OpenBA: An Open-sourced 15B Bilingual Asymmetric seq2seq Model
  Pre-trained from Scratch

Juntao Li; Zecheng Tang; Yuyang Ding; Pinzheng Wang; Pei Guo; Wangjie You; Dan Qiao; Wenliang Chen; Guohong Fu; Qiaoming Zhu; Guodong Zhou; Min Zhang

Paper · arXiv 2309.10706

OpenBA: An Open-sourced 15B Bilingual Asymmetric seq2seq Model Pre-trained from Scratch

Juntao Li, Zecheng Tang, Yuyang Ding, Pinzheng Wang, Pei Guo, Wangjie You, Dan Qiao, Wenliang Chen, Guohong Fu, Qiaoming Zhu, Guodong Zhou, Min Zhang

17 upvotesSeptember 19, 2023arXiv 预印本

AI 摘要

OpenBA, a 15B bilingual asymmetric seq2seq model, achieves competitive performance on natural language processing tasks using an efficient three-stage training strategy and minimal pre-training data.

bilingual asymmetric seq2seq modelOpenBAthree-stage training strategyBilingual Flan data collectionempirical observationsmodel architecture designtraining objectivespre-training data processingHuggingface Transformers Library

Abstract

Large language models (LLMs) with billions of parameters have demonstrated outstanding performance on various natural language processing tasks. This report presents OpenBA, an open-sourced 15B bilingual asymmetric seq2seq model, to contribute an LLM variant to the Chinese-oriented open-source model community. We enhance OpenBA with effective and efficient techniques as well as adopt a three-stage training strategy to train the model from scratch. Our solution can also achieve very competitive performance with only 380B tokens, which is better than LLaMA-70B on the BELEBELE benchmark, BLOOM-176B on the MMLU benchmark, GLM-130B on the C-Eval (hard) benchmark. This report provides the main details to pre-train an analogous model, including pre-training data processing, Bilingual Flan data collection, the empirical observations that inspire our model architecture design, training objectives of different stages, and other enhancement techniques. We have refactored our code to follow the design principles of the Huggingface Transformers Library, making it more convenient for developers to use, and released checkpoints of different training stages at https://huggingface.co/openBA. More details of our project are available at https://github.com/OpenNLG/openBA.git.