arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2210.02414cs.CLcs.AIcs.LG
收藏

GLM-130B:一个开放的双语预训练模型

GLM-130B: An Open Bilingual Pre-trained Model

Tsinghua University(清华大学)
Aohan Zeng, Xiao Liu, Zhengxiao Du, Zihan Wang, Hanyu Lai, Ming Ding, Zhuoyi Yang, Yifan Xu, Wendi Zheng, Xiao Xia, Weng Lam Tam, Zixuan Ma, Yufei Xue, Jidong Z… 展开作者

Aohan Zeng, Xiao Liu, Zhengxiao Du, Zihan Wang, Hanyu Lai, Ming Ding, Zhuoyi Yang, Yifan Xu, Wendi Zheng, Xiao Xia, Weng Lam Tam, Zixuan Ma, Yufei Xue, Jidong Zhai, Wenguang Chen, Peng Zhang, Yuxiao Dong, Jie Tang

AI总结:

本文介绍了GLM-130B,一个1300亿参数的双语预训练模型,通过独特设计克服训练挑战,在多项基准上超越GPT-3和ERNIE TITAN,并实现INT4量化以降低推理成本。

摘要:

我们介绍了GLM-130B,一个拥有1300亿参数的双语(英语和中文)预训练语言模型。这是一次尝试,旨在开源一个至少与GPT-3(davinci)同样优秀的1000亿规模模型,并揭示如何成功预训练这种规模的模型。在这一努力过程中,我们面临了众多意想不到的技术和工程挑战,特别是在损失尖峰和发散方面。本文介绍了GLM-130B的训练过程,包括其设计选择、兼顾效率和稳定性的训练策略以及工程努力。最终的GLM-130B模型在众多流行的英语基准测试上显著优于GPT-3 175B(davinci),而在OPT-175B和BLOOM-176B中并未观察到这种性能优势。在相关基准测试中,它也持续且显著地优于最大的中文语言模型ERNIE TITAN 3.0 260B。最后,我们利用GLM-130B独特的缩放特性实现了无需后训练的INT4量化,几乎没有性能损失,使其成为1000亿规模模型中的首个,更重要的是,允许在4×RTX 3090(24G)或8×RTX 2080 Ti(11G)GPU上进行有效推理,这是使用1000亿规模模型所需的最低成本GPU。GLM-130B模型权重可公开访问,其代码、训练日志、相关工具包和经验教训已在https://github.com/THUDM/GLM-130B/开源。

英文摘要:

We introduce GLM-130B, a bilingual (English and Chinese) pre-trained language model with 130 billion parameters. It is an attempt to open-source a 100B-scale model at least as good as GPT-3 (davinci) and unveil how models of such a scale can be successfully pre-trained. Over the course of this effort, we face numerous unexpected technical and engineering challenges, particularly on loss spikes and divergence. In this paper, we introduce the training process of GLM-130B including its design choices, training strategies for both efficiency and stability, and engineering efforts. The resultant GLM-130B model offers significant outperformance over GPT-3 175B (davinci) on a wide range of popular English benchmarks while the performance advantage is not observed in OPT-175B and BLOOM-176B. It also consistently and significantly outperforms ERNIE TITAN 3.0 260B -- the largest Chinese language model -- across related benchmarks. Finally, we leverage a unique scaling property of GLM-130B to reach INT4 quantization without post training, with almost no performance loss, making it the first among 100B-scale models and more importantly, allowing its effective inference on 4$\times$RTX 3090 (24G) or 8$\times$RTX 2080 Ti (11G) GPUs, the most affordable GPUs required for using 100B-scale models. The GLM-130B model weights are publicly accessible and its code, training logs, related toolkit, and lessons learned are open-sourced at \url{https://github.com/THUDM/GLM-130B/}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑