Generative Multimodal Models are In-Context Learners

arXiv:2312.13286v2 Announce Type: replace
Abstract: The human ability to easily solve multimodal tasks in context (i.e., with only a few demonstrations or simple instructions), is what current multimodal systems have largely struggled to imitate. In this work, we demonstrate that the task-agnostic in-context learning capabilities of large multimodal models can be significantly enhanced by effective scaling-up. We introduce Emu2, a generative multimodal model with 37 billion parameters, trained on large-scale multimodal sequences with a unified autoregressive objective. Emu2 exhibits strong …

abstract arxiv capabilities context context learning cs.cv current generative human in-context learning large multimodal models multimodal multimodal models multimodal systems simple solve systems tasks type work