Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.18855 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908378787217408 |
|---|---|
| author | Wang, Peiqi Peng, ShengYun Zhang, Xuewen Yu, Hanchao Yang, Yibo Huang, Lifu Liu, Fujun Wang, Qifan |
| author_facet | Wang, Peiqi Peng, ShengYun Zhang, Xuewen Yu, Hanchao Yang, Yibo Huang, Lifu Liu, Fujun Wang, Qifan |
| contents | This work investigates the optimal allocation of inference compute across three key scaling factors in video vision language models: language model size, frame count, and the number of visual tokens per frame. While prior works typically focuses on optimizing model efficiency or improving performance without considering resource constraints, we instead identify optimal model configuration under fixed inference compute budgets. We conduct large-scale training sweeps and careful parametric modeling of task performance to identify the inference compute-optimal frontier. Our experiments reveal how task performance depends on scaling factors and finetuning data size, as well as how changes in data size shift the compute-optimal frontier. These findings translate to practical tips for selecting these scaling factors. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_18855 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Inference Compute-Optimal Video Vision Language Models Wang, Peiqi Peng, ShengYun Zhang, Xuewen Yu, Hanchao Yang, Yibo Huang, Lifu Liu, Fujun Wang, Qifan Computer Vision and Pattern Recognition Computation and Language This work investigates the optimal allocation of inference compute across three key scaling factors in video vision language models: language model size, frame count, and the number of visual tokens per frame. While prior works typically focuses on optimizing model efficiency or improving performance without considering resource constraints, we instead identify optimal model configuration under fixed inference compute budgets. We conduct large-scale training sweeps and careful parametric modeling of task performance to identify the inference compute-optimal frontier. Our experiments reveal how task performance depends on scaling factors and finetuning data size, as well as how changes in data size shift the compute-optimal frontier. These findings translate to practical tips for selecting these scaling factors. |
| title | Inference Compute-Optimal Video Vision Language Models |
| topic | Computer Vision and Pattern Recognition Computation and Language |
| url | https://arxiv.org/abs/2505.18855 |