kangxiatao/kgat-pytorch-master

Knowledge Graph Attention Network

kgat-pytorch-master's Issues

报错

您好，
我训练的是douban250的数据，但是它好像没有npz预训练，所以我用其他三个数据集的预训练npz，请问训练别的数据集也是可以用用一个mf.npz吗，我这边报错了
2023-11-09 20:30:34,516 - root - INFO - Namespace(K=20, aggregation_type='bi-interaction', cf_batch_size=1024, cf_l2loss_lambda=1e-05, cf_print_every=1, conv_dim_list='[64, 32, 16]', data_dir='datasets/', data_name='douban250', entity_dim=64, evaluate_every=1, kg_batch_size=1024, kg_l2loss_lambda=1e-05, kg_print_every=1, local_rank=0, lr=0.0001, mess_dropout='[0.1, 0.1, 0.1]', n_epoch=1000, pretrain_embedding_dir='datasets/pretrain/', pretrain_model_path='trained_model/KGAT/amazon-book/entitydim64_relationdim64_bi-interaction_64-32-16_lr0.0001_pretrain1/model_epoch1.pth', relation_dim=64, save_dir='trained_model/KGAT/douban250/entitydim64_relationdim64_bi-interaction_64-32-16_lr0.0001_pretrain1/', seed=2020, stopping_steps=10, test_batch_size=1024, use_graph=1, use_pretrain=1)
All logs will be saved to trained_model/KGAT/douban250/entitydim64_relationdim64_bi-interaction_64-32-16_lr0.0001_pretrain1/log1.log
device: cpu n_gpu: 0
load data ...
-- datasets/douban250 --
2023-11-09 20:30:34 -- cf data finish --
2023-11-09 20:30:34 -- kg data load --
2023-11-09 20:30:36 -- kg data finish --
70679
Traceback (most recent call last):
File "D:/KGAT-pytorch-master-main/main_kgat.py", line 276, in
train(args)
File "D:/KGAT-pytorch-master-main/main_kgat.py", line 86, in train
data = DataLoaderKGAT(args, logging)
File "D:\KGAT-pytorch-master-main\utility\loader_kgat.py", line 90, in init
self.load_pretrained_data()
File "D:\KGAT-pytorch-master-main\utility\loader_kgat.py", line 352, in load_pretrained_data
assert self.user_pre_embed.shape[0] == self.n_users
AssertionError
然后请问kg_final.txt要怎么生成？

关于calc_kg_loss中计算pos_score时的报错。

您好，有一个问题想请教一下。
当我运行到这一步的时候：
kg_batch_loss = model('calc_kg_loss', kg_batch_head, kg_batch_relation, kg_batch_pos_tail, kg_batch_neg_tail)
出现了报错。
然后我尝试在calc_kg_loss方法中，print出pos_score，定位到了问题出在计算pos_score这一步。

部分报错信息如下：
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [64,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [65,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [66,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [67,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [68,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [69,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [70,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [71,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [72,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [73,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [74,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [75,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [76,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [77,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [78,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [79,0,0] Assertion srcIndex < srcSelectDimSize failed.
/pytorch/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [13,0,0], thread: [80,0,0] Assertion srcIndex < srcSelectDimSize failed.
。。。。。
File "train.py", line 137, in
kg_batch_neg_tail)
File , line 727, in _call_impl
result = self.forward(*input, **kwargs)
File , line 190, in forward
return self.calc_kg_loss(*input)
File , line 168, in calc_kg_loss

前面的r_mul_h，r_embed，r_mul_pos_t，打印出来size都是（256，64）。
但是这一步计算torch.sum(torch.pow(r_mul_h + r_embed - r_mul_pos_t, 2), dim=1) 无法正常执行了，能否帮忙看一下问题出现在哪里呢？非常感谢！

训练顺序的疑问

在main_kgat.py中，我注意到您是先求attention，然后训练kg_Loss，最后训练cf_Loss.

这么做的原因是因为什么呢？

而且我发现当不使用--use_pretrain之后，效果衰减的很厉害。几乎可以称得上训练不了。

KGAT-高阶传播

您好，我对高阶传播这部分代码有疑问，能不能麻烦您解释一下
for idx, layer in enumerate(self.aggregator_layers):
ego_embed = layer(ego_embed, self.A_in)
norm_embed = F.normalize(ego_embed, p=2, dim=1)
all_embed.append(norm_embed)
请问论文中的公式10就是ego_embed = layer(ego_embed, self.A_in)吗，那么论文中的邻居结点是如何体现高阶传播呢，怎么好像都是一阶节点在学习embedding呢。
期待您的回复。

关于数据集中kg_final.txt的问题

请问您自己制作的数据集中的知识图谱信息是怎么做的呢？尤其是kg_final.txt。谢谢！

请问ndcg20指标为什么跟原论文相比差了很多？谢谢~

请问ndcg20指标为什么跟原论文相比差了很多？谢谢~

关于issue

各位有issue和私信的学者，实在不好意思，后面基本上没回复。我这个课题有点久远了，基本都忘了，现在在做其他方向，有点无力。KGAT的整体代码还是蛮清晰的，里面的注释点和论文都有对应，可以debug慢慢读一遍。祝各位多发paper，少掉头发、基金职称两开花！