Mixture of Transformers

Software / App

An architecture that uses independent sets of transformer parameters for each modality (text, image, audio) to improve efficiency and generation quality.

Mentioned in 1 video

Videos Mentioning Mixture of Transformers

Stanford CS25: Transformers United V6 I From Language Models to Native Multimodal Intelligence

Stanford CS25: Transformers United V6 I From Language Models to Native Multimodal Intelligence

Stanford Online

An architecture that uses independent sets of transformer parameters for each modality (text, image, audio) to improve efficiency and generation quality.