You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank you for your work. Besides WN18, you use FB15K for experimental evaluations. Can you release FB15K.bin and detail parameters to train the model for FB15K.
The text was updated successfully, but these errors were encountered:
The dataset used in this repo seems to be the same that A.Bordes used in his paper talking about TransE: https://everest.hds.utc.fr/doku.php?id=en:transe
To generate the bin, you have to parse the original dataset to get the same structure as used on WN18.bin. Check out Python Pickle to read and save this binary files.
Hi,
Thanks for such a nice code for the method.
I was able to reproduce the results for WN18 and was trying to reproduce the results for FB15k dataset.
Will it be possible to share the parameters and other settings for FB15k as well?
With settings similar to WN18, I wasn't not getting the same numbers.
Hello,
Thank you for sharing your code here.
I cannot get the expected result for FB15K dataset with the parameters similar to WN18. Could you share your parameter settings for FB15K?
Thank you for your work. Besides WN18, you use FB15K for experimental evaluations. Can you release FB15K.bin and detail parameters to train the model for FB15K.
The text was updated successfully, but these errors were encountered: