GenBench

GenBench是由未来产业研究中心AI实验室开发的综合性基准数据集，专注于评估基因组基础模型的效能。该数据集包含43个真实世界的数据集，覆盖从局部到全局的基因组任务，特别关注编码区域、非编码区域和基因组结构。GenBench通过系统评估，旨在标准化基因组模型的评估，并推动该领域的进一步发展。数据集的应用领域广泛，包括基因定位、功能预测、调控元件识别等，旨在解决基因组数据处理中的长序列处理、训练策略混乱和模型设计不明确等问题。

GenBench is a comprehensive benchmark dataset developed by the AI Lab of the Future Industry Research Center, focusing on evaluating the performance of genomic foundation models. This dataset encompasses 43 real-world datasets, covering genomic tasks ranging from local to global scales, with particular emphasis on coding regions, non-coding regions, and genomic structures. GenBench aims to standardize the evaluation of genomic models through systematic assessments and promote further progress in this field. The dataset has a wide range of application scenarios, including gene localization, function prediction, regulatory element recognition and more, and is designed to address key issues in genomic data processing such as long sequence processing, inconsistent training strategies, and ambiguous model design.