Optimization Hyper-parameter Laws for Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916029392748544 |
|---|---|
| author | Xie, Xingyu Ding, Kuangyu Yan, Shuicheng Toh, Kim-Chuan Wei, Tianwen |
| author_facet | Xie, Xingyu Ding, Kuangyu Yan, Shuicheng Toh, Kim-Chuan Wei, Tianwen |
| contents | Large Language Models have driven significant AI advancements, yet their training is resource-intensive and highly sensitive to hyper-parameter selection. While scaling laws provide valuable guidance on model size and data requirements, they fall short in choosing dynamic hyper-parameters, such as learning-rate (LR) schedules, that evolve during training. To bridge this gap, we present Optimization Hyper-parameter Laws (Opt-Laws), a framework that predicts final training loss as a function of LR schedule, model size, and data size. Grounded in SDE-based convergence and escape analyses, Opt-Laws yield interpretable convergence and escape features that predict final training loss across model scales, enabling schedule pre-selection from small-scale experiments. Empirically, Opt-Laws achieve a 94% Top-2 hit rate for identifying near-optimal schedule candidates on held-out configurations, correctly identify the best-performing schedule family in all five evaluated out-of-family settings, and detect training divergence with F1 = 0.92. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_04777 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Optimization Hyper-parameter Laws for Large Language Models Xie, Xingyu Ding, Kuangyu Yan, Shuicheng Toh, Kim-Chuan Wei, Tianwen Machine Learning Optimization and Control Large Language Models have driven significant AI advancements, yet their training is resource-intensive and highly sensitive to hyper-parameter selection. While scaling laws provide valuable guidance on model size and data requirements, they fall short in choosing dynamic hyper-parameters, such as learning-rate (LR) schedules, that evolve during training. To bridge this gap, we present Optimization Hyper-parameter Laws (Opt-Laws), a framework that predicts final training loss as a function of LR schedule, model size, and data size. Grounded in SDE-based convergence and escape analyses, Opt-Laws yield interpretable convergence and escape features that predict final training loss across model scales, enabling schedule pre-selection from small-scale experiments. Empirically, Opt-Laws achieve a 94% Top-2 hit rate for identifying near-optimal schedule candidates on held-out configurations, correctly identify the best-performing schedule family in all five evaluated out-of-family settings, and detect training divergence with F1 = 0.92. |
| title | Optimization Hyper-parameter Laws for Large Language Models |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2409.04777 |