danielegrattarola/keras-gat issues and pull requests

#36 - Why is my accuracy consistently around 0.2 in training

Issue - State: open - Opened by hnulmz3 about 1 year ago - 1 comment

#35 - Input shape

Issue - State: closed - Opened by Kareshi-X over 2 years ago - 2 comments

#34 - How to fix this issues

Issue - State: open - Opened by weber-tsao almost 3 years ago - 2 comments

#33 - 遇到这个问题怎么解决啊？？？求大佬解答！！

Issue - State: open - Opened by cwl1999 almost 3 years ago - 1 comment

#32 - How to use predict func after training model ?

Issue - State: open - Opened by monkey1302 almost 3 years ago - 2 comments

#31 - Question about the image segmentation (use GAT)

Issue - State: open - Opened by tanjia123456 almost 4 years ago

#30 - Performance in Cora

Issue - State: closed - Opened by JhuoW almost 4 years ago - 2 comments

#29 - Visualization Qesition

Issue - State: closed - Opened by tanjia123456 almost 4 years ago - 2 comments

#28 - Why I got loss value nan in working with citeseer data

Issue - State: closed - Opened by gaudelbijay about 4 years ago - 7 comments

#27 - matrix A

Issue - State: open - Opened by TransformersWsz over 4 years ago - 2 comments

#26 - Cannot approach the accuracy ~83% on cora

Issue - State: closed - Opened by stridence over 4 years ago - 2 comments

#25 - batch dimension

Issue - State: closed - Opened by xhtian95 about 5 years ago - 1 comment

#24 - random seed

Issue - State: open - Opened by xhtian95 about 5 years ago - 1 comment

#23 - Getting NaN loss

Issue - State: closed - Opened by duncster94 over 5 years ago - 5 comments

#22 - Only 9700 parameters and 13700 samples result in OOM on a TITAN Xp

Issue - State: closed - Opened by Imorton-zd over 5 years ago - 1 comment

#21 - self attention issue

Issue - State: closed - Opened by FrankCAN almost 6 years ago - 1 comment

#20 - there's actually no need to do the activation early in case of "conca…

Pull Request - State: closed - Opened by mawright almost 6 years ago - 1 comment

#19 - Adjust l2 norm penalty in example

Pull Request - State: closed - Opened by mawright almost 6 years ago - 1 comment

#18 - PPI

Issue - State: closed - Opened by xuhaiyun43 almost 6 years ago - 1 comment

#18 - PPI

Issue - State: closed - Opened by xuhaiyun43 almost 6 years ago - 1 comment

#17 - The example script still does not reproduce the results

Issue - State: closed - Opened by randomboolean about 6 years ago - 6 comments

#16 - How to train the model through minibatch of samples

Issue - State: closed - Opened by qzshadow about 6 years ago - 3 comments

#16 - How to train the model through minibatch of samples

Issue - State: closed - Opened by qzshadow about 6 years ago - 3 comments

#15 - How to perform inductive learning on PPI?

Issue - State: closed - Opened by Jhy1993 about 6 years ago - 3 comments

#14 - citeseer dataset loss is nan

Issue - State: closed - Opened by CEfanmin about 6 years ago - 3 comments

#14 - citeseer dataset loss is nan

Issue - State: closed - Opened by CEfanmin about 6 years ago - 3 comments

#13 - Comparison with original GCN

Issue - State: closed - Opened by wzm2256 over 6 years ago - 1 comment

#12 - plot the first GAT layer

Issue - State: closed - Opened by CEfanmin over 6 years ago - 1 comment

#12 - plot the first GAT layer

Issue - State: closed - Opened by CEfanmin over 6 years ago - 1 comment

#11 - Reporting 0.8350 accuracy (epochs 1000) after fixing a bug.

Issue - State: closed - Opened by vermaMachineLearning over 6 years ago - 2 comments

#11 - Reporting 0.8350 accuracy (epochs 1000) after fixing a bug.

Issue - State: closed - Opened by vermaMachineLearning over 6 years ago - 2 comments

#10 - Not an issue but a suggestion for making code 4 X Faster

Issue - State: closed - Opened by vermaMachineLearning over 6 years ago - 1 comment

#10 - Not an issue but a suggestion for making code 4 X Faster

Issue - State: closed - Opened by vermaMachineLearning over 6 years ago - 1 comment

#9 - call() got an unexpected keyword argument 'mask'

Issue - State: closed - Opened by AYG-DL over 6 years ago - 2 comments

#9 - call() got an unexpected keyword argument 'mask'

Issue - State: closed - Opened by AYG-DL over 6 years ago - 2 comments

#8 - * Rewrite attention computations so we only do each <a,Wh> once

Pull Request - State: closed - Opened by danielegrattarola over 6 years ago

#8 - * Rewrite attention computations so we only do each <a,Wh> once

Pull Request - State: closed - Opened by danielegrattarola over 6 years ago

#7 - Performances

Issue - State: closed - Opened by Diego999 over 6 years ago - 6 comments

#7 - Performances

Issue - State: closed - Opened by Diego999 over 6 years ago - 6 comments

#6 - Result of example script with 8 attention heads

Issue - State: closed - Opened by matthias-samwald over 6 years ago - 1 comment

#6 - Result of example script with 8 attention heads

Issue - State: closed - Opened by matthias-samwald over 6 years ago - 1 comment

#5 - What is behavior of K.dot?

Issue - State: closed - Opened by machanic over 6 years ago - 6 comments

#4 - add LeakyReLU nonlinearty and change dropout rate

Pull Request - State: closed - Opened by dawnranger almost 7 years ago - 1 comment

#4 - add LeakyReLU nonlinearty and change dropout rate

Pull Request - State: closed - Opened by dawnranger almost 7 years ago - 1 comment

#3 - fix loss not improve bug and add earlystopping

Pull Request - State: closed - Opened by dawnranger almost 7 years ago - 1 comment

#3 - fix loss not improve bug and add earlystopping

Pull Request - State: closed - Opened by dawnranger almost 7 years ago - 1 comment

GitHub / danielegrattarola/keras-gat issues and pull requests