6 datasets found

The QUBIQ dataset is a multi-rater, multi-center, multi-modality dataset that includes both 2D and 3D segmentation tasks. It features various modalities such as MRI and CT;...

The dataset used in the paper is the Pick-a-Pic dataset, which consists of 87,687 pairs of text prompts and images.

The dataset used in the paper is the PartiPrompts dataset, which consists of 851,293 pairs of text prompts and images.

Human preference data is collected for reward modeling.

Online human behavior data is collected from Baidu Search. Human preference data is collected for reward modeling.

The dataset is used for human alignment in large language models.