thank you for great works [1,3,4] among many others. I was wondering whether one could find the best performing hyper parameters for Hyper along with the number of parameters of it.
I reckon that it would be beneficial for everyone if a table reporting best performing hyper parameters and respective number of model parameters --including Hyper, MuRP and TuckER were available.
Hi Caglar, really sorry for the late response, I must have somehow missed this. Unfortunately I haven't had time lately to do this, but I will do it if I find some time.
Dear Ivana,
thank you for great works [1,3,4] among many others. I was wondering whether one could find the best performing hyper parameters for Hyper along with the number of parameters of it.
I reckon that it would be beneficial for everyone if a table reporting best performing hyper parameters and respective number of model parameters --including Hyper, MuRP and TuckER were available.
Cheers,
[1] Hypernetwork Knowledge Graph Embeddings [2] Multi-relational Poincaré graph embeddings [3] TuckER: Tensor Factorization for Knowledge Graph Completion