COSMO: COntrastive Streamlined MultimOdal Model with Interleaved Pre-Training
This work addresses the limited availability of high-quality long-text video datasets for vision-language models, though it is incremental in combining existing techniques like contrastive loss with model partitioning.
The paper tackles the challenge of aligning autoregressive vision-language models for extended textual contexts by introducing COSMO, a framework that integrates contrastive loss and partitions the language model into unimodal and multimodal components, achieving a performance improvement from 57.2% to 65% in a 4-shot flickr captioning task with 34% learnable parameters.
In the evolution of Vision-Language Pre-training, shifting from short-text comprehension to encompassing extended textual contexts is pivotal. Recent autoregressive vision-language models like \cite{flamingo, palme}, leveraging the long-context capability of Large Language Models, have excelled in few-shot text generation tasks but face challenges in alignment tasks. Addressing this gap, we introduce the contrastive loss into text generation models, presenting the COntrastive-Streamlined MultimOdal framework (\ModelName), strategically partitioning the language model into dedicated unimodal text processing and adept multimodal data handling components. \ModelName, our unified framework, merges unimodal and multimodal elements, enhancing model performance for tasks involving textual and visual data while notably reducing learnable parameters. However, these models demand extensive long-text datasets, yet the availability of high-quality long-text video datasets remains limited. To bridge this gap, this work introduces \VideoDatasetName, an inaugural interleaved video-text dataset featuring comprehensive captions, marking a significant step forward. Demonstrating its impact, we illustrate how \VideoDatasetName{} enhances model performance in image-text tasks. With 34% learnable parameters and utilizing 72\% of the available data, our model demonstrates significant superiority over OpenFlamingo~\cite{openflamingo}. For instance, in the 4-shot flickr captioning task, performance notably improves from 57.2% to 65.\%. The contributions of \ModelName{} and \VideoDatasetName{} are underscored by notable performance gains across 14 diverse downstream datasets encompassing both image-text and video-text tasks.