Multimodal Flow: Unified Flow Modeling of Language and Vision in Embedding Spaces

We present Multimodal Flow, a fully continuous generative model of language and vision. Most unified multimodal models either model both language and …

Continue in Briflio to read, react, comment, and share.