kimiyoung / planetoid Goto Github PK

Hi,

could you please provide the pre-processing script? Would require that to extend to other datasets.

thanks,
Deep

rzai@rzai00:/prj/planetoid$ python test_trans.py
Using gpu device 0: GeForce GTX 1080 (CNMeM is disabled, cuDNN 5005)
Traceback (most recent call last):
File "test_trans.py", line 3, in
from trans_model import trans_model as model
File "/home/rzai/prj/planetoid/trans_model.py", line 2, in
import lasagne
File "/usr/local/lib/python2.7/dist-packages/lasagne/init.py", line 19, in
from . import layers
File "/usr/local/lib/python2.7/dist-packages/lasagne/layers/init.py", line 7, in
from .pool import *
File "/usr/local/lib/python2.7/dist-packages/lasagne/layers/pool.py", line 6, in
from theano.tensor.signal import downsample
ImportError: cannot import name downsample
rzai@rzai00:/prj/planetoid$

Hello, what do files ".x .y .tx .ty .allx .ally .graph .index" store？

run the code cost a little effort in environment setting .

Can you please elaborate in detail that how can we generate the test.index file, while using the code with image dataset?
Please elaborate the procedure to generate the test.index file.

Thanks in advance...

Hello, could you please provide the original dataset before your preprocessing ? It seems it does not match the dataset in the following link https://linqs.soe.ucsc.edu/data. It seems your train/val/test splits are well chosen. Some labels are also different from the original dataset in https://linqs.soe.ucsc.edu/data. Do you have any ideas about this ? Thanks.

想问下是怎么生成的.x .y 这些文件的呢？是有一个预处理过程吗？我用原版的cora数据集训练出来只有65左右的精确度（400epoch），是因为预处理的问题吗？
可以的话希望能分享下生成.x .y文件的代码，好多人都想要

Hi, could you tell me how can I process other original datasets into the format files of 'x', 'y', 'allx', 'ally' etc like yours? Thanks a lot!

I sincerely appreciate your efforts in curating diverse datasets for the graph field.

I am trying to analyze the semantics of nodes and their connectivity in the datasets.

So, is there any dictionary for bag-of-word embedding that matches an index to a word?
Or, can I get the source code for preprocessing the bag-of-word embedding?

Hi! I am trying to figure out the .test.index file. Is there a reason why the nodes are not ordered, does the order hold some meaning?

Dear Sir,

Could you please upload the version of code compatible with Python3?

Regards,
Mabi

Hi there! I use the CORA dataset to train the model and explain it using model interpretation methods. But in order to fully understand the operation of the model, the names of the features that are encoded in the tensor data.x (cora.x) are missing. Where can I find a dictionary with the names of the encoded features?

If I want to get the embedding of cora ,how can i get it?

The shape of allx is 18717*500 shouldn't it be 19717x500?

Running test_trans and test_ind.py on the CITESEER dataset both yield (a few % points) better performance than reported in the paper. Any idea why that would be? Is the implementation here slightly different, or is it just a function of train/test split?

EDIT: Same goes for CORA, but I haven't been able to reproduce PUBMED -- got to ~ 0.64 and then I killed it, so perhaps I didn't let it run long enough. At that point accuracy was increasing very slowly, though.

When I use citeseer data, it will report error that index 3312 is out of bounds for dim with size 3312. Or expected dim 0 size 3327,got 3312. I don't know whether citeseer has 3312 or 3327 nodes. Please help me to solve the question.Thanks

hello, when I run the test_trans.py, the error message occurs:

File "F:/googledownload/planetoid-master/test_trans.py", line 49, in
OBJECTS.append(cPickle.load(open("data/trans.{}.{}".format(DATASET, NAMES[i]),"rb+")))
UnicodeDecodeError: 'ascii' codec can't decode byte 0x85 in position 16: ordinal not in range(128)

do you know how to solve it?
my python version is 3.6

While true is an infinite loop.
How many times does it take to cycle to achieve the accuracy of the paper?