tsinghua University & Ant Research
Haibin Lin
ByteDance
Michael Luo
UC Berkeley
Zijian Zhang
Tsinghua University & NVIDIA
Jiaxuan Gao
Tsinghua University & Ant Research
Chenyang Zhao
UCLA & SGLang
Song Han
MIT & NVIDIA