search

lienwc / DKGE

This is the source code of DKGE.
26 stars 11 forks source link

Regarding results of the cores #1

Open YuyangWei opened 4 years ago

YuyangWei commented 4 years ago

Hi! Thanks for open sourcing your code. But, I have a question after running your source core. The results of your train.py are three files including all_parameters, entity_o and relation_o. In the entity_o, the content is “19672: [NaN, NaN, NaN, NaN,…,NaN]”. In general, every dimension of embedding is float or double instead of NaN. Does the result accord with your expectations? result of train In addition, I am also very interested in your comparison algorithm puTransE. but the author of puTransE does not release source codes, and I wondering if you could kindly send me the source program of puTransE, datasets and the necessary information about it. I promise they will be used only for research purposed. Thank you very much for your kind consideration and I am looking forward to your early reply.

fangyiming commented 3 years ago

Hello, do you know his running environment and running commands? If you do, please tell me. Thank you

fangyiming commented 3 years ago

Hi! Thanks for open sourcing your code. But, I have a question after running your source core. The results of your train.py are three files including all_parameters, entity_o and relation_o. In the entity_o, the content is “19672: [NaN, NaN, NaN, NaN,…,NaN]”. In general, every dimension of embedding is float or double instead of NaN. Does the result accord with your expectations? result of train In addition, I am also very interested in your comparison algorithm puTransE. but the author of puTransE does not release source codes, and I wondering if you could kindly send me the source program of puTransE, datasets and the necessary information about it. I promise they will be used only for research purposed. Thank you very much for your kind consideration and I am looking forward to your early reply.

Hello, do you know his running environment and running commands? If you do, please tell me. Thank you