search

mayhugotong / GenTKG

This is the official impletation repository of NAACL findings paper, GenTKG: Generative Forecasting on Temporal Knowledge Graph with Large Language Models. (https://arxiv.org/abs/2310.07793).
32 stars 5 forks source link

Update links for 'data' and 'model' on GDrive #6

Open komarovala opened 2 months ago

komarovala commented 2 months ago

I got a lot of errors related to the data structure and model, here is one of them: photo_2024-09-22_10-58-34

I downloaded the structure from this link There are no files on the links for downloading via gdown. Could you update the links to Google Drive?

mayhugotong commented 2 months ago

Hi the link is valid. Can you please check your gdown usage?

komarovala commented 2 months ago

Certainly:

gdowm

komarovala commented 2 months ago

Plus if go by link photo_2024-09-22_19-51-05

mayhugotong commented 1 month ago

Hi komarovala,

Our link is public and valid. Can you please try this: https://drive.google.com/drive/folders/1pZwppUnjLAfdzu30lKaxZGM3AC-x5We-?usp=sharing

IonutIga commented 2 weeks ago

Hello, I was trying to run the code as presented on the home page for icews14, but I get into the same issue as stated here in the beginning. After inspecting the code, I do understand that in the _data_utils/ruleslearning/grapher.py file, _createstore function will create a file similar to what is already provided in the drive/data/original/icews14/train.txt, where each quad is written using their members' corresponding ID numbers. In other words, the train and test files provided for each original dataset is expected, by the create_store function, to be in the lexical format and not in the id one, but is not the case here. Can it be the reason why the above index error is received? If so, could you please provide the lexical form of the original datasets? Thank you.