@urielsinger
arxiv.org Exploring the Design Space of Transition Matching Transition Matching (TM) is an emerging paradigm for generative modeling that generalizes diffusion and flow-matching models as well as continuous-state autoregressive models. TM, similar to... ππ¬We introduce TMD (Transition Matching Distillation): 480p videos generated from text prompts in < 3 NFEs!
1οΈβ£Main backbone for feature extraction and lightweight head for iterative refinement
2οΈβ£Distilled from Wan2.1 14B T2V combining MeanFlow & DMD2
πresearch.nvidia.com/labs/genair/tmdView quoted post on X
[1/n]
New paper alert! π
Excited to introduce ππ«ππ§π¬π’ππ’π¨π§ πππππ‘π’π§π (ππ)! We're replacing short-timestep kernels from Flow Matching/Diffusion with... a generative modelπ€―, achieving SOTA text-2-image generation!
@urielsinger @itai_gat @lipmanyaView quoted post on X
This work was done during my internship at @AIatMeta πView quoted post on X
Huge thanks to my amazing collaborators @urielsinger @amit_zhr @YKirstain @adam_polyak90 Yaniv Taigman @liorwolf and @ShellySheynin
Check out the project page for many more results and details: hila-chefer.github.io/videojam-paperβ¦
Meta presents Video Editing via Factorized Diffusion Distillation
We introduce Emu Video Edit (EVE), a model that establishes a new state-of-the art in video editing without relying on any supervised video editing data. To develop EVE we separately train an image editingView quoted post on X
