malllabiisc/CompGCN

Issues

• Code for graph classification task

  #29 opened 3 years ago by Kevin0726
  2

• How to get trained embedding matrix

  #27 opened 3 years ago by SRL94
  3

• A problem about replace head entity

  #43 opened a year ago by linxiang1911
  0

• head-batch and tail-batch? In line 293 of the run.py file.

  #42 opened a year ago by 5zongyuan
  0

• ValueError: Unable to configure handler 'file_handler'

  #36 opened 2 years ago by HJX-zhanS
  3

• OSError:libtorch_cpu.so:cannot open shared object file : No such file or diectory

  #38 opened 2 years ago by VachelChen
  1

• Who can reproduce the reported results on the dataset WN18RR?

  #39 opened 2 years ago by pp413
  0

• Just a suggestion on Fourier transformation APIs for users whose pytorch version >= 1.8.

  #35 opened 2 years ago by nxznm
  0

• Which evaluation protocol do you use

  #34 opened 3 years ago by jwzhi
  0

• How do you calculate mrr by relation category

  #33 opened 3 years ago by jwzhi
  0

• Random adjacency matrices do not affect the performance

  #32 opened 3 years ago by zhanqiuzhang
  1

• About why you pre-calculate all node representations

  #15 opened 4 years ago by Punchwes
  4

• What's the parameters for WN18RR？

  #31 opened 3 years ago by sherluck314
  1

• Problem for running

  #30 opened 3 years ago by MangoSue
  1

• Code for graph classification

  #11 opened 4 years ago by freyaya123
  1

• What are the best superparameters for WN18RR?

  #24 opened 3 years ago by diligentlee
  5

• What is the label in run_epoch

  #28 opened 3 years ago by SRL94
  1

• confused about distmult

  #25 opened 3 years ago by junkangwu
  1

• Regarding number of GCN updates

  #26 opened 3 years ago by Megh-Thakkar
  4

• efficiency about batch training

  #21 opened 3 years ago by junkangwu
  4

• About cuda version

  #23 opened 3 years ago by yunzi-94
  1

• The whole graph of the adjacency matrix created with the training set cannot contain some entities that have not appeared in the training set.

  #22 opened 4 years ago by WeidongLi-KG
  1

• Why performance on head data hugely worse than tail data?

  #20 opened 4 years ago by renli1024
  1

• Code for entity classification task

  #2 opened 4 years ago by PlusRoss
  6

• Why do you use outgoing edges as neighbor?

  #19 opened 4 years ago by renli1024
  1

• Small details in Transe

  #18 opened 4 years ago by JiyaoWei
  1

• What does gamma (margin) mean in your hyper parameter?

  #17 opened 4 years ago by ZifengDing
  2

• 3 questions or suggestions

  #16 opened 4 years ago by PolarisRisingWar
  1

• The scripts to reproduce the results.

  #12 opened 4 years ago by jianyucai
  1

• CompGCN with TransE decoder does not converge

  #9 opened 4 years ago by migalkin
  3

• Data leakage?

  #14 opened 4 years ago by guolingbing
  1

• What about the lowest GPU needed?

  #13 opened 4 years ago by leiloong
  1

• Model efficiency

  #5 opened 4 years ago by zhifei1993
  1

• Some questions about the results of W-GCN represented in the paper

  #8 opened 4 years ago by jianyucai
  1

• Just a small suggestion.

  #10 opened 4 years ago by seanswyi
  1

• Is 1.1.2 the correct version for `torch-scatter`?

  #7 opened 4 years ago by seanswyi
  1

• Run with relation basis vectors

  #6 opened 4 years ago by gsp2014
  2

• I got an error when i run the code. What is the "run_gcn"?

  #1 opened 4 years ago by 00GitHub
  1