遇见数据集

aisingapore/nlu-metaphor

收藏
Hugging Face2024-12-20 更新2024-12-21 收录
官方服务:

资源简介:

SEA Metaphor数据集用于评估模型在解释具有不同含义的比喻短语方面的能力。该数据集从Multilingual-Fig-QA中采样,包含印度尼西亚语、爪哇语和巽他语。数据集按语言划分,并提供了fewshot示例。支持的任务包括文本生成和问答。数据来源于Multilingual-Fig-QA,许可证为MIT。

The SEA Metaphor dataset evaluates a models ability to interpret paired figurative phrases with divergent meanings. It is sampled from Multilingual-Fig-QA for Indonesian, Javanese, and Sundanese. The dataset is split by language, with additional splits containing fewshot examples. Supported tasks include text generation and question answering. The data source is Multilingual-Fig-QA, and the license is MIT.

提供机构:
aisingapore
二维码
社区交流群
二维码
科研交流群
商业服务