README
Intro
Sky-NER is a set of GPT-generated data for Chinese named entity recognition. Sky-NER is constructed by following the recipe in UniversalNER and based on the SkyPile-Corpus. It was collected by prompting gpt-3.5-turbo-0125 and augmented by negative sampling.
For more detailed data preprocessing steps, please refer to our paper and github repo.
The data collection prompt is as follows:
给定一段文本，你的任务是抽取所有实体并识别它们的实体类别。输出应为以下JSON格式：[{"实体1": "实体1的类别"}, ...]。
License
Attribution-NonCommercial 4.0 International
Similar repos
README
Intro
Sky-NER is a set of GPT-generated data for Chinese named entity recognition. Sky-NER is constructed by following the recipe in UniversalNER and based on the SkyPile-Corpus. It was collected by prompting gpt-3.5-turbo-0125 and augmented by negative sampling.
For more detailed data preprocessing steps, please refer to our paper and github repo.
The data collection prompt is as follows:
给定一段文本，你的任务是抽取所有实体并识别它们的实体类别。输出应为以下JSON格式：[{"实体1": "实体1的类别"}, ...]。
License
Attribution-NonCommercial 4.0 International