search

migalkin / StarE

EMNLP 2020: Message Passing for Hyper-Relational Knowledge Graphs
MIT License
85 stars 17 forks source link

How to activate the ATTENTION mechanism? #4

Closed bigdatae closed 3 years ago

bigdatae commented 3 years ago

Hi,

I'd like to calculate the value of attention using 'jf17k', but when I change STAREARGS in run.py to 'ATTENTION': True, there is a code error

RuntimeError: The expanded size of the tensor (4) must match the existing size (61400) at non-singleton dimension 1. Target sizes: [61400, 4, 50]. Tensor sizes: [1, 61400, 1]

I cannt find the error in the codes.