jiaweisheng / faan Goto Github PK

i see you trained the node embedding by torch.Embedding() rather than from relation2vec.ComplEx
and i find that the relation in train_task.json is not contained in relation2id or relation2vec
it seem that the relation in relation2id, relation2vec and pre_trained embedding are same, but different with relations in train_task.json,
so my question is, is the pre-trained embedding used in this papper?
Anticipates your answer very much, thank you.

I got a CUDA error when I run trainer.py, anyone can help me figure it out?

Hi, I got this error:

Traceback (most recent call last):
File "trainer.py", line 434, in
trainer.train()
File "trainer.py", line 226, in train
nn.utils.clip_grad_norm(self.parameters, self.grad_clip)
File "/python3.6/site-packages/torch/nn/utils/clip_grad.py", line 47, in clip_grad_norm
return clip_grad_norm_(parameters, max_norm, norm_type)
File "/python3.6/site-packages/torch/nn/utils/clip_grad.py", line 30, in clip_grad_norm_
total_norm = torch.norm(torch.stack([torch.norm(p.grad.detach(), norm_type) for p in parameters]), norm_type)
RuntimeError: stack expects a non-empty TensorList

According to the code you provided, when training FAAN, you set the embedding dimension of NELL in the parameters to 100 and the embedding dimension of Wiki to 50, while the opposite is true in the paper.
"The embedding dimensionality is set to 50 and 100 for NELL and Wiki, respectively."

Hello, Jiawei

Can you help me understand the method of limiting the candidate entities, in "One-Shot Relational Learning for Knowledge Graphs". Many people cited your paper, but they have not given a specific parameters.

How did you make specific restrictions in your experiment? Can you show your code for generating the candidate set?

Looking forward to your reply!

When I run this code on several different servers, I find that the all results of Wiki dataset are about 3% lower than the result in the paper, but the result of NELL dataset is correct, and the random seed of both are 19950922 as the code. So I want know what's the problem. Look forward to your reply.