Visual instruction tuning with polite flamingo

D Chen, J Liu, W Dai, B Wang - … of the AAAI Conference on Artificial …, 2024 - ojs.aaai.org
Proceedings of the AAAI Conference on Artificial Intelligence, 2024ojs.aaai.org
Recent research has demonstrated that the multi-task fine-tuning of multi-modal Large
Language Models (LLMs) using an assortment of annotated downstream vision-language
datasets significantly enhances their performance. Yet, during this process, a side effect,
which we termed as the" multi-modal alignment tax", surfaces. This side effect negatively
impacts the model's ability to format responses appropriately-for instance, its" politeness"-
due to the overly succinct and unformatted nature of raw annotations, resulting in reduced …
Abstract
Recent research has demonstrated that the multi-task fine-tuning of multi-modal Large Language Models (LLMs) using an assortment of annotated downstream vision-language datasets significantly enhances their performance. Yet, during this process, a side effect, which we termed as the" multi-modal alignment tax", surfaces. This side effect negatively impacts the model's ability to format responses appropriately-for instance, its" politeness"-due to the overly succinct and unformatted nature of raw annotations, resulting in reduced human preference. In this paper, we introduce Polite Flamingo, a multi-modal response rewriter that transforms raw annotations into a more appealing," polite" format. Polite Flamingo is trained to reconstruct high-quality responses from their automatically distorted counterparts and is subsequently applied to a vast array of vision-language datasets for response rewriting. After rigorous filtering, we generate the PF-1M dataset and further validate its value by fine-tuning a multi-modal LLM with it. Combined with novel methodologies including U-shaped multi-stage tuning and multi-turn augmentation, the resulting model, Clever Flamingo, demonstrates its advantages in both multi-modal understanding and response politeness according to automated and human evaluations. Code and dataset are available at https://github. com/ChenDelong1999/polite-flamingo
ojs.aaai.org
以上显示的是最相近的搜索结果。 查看全部搜索结果