Tag
This paper evaluates state-of-the-art tabular models as encoders in image-tabular multimodal learning, addressing the challenge of using In-Context Learning models that require labels for embedding training and test instances.