Cramming: Training a Language Model on a single GPU in one day.

Jonas Geiping, Tom Goldstein
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:11117-11143, 2023.

Abstract

Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day? We investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU. Aside from re-analyzing nearly all components of the pretraining pipeline for this scenario and providing a modified pipeline with performance close to BERT, we investigate why scaling down is hard, and which modifications actually improve performance in this scenario. We provide evidence that even in this constrained setting, performance closely follows scaling laws observed in large-compute settings. Through the lens of scaling laws, we categorize a range of recent improvements to training and architecture and discuss their merit and practical applicability (or lack thereof) for the limited compute setting. We provide code to reproduce all experiments at github.com/JonasGeiping/cramming .

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-geiping23a, title = {Cramming: Training a Language Model on a single {GPU} in one day.}, author = {Geiping, Jonas and Goldstein, Tom}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {11117--11143}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/geiping23a/geiping23a.pdf}, url = {https://proceedings.mlr.press/v202/geiping23a.html}, abstract = {Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day? We investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU. Aside from re-analyzing nearly all components of the pretraining pipeline for this scenario and providing a modified pipeline with performance close to BERT, we investigate why scaling down is hard, and which modifications actually improve performance in this scenario. We provide evidence that even in this constrained setting, performance closely follows scaling laws observed in large-compute settings. Through the lens of scaling laws, we categorize a range of recent improvements to training and architecture and discuss their merit and practical applicability (or lack thereof) for the limited compute setting. We provide code to reproduce all experiments at github.com/JonasGeiping/cramming .} }
Endnote
%0 Conference Paper %T Cramming: Training a Language Model on a single GPU in one day. %A Jonas Geiping %A Tom Goldstein %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-geiping23a %I PMLR %P 11117--11143 %U https://proceedings.mlr.press/v202/geiping23a.html %V 202 %X Recent trends in language modeling have focused on increasing performance through scaling, and have resulted in an environment where training language models is out of reach for most researchers and practitioners. While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day? We investigate the downstream performance achievable with a transformer-based language model trained completely from scratch with masked language modeling for a single day on a single consumer GPU. Aside from re-analyzing nearly all components of the pretraining pipeline for this scenario and providing a modified pipeline with performance close to BERT, we investigate why scaling down is hard, and which modifications actually improve performance in this scenario. We provide evidence that even in this constrained setting, performance closely follows scaling laws observed in large-compute settings. Through the lens of scaling laws, we categorize a range of recent improvements to training and architecture and discuss their merit and practical applicability (or lack thereof) for the limited compute setting. We provide code to reproduce all experiments at github.com/JonasGeiping/cramming .
APA
Geiping, J. & Goldstein, T.. (2023). Cramming: Training a Language Model on a single GPU in one day.. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:11117-11143 Available from https://proceedings.mlr.press/v202/geiping23a.html.

Related Material