Skip to content

TransG #1

@codeants2012

Description

@codeants2012

TransG 模型对FB15K数据集进行训练时,为什么达不到论文中所提到的效果,我的参数完全按照论文中的参数设定的,请问TransG类的其他参数如何设置?
我训练的效果如下:
这是left:
1:0.787124
2:0.883082
3:0.244106
4:0.565695
Raw.BestMEANS = 265.16
Raw.BestMRR = 0.205449
Raw.BestHITS = 0.411251
Filter.BestMEANS = 152.369
Filter.BestMRR= 0.331879
Filter.BestHITS = 0.548831
total = 59071

这是right:
1:0.76324
2:0.286448
3:0.902593
4:0.592057
Raw.BestMEANS = 194.49
Raw.BestMRR = 0.265348
Raw.BestHITS = 0.489513
Filter.BestMEANS = 122.189
Filter.BestMRR= 0.390626
Filter.BestHITS = 0.614701
total = 59071
与论文中的相差有点大。我的n_cluster = 20; step_before = 10,variance_bound = 0.01,sot = false, be_weight_normalized = true; @BookmanHan

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions