The last several years have witnessed remarkable progress in video-and-language (VidL) understanding. However, most modern VidL approaches use complex and specialized model architectures and sophisticated pretraining protocols, making the reproducibility, analysis and comparisons of these frameworks difficult. Hence, instead of proposing yet another new VidL model, this paper conducts a thorough empirical study demystifying the most important factors in the VidL model design. Among the factors that we investigate are (i) the spatiotemporal architecture design, (ii) the multimodal fusion schemes, (iii) the pretraining objectives, (iv) the choice of pretraining data, (v) pretraining and finetuning protocols, and (vi) dataset and model scaling. Our empirical study reveals that the most important design factors include: temporal modeling, video-to-text multimodal fusion, masked modeling objectives, and joint training on images and videos. Using these empirical insights, we then develop a step-by-step recipe, dubbed VindLU, for effective VidL pretraining. Our final model trained using our recipe achieves comparable or better than state-of-the-art results on several VidL tasks without relying on external CLIP pretraining. In particular, on the text-to-video retrieval task, our approach obtains 61.2% on DiDeMo, and 55.0% on ActivityNet, outperforming current SOTA by 7.8% and 6.1% respectively. Furthermore, our model also obtains state-of-the-art video question-answering results on ActivityNet-QA, MSRVTT-QA, MSRVTT-MC and TVQA. Our code and pretrained models are publicly available at: https://github.com/klauscc/VindLU.

翻译：过去数年间，视频与语言理解领域取得了显著进展。然而，大多数现代视频与语言方法采用复杂且专用的模型架构以及精密的预训练协议，导致这些框架的可复现性、分析和比较变得困难。因此，本文不再提出新的视频与语言模型，而是通过彻底的实证研究来揭示视频与语言模型设计中最关键的因素。我们研究的因素包括：(i) 时空架构设计，(ii) 多模态融合方案，(iii) 预训练目标，(iv) 预训练数据的选择，(v) 预训练与微调协议，以及 (vi) 数据集与模型规模化。实证研究表明，最重要的设计因素包括：时序建模、视频到文本多模态融合、掩码建模目标以及图像与视频联合训练。基于这些实证洞见，我们逐步提出了一套名为 VindLU 的指南，用于高效视频与语言预训练。采用该指南训练的最终模型，在不依赖外部CLIP预训练的情况下，在多个视频与语言任务上取得了与当前最优水平相当或更优的结果。具体而言，在文本到视频检索任务中，我们的方法在DiDeMo上达到61.2%，在ActivityNet上达到55.0%，分别超越当前最优方法7.8%和6.1%。此外，模型在ActivityNet-QA、MSRVTT-QA、MSRVTT-MC和TVQA的视频问答任务上也取得了最优结果。我们的代码与预训练模型已公开于：https://github.com/klauscc/VindLU。