Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2501.06954 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909613872381952 |
|---|---|
| author | Xu, Shiyun Bu, Zhiqi Zhang, Yiliang Barnett, Ian |
| author_facet | Xu, Shiyun Bu, Zhiqi Zhang, Yiliang Barnett, Ian |
| contents | Differential learning rate (DLR), a technique that applies different learning rates to different model parameters, has been widely used in deep learning and achieved empirical success via its various forms. For example, parameter-efficient fine-tuning (PEFT) applies zero learning rates to most parameters so as to significantly save the computational cost.
At the core, DLR leverages the observation that different parameters can have different loss curvature, which is hard to characterize in general. We propose the Hessian-informed differential learning rate (Hi-DLR), an efficient approach that solves the hyperparameter optimization (HPO) of learning rates and captures the loss curvature for any model and optimizer adaptively. Given a proper grouping of parameters, we empirically demonstrate that Hi-DLR can improve the convergence by dynamically determining the learning rates during the training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_06954 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Hessian-informed hyperparameter optimization for differential learning rate Xu, Shiyun Bu, Zhiqi Zhang, Yiliang Barnett, Ian Machine Learning Differential learning rate (DLR), a technique that applies different learning rates to different model parameters, has been widely used in deep learning and achieved empirical success via its various forms. For example, parameter-efficient fine-tuning (PEFT) applies zero learning rates to most parameters so as to significantly save the computational cost. At the core, DLR leverages the observation that different parameters can have different loss curvature, which is hard to characterize in general. We propose the Hessian-informed differential learning rate (Hi-DLR), an efficient approach that solves the hyperparameter optimization (HPO) of learning rates and captures the loss curvature for any model and optimizer adaptively. Given a proper grouping of parameters, we empirically demonstrate that Hi-DLR can improve the convergence by dynamically determining the learning rates during the training. |
| title | A Hessian-informed hyperparameter optimization for differential learning rate |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2501.06954 |