Efficient Large Scale Language Modeling with Mixtures of Experts
Mikel Artetxe | Shruti Bhosale | Naman Goyal | Todor Mihaylov | Myle Ott | Sam Shleifer | Xi Victoria Lin | Jingfei Du | Srinivasan Iyer | Ramakanth Pasunuru | Giridharan Anantharaman | Xian Li | Shuohui Chen | Halil Akin | Mandeep Baines | Louis Martin | Xing Zhou | Punit Singh Koura | Brian O’Horo | Jeffrey Wang | Luke Zettlemoyer | Mona Diab | Zornitsa Kozareva | Veselin Stoyanov
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Mixture of Experts layers (MoEs) enable efficient scaling of language models through conditional computation. This paper presents a detailed empirical study of how autoregressive MoE language models scale in comparison with dense models in a wide range of settings: in- and out-of-domain language modeling, zero- and few-shot priming, and full-shot fine-tuning. With the exception of fine-tuning, we find MoEs to be substantially more compute efficient. At more modest training budgets, MoEs can match the performance of dense models using ~4 times less compute. This gap narrows at scale, but our largest MoE model (1.1T parameters) consistently outperforms a compute-equivalent dense model (6.7B parameters). Overall, this performance gap varies greatly across tasks and domains, suggesting that MoE and dense models generalize differently in ways that are worthy of future study. We make our code and models publicly available for research use.
- Mikel Artetxe 1
- Shruti Bhosale 1
- Naman Goyal 1
- Todor Mihaylov 1
- Myle Ott 1
- show all...