Emu: Generative Pretraining in Multimodality
This work addresses the need for versatile multimodal AI systems that can handle diverse data types, though it appears incremental as it builds on existing Transformer and multimodal pretraining paradigms.
The authors tackled the problem of multimodal generation by introducing Emu, a Transformer-based foundation model that can generate images and texts from any single-modality or multimodal inputs, achieving superb performance in zero-shot/few-shot tasks like image captioning and text-to-image generation compared to state-of-the-art models.
We present Emu, a Transformer-based multimodal foundation model, which can seamlessly generate images and texts in multimodal context. This omnivore model can take in any single-modality or multimodal data input indiscriminately (e.g., interleaved image, text and video) through a one-model-for-all autoregressive training process. First, visual signals are encoded into embeddings, and together with text tokens form an interleaved input sequence. Emu is then end-to-end trained with a unified objective of classifying the next text token or regressing the next visual embedding in the multimodal sequence. This versatile multimodality empowers the exploration of diverse pretraining data sources at scale, such as videos with interleaved frames and text, webpages with interleaved images and text, as well as web-scale image-text pairs and video-text pairs. Emu can serve as a generalist multimodal interface for both image-to-text and text-to-image tasks, and supports in-context image and text generation. Across a broad range of zero-shot/few-shot tasks including image captioning, visual question answering, video question answering and text-to-image generation, Emu demonstrates superb performance compared to state-of-the-art large multimodal models. Extended capabilities such as multimodal assistants via instruction tuning are also demonstrated with impressive performance.