emoji prediction task. With respect to the combination of textual and visual modalities, we adopt a middle fusion ap-.
確定! 回上一頁