CoCa: Contrastive Captioners are Image-Text Foundation Models
CoCa：对比描述器是图像文本基础模型.
VinVL: Revisiting Visual Representations in Vision-Language Models
VinVL：重新回归视觉语言模型中的视觉表示.
SimVLM: Simple Visual Language Model Pretraining with Weak Supervision
SimVLM：弱监督的简单视觉语言模型预训练.
GIT: A Generative Image-to-text Transformer for Vision and Language
GIT：视觉和语言的通用图像到文本Transformer.
VLMo: Unified Vision-Language Pre-Training with Mixture-of-Modality-Experts
VLMo：使用模态混合专家的统一视觉语言预训练.
Align before Fuse: Vision and Language Representation Learning with Momentum Distillation
融合前对齐：使用动量蒸馏进行视觉和语言表示学习.