Run on Different Dataset

[MayarOssama]

Hello, Thanks for sharing your work.
How can I run your approach on different datasets?
TIA

[tanyuqian]

Hi! Thanks for noticing our work and sorry for late reply.

To run your dataset, you only need to prepare a json file with "init_prompts" and "seed_ent_tuples" (the examples can be found in relation_info/ dir) and run search_prompts.py (will add "prompts" in your json file) and main.py (search for knowledge tuples) with --rel_set as your customized name (detailed commands can be found in README).

Please let us know if you run into any errors. BTW, we are preparing a demo website lmnet.io (still under development). Feel free to check it out!

[puraminy]

where are data used to evaluate your model? I mean the data and the true answer, in the paper you evaluated it.