Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.15271 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917993089335296 |
|---|---|
| author | Chen, Guo Li, Zhiqi Wang, Shihao Jiang, Jindong Liu, Yicheng Lu, Lidong Huang, De-An Byeon, Wonmin Le, Matthieu Rintamaki, Tuomas Poon, Tyler Ehrlich, Max Rintamaki, Tuomas Poon, Tyler Lu, Tong Wang, Limin Catanzaro, Bryan Kautz, Jan Tao, Andrew Yu, Zhiding Liu, Guilin |
| author_facet | Chen, Guo Li, Zhiqi Wang, Shihao Jiang, Jindong Liu, Yicheng Lu, Lidong Huang, De-An Byeon, Wonmin Le, Matthieu Rintamaki, Tuomas Poon, Tyler Ehrlich, Max Rintamaki, Tuomas Poon, Tyler Lu, Tong Wang, Limin Catanzaro, Bryan Kautz, Jan Tao, Andrew Yu, Zhiding Liu, Guilin |
| contents | We introduce Eagle 2.5, a family of frontier vision-language models (VLMs) for long-context multimodal learning. Our work addresses the challenges in long video comprehension and high-resolution image understanding, introducing a generalist framework for both tasks. The proposed training framework incorporates Automatic Degrade Sampling and Image Area Preservation, two techniques that preserve contextual integrity and visual details. The framework also includes numerous efficiency optimizations in the pipeline for long-context data training. Finally, we propose Eagle-Video-110K, a novel dataset that integrates both story-level and clip-level annotations, facilitating long-video understanding. Eagle 2.5 demonstrates substantial improvements on long-context multimodal benchmarks, providing a robust solution to the limitations of existing VLMs. Notably, our best model Eagle 2.5-8B achieves 72.4% on Video-MME with 512 input frames, matching the results of top-tier commercial model such as GPT-4o and large-scale open-source models like Qwen2.5-VL-72B and InternVL2.5-78B. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_15271 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Eagle 2.5: Boosting Long-Context Post-Training for Frontier Vision-Language Models Chen, Guo Li, Zhiqi Wang, Shihao Jiang, Jindong Liu, Yicheng Lu, Lidong Huang, De-An Byeon, Wonmin Le, Matthieu Rintamaki, Tuomas Poon, Tyler Ehrlich, Max Rintamaki, Tuomas Poon, Tyler Lu, Tong Wang, Limin Catanzaro, Bryan Kautz, Jan Tao, Andrew Yu, Zhiding Liu, Guilin Computer Vision and Pattern Recognition We introduce Eagle 2.5, a family of frontier vision-language models (VLMs) for long-context multimodal learning. Our work addresses the challenges in long video comprehension and high-resolution image understanding, introducing a generalist framework for both tasks. The proposed training framework incorporates Automatic Degrade Sampling and Image Area Preservation, two techniques that preserve contextual integrity and visual details. The framework also includes numerous efficiency optimizations in the pipeline for long-context data training. Finally, we propose Eagle-Video-110K, a novel dataset that integrates both story-level and clip-level annotations, facilitating long-video understanding. Eagle 2.5 demonstrates substantial improvements on long-context multimodal benchmarks, providing a robust solution to the limitations of existing VLMs. Notably, our best model Eagle 2.5-8B achieves 72.4% on Video-MME with 512 input frames, matching the results of top-tier commercial model such as GPT-4o and large-scale open-source models like Qwen2.5-VL-72B and InternVL2.5-78B. |
| title | Eagle 2.5: Boosting Long-Context Post-Training for Frontier Vision-Language Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.15271 |