deshanksuman/WSD_DATASET_FEWS
收藏Hugging Face2025-03-21 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/deshanksuman/WSD_DATASET_FEWS
下载链接
链接失效反馈官方服务：
资源简介：
FEWS词义消歧数据集是一个经过预处理和格式化的数据集，旨在直接用于训练和微调用于词义消歧的语言模型。数据集中的每个上下文中的模糊词都被`<WSD>`标签包围，以便模型在训练和推理过程中专注于特定的模糊词。数据集按照alpaca_prompt的格式组织，包括指令、输入和输出部分。该数据集适用于词义消歧任务的模型微调、评估词义消歧性能以及跨语言语义消歧的研究。
The FEWS Dataset for Word Sense Disambiguation is a preprocessed and formatted dataset intended for direct use in training and fine-tuning language models for word sense disambiguation tasks. Each ambiguous word in the context is enclosed with `<WSD>` tags to focus the model on specific words during training and inference. The dataset is organized according to the alpaca_prompt format, which includes Instruction, Input, and Output sections. It is suitable for fine-tuning models for WSD tasks, evaluating WSD performance, and research on cross-lingual semantic disambiguation.
提供机构：
deshanksuman
