CLIPPO: Image-and-Language Understanding from Pixels Only
This work addresses the need for more unified and efficient multimodal models, offering a novel approach that eliminates modality-specific components, which is incremental but impactful for simplifying AI systems.
The paper tackles the problem of multimodal models requiring separate components for different modalities by introducing CLIPPO, a single pixel-based encoder that processes both images and text rendered as images, achieving competitive performance in tasks like image retrieval and zero-shot classification with half the parameters of CLIP-style models, and outperforming prior work in natural language understanding and multilingual retrieval.
Multimodal models are becoming increasingly effective, in part due to unified components, such as the Transformer architecture. However, multimodal models still often consist of many task- and modality-specific pieces and training procedures. For example, CLIP (Radford et al., 2021) trains independent text and image towers via a contrastive loss. We explore an additional unification: the use of a pure pixel-based model to perform image, text, and multimodal tasks. Our model is trained with contrastive loss alone, so we call it CLIP-Pixels Only (CLIPPO). CLIPPO uses a single encoder that processes both regular images and text rendered as images. CLIPPO performs image-based tasks such as retrieval and zero-shot image classification almost as well as CLIP-style models, with half the number of parameters and no text-specific tower or embedding. When trained jointly via image-text contrastive learning and next-sentence contrastive learning, CLIPPO can perform well on natural language understanding tasks, without any word-level loss (language modelling or masked language modelling), outperforming pixel-based prior work. Surprisingly, CLIPPO can obtain good accuracy in visual question answering, simply by rendering the question and image together. Finally, we exploit the fact that CLIPPO does not require a tokenizer to show that it can achieve strong performance on multilingual multimodal retrieval without modifications.