trl-lib/llava-instruct-mix

LLaVA Instruct Mix数据集是一个对话式语言建模数据集，包含了与文本相关的图片、提示信息以及模型响应的完成信息。数据集的结构设计使模型能够从对话上下文中学习，从而更好地根据视觉输入生成描述性文本。

The LLaVA Instruct Mix dataset is a conversational language-modeling dataset that includes images associated with the text, prompt messages, and the models response completions. The structure of the dataset is designed to enable models to learn from the context of conversations, enhancing their ability to generate descriptive text based on visual inputs.