FlashVideo: Flowing Fidelity to Detail for Efficient High-Resolution Video Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917233659215872 |
|---|---|
| author | Zhang, Shilong Li, Wenbo Chen, Shoufa GE, Chongjian Sun, Peize Zhang, Yifu Jiang, Yi Yuan, Zehuan Peng, Bingyue Luo, Ping |
| author_facet | Zhang, Shilong Li, Wenbo Chen, Shoufa GE, Chongjian Sun, Peize Zhang, Yifu Jiang, Yi Yuan, Zehuan Peng, Bingyue Luo, Ping |
| contents | DiT models have achieved great success in text-to-video generation, leveraging their scalability in model capacity and data scale. High content and motion fidelity aligned with text prompts, however, often require large model parameters and a substantial number of function evaluations (NFEs). Realistic and visually appealing details are typically reflected in high-resolution outputs, further amplifying computational demands-especially for single-stage DiT models. To address these challenges, we propose a novel two-stage framework, FlashVideo, which strategically allocates model capacity and NFEs across stages to balance generation fidelity and quality. In the first stage, prompt fidelity is prioritized through a low-resolution generation process utilizing large parameters and sufficient NFEs to enhance computational efficiency. The second stage achieves a nearly straight ODE trajectory between low and high resolutions via flow matching, effectively generating fine details and fixing artifacts with minimal NFEs. To ensure a seamless connection between the two independently trained stages during inference, we carefully design degradation strategies during the second-stage training. Quantitative and visual results demonstrate that FlashVideo achieves state-of-the-art high-resolution video generation with superior computational efficiency. Additionally, the two-stage design enables users to preview the initial output and accordingly adjust the prompt before committing to full-resolution generation, thereby significantly reducing computational costs and wait times as well as enhancing commercial viability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_05179 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FlashVideo: Flowing Fidelity to Detail for Efficient High-Resolution Video Generation Zhang, Shilong Li, Wenbo Chen, Shoufa GE, Chongjian Sun, Peize Zhang, Yifu Jiang, Yi Yuan, Zehuan Peng, Bingyue Luo, Ping Computer Vision and Pattern Recognition DiT models have achieved great success in text-to-video generation, leveraging their scalability in model capacity and data scale. High content and motion fidelity aligned with text prompts, however, often require large model parameters and a substantial number of function evaluations (NFEs). Realistic and visually appealing details are typically reflected in high-resolution outputs, further amplifying computational demands-especially for single-stage DiT models. To address these challenges, we propose a novel two-stage framework, FlashVideo, which strategically allocates model capacity and NFEs across stages to balance generation fidelity and quality. In the first stage, prompt fidelity is prioritized through a low-resolution generation process utilizing large parameters and sufficient NFEs to enhance computational efficiency. The second stage achieves a nearly straight ODE trajectory between low and high resolutions via flow matching, effectively generating fine details and fixing artifacts with minimal NFEs. To ensure a seamless connection between the two independently trained stages during inference, we carefully design degradation strategies during the second-stage training. Quantitative and visual results demonstrate that FlashVideo achieves state-of-the-art high-resolution video generation with superior computational efficiency. Additionally, the two-stage design enables users to preview the initial output and accordingly adjust the prompt before committing to full-resolution generation, thereby significantly reducing computational costs and wait times as well as enhancing commercial viability. |
| title | FlashVideo: Flowing Fidelity to Detail for Efficient High-Resolution Video Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2502.05179 |