Kandinsky Video — a new text-to-video generation model

SoTA quality among open-source solutions

Kandinsky Video is a text-to-video generation model, which is based on the FusionFrames architecture, consisting of two main stages: keyframe generation and interpolation. Our approach for temporal conditioning allows us to generate videos with high-quality appearance, smoothness and dynamics.

Pipeline

The encoded text prompt enters the U-Net keyframe generation model with temporal layers or blocks, and then the sampled latent keyframes are sent to the latent interpolation model in such a way as to predict three interpolation frames between two keyframes. A temporal MoVQ-GAN decoder is used to get the final video result.

Architecture details

• Text encoder (Flan-UL2) - 8.6B
• Latent Diffusion U-Net3D - 4.0B
• MoVQ encoder/decoder - 256M

BibTeX

If you use our work in your research, please cite our publication:

@article{arkhipkin2023fusionframes,
  title     = {FusionFrames: Efficient Architectural Aspects for Text-to-Video Generation Pipeline},
  author    = {Arkhipkin, Vladimir and Shaheen, Zein and Vasilev, Viacheslav and Dakhova, Elizaveta and Kuznetsov, Andrey and Dimitrov, Denis},
  journal   = {arXiv preprint arXiv:2311.13073},
  year      = {2023}, 
}