-
如果一个图节点只有一度邻居,那这个节点可以进行graphsage训练吗?
-
https://nbviewer.org/github/alibaba/GraphScope/blob/main/tutorials/zh/7_unsupervised_learning_with_graphsage.ipynb
![image](https://user-images.githubusercontent.com/38930973/149502351-bf68038c-201f-…
-
请教一个问题:在graphscope 分支下,使用vineyard 分布式的存一张大图,在大图之上跑分布式训练, 每个trainer 理论上讲都可以touch 到全图的数据,这样是否会导致 vineyard 数据分片在不同节点间频繁拷贝?
-
请教一个问题:使用vineyard 分布式的存一张大图,在大图之上跑分布式训练, 每个trainer 理论上讲都可以touch 到全图的数据,这样是否会导致 vineyard 数据分片在不同节点间频繁拷贝?
-
Thanks for your work! I find the shifting skill simple and interesting, however, it seems that the implementation in the code is different from what is in the paper.
Paper
Code
https://github…
-
"from graphlearn.python.models import *" command can be imported normally,
but "import graphlearn as gl && gl.module" will raise AttributeError:
'module' object has no attribute 'LearningBasedMode…
-
你好,
我使用graph-learn在一份较大的图数据上测试graphsage的结果。图有6千万的节点,8千万的边,初始化的node feature使用128维的向量,最终输出也是128维向量。任务在kubernetes上运行。
目前遇到两个问题:
1,在batch size 设置>5的情况下,worker节点刚开始跑就会报ResourceExhaustedError。但此时不论是worke…
-
hi,dear
when I run
`$ python cora.py`
Traceback (most recent call last):
File "cora.py", line 26, in
from utils import download, extract
File "/data/prod/xulm1/graph-learn/examples/dat…
-
I see that GraphLearn is serially sampling for each node, is there any plan to support parallel neighbor sampling in the near furture?
-
After patch the fix(#4 #11), I rebuild/reinstall graph-learn and use below commands to start dist_train.py, problem still.
```bash
PS_HOSTS="127.0.0.1:2300,127.0.0.1:2311"
WK_HOSTS="127.0.0.1:22…