We release our dataset to ensure that everyone can replicate our experimental conclusions.
Directory Description
|-- dataset generate / method(prompts) to generate data
--|-- dataset / data resource
|-- llava training / training code
|-- tuning script / tuing parameters
