The first project I led at Netflix Eyeline Studios is headed to #CVPR2025 with 5,5,4 review scores: 🌊Go-with-the-Flow🌊 warps noise for effortless motion control in video diffusion — no pipeline changes, same compute. Direct camera/object motion, transfer movement between videos, first frame editing, and even tackle text-to-video 3D scenes, all with a simple fine-tuning. Paper: arxiv.org/pdf/2501.08331 Web: eyeline-research.github.io/G… Code: github.com/Eyeline-Research/… Models: huggingface.co/Eyeline-Resea… My great honor to work with the team: @RyanBurgert, @Yuancheng_Xu0, @wenqi_xian, Oliver Pilarski, Pascal Clausen, Mingming He, @maleewahaha, @yitong_deng, Lingxiao Li, Mohsen Mousavi, @ryoo_michael, @debfx, from @eyelinestudios, @Scanline_VFX, @netflix, Stony Brook University, University of Maryland, and @Stanford. #MachineLearning #video #VideoGeneration #DiffusionModels #VideoDiffusionModels #OpenSource

The first project I led at Netflix Eyeline Studios is headed to #CVPR2025 with 5,5,4 review scores: 🌊Go-with-the-Flow🌊 warps noise for effortless motion control in video diffusion — no pipeline changes, same compute. Direct camera/object motion, transfer movement between videos, first frame editing, and even tackle text-to-video 3D scenes, all with a simple fine-tuning. Paper: arxiv.org/pdf/2501.08331 Web: eyeline-research.github.io/G… Code: github.com/Eyeline-Research/… Models: huggingface.co/Eyeline-Resea… My great honor to work with the team: @RyanBurgert, @Yuancheng_Xu0, @wenqi_xian, Oliver Pilarski, Pascal Clausen, Mingming He, @maleewahaha, @yitong_deng, Lingxiao Li, Mohsen Mousavi, @ryoo_michael, @debfx, from @eyelinestudios, @Scanline_VFX, @netflix, Stony Brook University, University of Maryland, and @Stanford. #MachineLearning #video #VideoGeneration #DiffusionModels #VideoDiffusionModels #OpenSource

Kudos to the research team at our sister company @eyelinestudios. Their latest research paper, 🌊Go-with-the-Flow 🌊, will be presented at #CVPR2025! Based on their research, we believe this could allow artists in the future to leverage these new techniques to direct the motion in generated videos, empowering creative control in a wide range of video applications: cut-and-drag animation, transferring movement between videos, first frame editing, camera control via depth warping, and text-to-video 3D scene creation. Kudos to the amazing team: @RyanBurgert, @Yuancheng_Xu0, @wenqi_xian, Oliver Pilarski, Pascal Clausen, Mingming He, @maleewahaha, @yitong_deng, Lingxiao Li, Mohsen Mousavi, @ryoo_michael, @debfx, @realNingYu, from @eyelinestudios, @Scanline_VFX, @netflix, Stony Brook University, University of Maryland, and @Stanford. ***This is part of the ongoing research and development at @eyelinestudios and we hope to see adoption in these techniques and workflows soon. Paper: arxiv.org/pdf/2501.08331 Web: eyeline-research.github.io/G… Code: github.com/Eyeline-Research/… Models: huggingface.co/Eyeline-Resea… #MachineLearning #video #VideoGeneration #DiffusionModels #VideoDiffusionModels #OpenSource