负采样问题

[LZ-LAB]

Yu Liu，你好。我想请教一下，文中提出的n-ary知识图谱的负采样在代码中哪一块体现？

[liuyuaa]

In practice, we did the negative sampling on all entities, i.e., all other entities except the predicted one are viewed as negative samples, which comes to the softmax-based loss function here:

RAM/model.py

Line 13 in fb4de01

def forward(self, pred1, tar1):

.