Few-Shot NER Dataset

该数据集是最大的少样本命名实体识别数据集，包含了66种细粒度实体类型，这些类型分布在8个粗粒度类别中。该数据集旨在支持两项任务：内部任务和跨任务，通过使用不同的源数据集和目标数据集来评估跨领域的适应能力。其规模属于大型，专注于少样本命名实体识别任务。

This is the largest few-shot named entity recognition (NER) dataset, encompassing 66 fine-grained entity types categorized into 8 coarse-grained classes. It is developed to support two tasks: intra-task and cross-task, which evaluate cross-domain adaptation capabilities by leveraging distinct source and target datasets. Boasting a large scale, this dataset specifically targets few-shot NER tasks.