Paper
Scaling Up Models and Data with $\texttt{t5x}$ and $\texttt{seqio}$
Authors
Adam Roberts, Hyung Won Chung, Anselm Levskaya, Gaurav Mishra, James Bradbury, D. Andor, Sharan Narang, Brian Lester, Colin Gaffney, Afroz Mohiuddin, Curtis Hawthorne, Aitor Lewkowycz, Alexandru Salcianu, Marc van Zee, Jacob Austin, Sebastian Goodman, Livio Baldini Soares, Haitang Hu, Sasha Tsvyashchenko, A. Chowdhery, Jasmijn Bastings, Jannis Bulian, Xavier García, Jianmo Ni, A. Chen, Kathleen Kenealy, J. Clark, Stephan Lee, Dan Garrette, J. Lee-Thorp, Colin Raffel, Noam Shazeer, Marvin Ritter, Maarten Bosma, Alexandre Passos, Jeremy B. Maitin-Shepard, Noah Fiedel, Mark Omernick, Brennan Saeta, Ryan Sepassi, A. Spiridonov, Joshua Newlan, Andrea Gesmundo
Abstract
Recent neural network-based language models have benefited greatly from scaling up the size of training datasets and the number of parameters in the models themselves. Scaling can be complicated due to various factors including the need to distribute computation on supercomputer clusters (e.g., TPUs), prevent bottlenecks when infeeding data, and ensure reproducible results. In this work, we present two software libraries that ease these issues: simplifies the process of building and training large language models at scale while maintaining ease of use, and provides a task-based API for simple creation of fast and reproducible training data and evaluation pipelines. These open-source libraries have been used to train models with hundreds of billions of parameters on datasets with multiple terabytes of training data.
Along with the libraries, we release configurations and instructions for T5-like encoder-decoder models as well as GPT-like decoder-only architectures.
and are open source and available at https://github.com/google-research/t5x and https://github.com/google/seqio, respectively.