FB15k-237-subset
收藏资源简介:
这是一个用于知识图谱补全的子集数据集,特别设计用于测试模型在不同场景(如传导性、归纳性和少样本学习)下的性能。
This is a subset dataset for knowledge graph completion, which is specifically designed to evaluate model performance across various scenarios including transductive, inductive, and few-shot learning.
CATS: Context-aware Inductive Knowledge Graph Completion with Latent Type Constraints and Subgraph Reasoning
数据集
-
数据集和LLM指令可以从以下链接下载:
-
将两个子文件夹 "datasets" 和 "instructions" 复制到项目目录中。
-
或者,可以通过执行
python build_instruction.py来构建LLM指令提示。
LLM设置
可以从以下链接下载LLM检查点:
请更新脚本 data_manager.py 中的 LLM_PATH 默认值为本地模型路径。
引用
如果发现此代码有用,请考虑引用以下论文:
@misc{li2024contextawareinductiveknowledgegraph, title={Context-aware Inductive Knowledge Graph Completion with Latent Type Constraints and Subgraph Reasoning}, author={Muzhi Li and Cehao Yang and Chengjin Xu and Zixing Song and Xuhui Jiang and Jian Guo and Ho-fung Leung and Irwin King}, year={2024}, eprint={2410.16803}, archivePrefix={arXiv}, primaryClass={cs.AI}, url={https://arxiv.org/abs/2410.16803}, }




