FlowBind enables efficient any-to-any multimodal generation via a shared latent space bridged by modality-specific invertible flows, matching prior quality with up to 6x fewer parameters and 10x faster training.
In this setup, a powerful large language model performs cross-modal sequence generation, with tokenized data of all modalities
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
FlowBind: Efficient Any-to-Any Generation with Bidirectional Flows
FlowBind enables efficient any-to-any multimodal generation via a shared latent space bridged by modality-specific invertible flows, matching prior quality with up to 6x fewer parameters and 10x faster training.