GRAWA: Gradient-based Weighted Averaging for Distributed Training of Deep Learning Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917606697467904 |
|---|---|
| author | Dimlioglu, Tolga Choromanska, Anna |
| author_facet | Dimlioglu, Tolga Choromanska, Anna |
| contents | We study distributed training of deep learning models in time-constrained environments. We propose a new algorithm that periodically pulls workers towards the center variable computed as a weighted average of workers, where the weights are inversely proportional to the gradient norms of the workers such that recovering the flat regions in the optimization landscape is prioritized. We develop two asynchronous variants of the proposed algorithm that we call Model-level and Layer-level Gradient-based Weighted Averaging (resp. MGRAWA and LGRAWA), which differ in terms of the weighting scheme that is either done with respect to the entire model or is applied layer-wise. On the theoretical front, we prove the convergence guarantee for the proposed approach in both convex and non-convex settings. We then experimentally demonstrate that our algorithms outperform the competitor methods by achieving faster convergence and recovering better quality and flatter local optima. We also carry out an ablation study to analyze the scalability of the proposed algorithms in more crowded distributed training environments. Finally, we report that our approach requires less frequent communication and fewer distributed updates compared to the state-of-the-art baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_04206 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GRAWA: Gradient-based Weighted Averaging for Distributed Training of Deep Learning Models Dimlioglu, Tolga Choromanska, Anna Machine Learning Distributed, Parallel, and Cluster Computing Optimization and Control We study distributed training of deep learning models in time-constrained environments. We propose a new algorithm that periodically pulls workers towards the center variable computed as a weighted average of workers, where the weights are inversely proportional to the gradient norms of the workers such that recovering the flat regions in the optimization landscape is prioritized. We develop two asynchronous variants of the proposed algorithm that we call Model-level and Layer-level Gradient-based Weighted Averaging (resp. MGRAWA and LGRAWA), which differ in terms of the weighting scheme that is either done with respect to the entire model or is applied layer-wise. On the theoretical front, we prove the convergence guarantee for the proposed approach in both convex and non-convex settings. We then experimentally demonstrate that our algorithms outperform the competitor methods by achieving faster convergence and recovering better quality and flatter local optima. We also carry out an ablation study to analyze the scalability of the proposed algorithms in more crowded distributed training environments. Finally, we report that our approach requires less frequent communication and fewer distributed updates compared to the state-of-the-art baselines. |
| title | GRAWA: Gradient-based Weighted Averaging for Distributed Training of Deep Learning Models |
| topic | Machine Learning Distributed, Parallel, and Cluster Computing Optimization and Control |
| url | https://arxiv.org/abs/2403.04206 |