# Transflower: probabilistic autoregressive dance generation with multimodal attention

Source: [Transflower: probabilistic autoregressive dance generation with multimodal attention](<https://www.sota2.com/research/paper/68fa6cc11e209131cea35d0d>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T00:18:38.162Z

Authors: Guillermo Valle-P\\'erez, Gustav Eje Henter, Jonas Beskow, Andr\\'e Holzapfel, Pierre-Yves Oudeyer, Simon Alexanderson

Published: 2021-06-25T20:14:28.000Z

## Abstract

Dance requires skillful composition of complex movements that follow rhythmic, tonal and timbral features of music. Formally, generating dance conditioned on a piece of music can be expressed as a problem of modelling a high-dimensional continuous motion signal, conditioned on an audio signal. In this work we make two contributions to tackle this problem. First, we present a novel probabilistic autoregressive architecture that models the distribution over future poses with a normalizing flow conditioned on previous poses as well as music context, using a multimodal transformer encoder. Second, we introduce the currently largest 3D dance-motion dataset, obtained with a variety of motion-capture technologies, and including both professional and casual dancers. Using this dataset, we compare our new model against two baselines, via objective metrics and a user study, and show that both the ability to model a probability distribution, as well as being able to attend over a large motion and music context are necessary to produce interesting, diverse, and realistic dance that matches the music.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Group Dance Generation on AIOZ-GDANCE 1.0 (test)](<https://www.sota2.com/research/sota/group-dance-generation-on-aioz-gdance-1-0-test.md>): Group Dance Generation; AIOZ-GDANCE 1.0 (test); FID: 37.73.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2106.13871>)

[PDF](<https://arxiv.org/pdf/2106.13871.pdf>)