The kernel for network/line.ipynb appears to have died. It will restart automatically.

[p-dre]

Hello I would like to use LINE very much within a Jupyter hub. If I run the following example code, everything works without problems and impressively fast.

app = gap.GraphApplication(dim=128)
app.load(file_name=gv.dataset.blogcatalog.train)
app.build()
app.train()

However, as soon as I use my own data, the core crashes during app.train(). I have selected a very small sample of 200 edges. The kernel for network/line.ipynb appears to have died. It will restart automatically. So i assume that the crash of the kernel is due to the way i enter my data

app = gap.GraphApplication(dim=128)
app.load(file_name = "data/edges_list_2021_weight.txt")
loading graph from data/edges_list_2021_weight.txt

<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
Graph<uint32>
------------------ Graph -------------------
#vertex: 358, #edge: 200
as undirected: yes, normalization: no
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

app.build()
[time] GraphApplication.build: 0.0535276 s

app.train()

the data looks like

439485003   58285254
891616492269293568  17864468
1124428778170662913 46673995
1158642693544775681 19802110
1643969455  433607851
71649699    143389019
1309811119490510855 1359884487702827010
721959981415854080  3071965750
55563609    52025567

and is created with

with open('data/edges_list_2021_test.csv', "r") as fin, open('data/edges_list_2021_weight.txt', "w") as fout:
            for line in fin:
                fout.write(line.replace(",", "\t"))

64 RAM 16 CPU NVIDIA A40 12 RAM