axolotl-ai-co/llava-instruct-mix-vsft-small

这是一个精简版本的`HuggingFaceH4/llava-instruct-mix-vsft`数据集，包含原数据集1%的训练和测试数据。数据集包含两个主要特征：messages和images。messages特征是一个列表，包含content和role两个子特征，其中content又是一个包含index、text和type的列表。images特征是一个图像序列。数据集分为训练集和测试集，训练集包含2592个样本，测试集包含136个样本。

A slim version of `HuggingFaceH4/llava-instruct-mix-vsft`. Contains 1% of train and test split. The dataset includes two main features: messages and images. The messages feature is a list containing two sub-features: content and role, where content is a list containing index, text, and type. The images feature is a sequence of images. The dataset is divided into a training set with 2592 examples and a test set with 136 examples.