FileNotFoundError: [Errno 2] No such file or directory: 'datasets/DB15K/text_features.pkl'

[fei-wang-1314]

Hi, Could you please provide the file: text_features.pkl in every dataset mentioned in your paper and other files needed, such FB15k-237_EntityTriples, etc.? Thank you very much.

Best regards,

[bbshu0801]

I asked via email. They lost all files since the server changing. :(

[fei-wang-1314]

@bbshu0801 Thank you.

[bbshu0801]

@bbshu0801 Thank you.

But I think you can merge train,val,test.txt by yourself. It will be same as the other EntityTriples files after merging.

[fei-wang-1314]

@bbshu0801 Thank you.

But I think you can merge train,val,test.txt by yourself. It will be same as the other EntityTriples files after merging.

Thank you for your time and reply. There still no text_features.pkl

[bbshu0801]

@bbshu0801 Thank you.

But I think you can merge train,val,test.txt by yourself. It will be same as the other EntityTriples files after merging.

Thank you for your time and reply. There still no text_features.pkl

text_features.pkl is produced by BERT which authors mentioned in their paper. This means that you also need to do this by yourself.

[fei-wang-1314]

@bbshu0801 Thank you.

But I think you can merge train,val,test.txt by yourself. It will be same as the other EntityTriples files after merging.

Thank you for your time and reply. There still no text_features.pkl

text_features.pkl is produced by BERT which authors mentioned in their paper. This means that you also need to do this by yourself.

I see. Thank you!

[cc1300386495]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

[bbshu0801]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

[cc1300386495]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

万分感谢！想再咨询您一个问题 ，您在训练encoder时，GAT网络计算loss那里的代码“loss = model.loss_func(train_indices, entity_embed, relation_embed)”，您改动了吗。我在电脑上运行到这里的时候报错“RuntimeError: CUDA error: device-side assert triggered”，不知道是我的显存不足，还是代码需要改动。

[bbshu0801]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

万分感谢！想再咨询您一个问题 ，您在训练encoder时，GAT网络计算loss那里的代码“loss = model.loss_func(train_indices, entity_embed, relation_embed)”，您改动了吗。我在电脑上运行到这里的时候报错“RuntimeError: CUDA error: device-side assert triggered”，不知道是我的显存不足，还是代码需要改动。

这句我没改，也没报错哦

[cc1300386495]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

万分感谢！想再咨询您一个问题 ，您在训练encoder时，GAT网络计算loss那里的代码“loss = model.loss_func(train_indices, entity_embed, relation_embed)”，您改动了吗。我在电脑上运行到这里的时候报错“RuntimeError: CUDA error: device-side assert triggered”，不知道是我的显存不足，还是代码需要改动。

这句我没改，也没报错哦

好滴，那看来就是我显存比较小了~谢谢您啦

[sun8982]

我通过电子邮件问道。自从服务器更改以来，他们丢失了所有文件。:(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

万分感谢！想再咨询您一个问题 ，您在训练encoder时，GAT网络计算loss那里的代码“loss = model.loss_func（train_indices， entity_embed， relation_embed）”，您改动了吗。我在电脑上运行到这里的时候报错“RuntimeError： CUDA error： device-side assert triggered”，不知道是我的显存不足，还是代码需要改动。

这句我没改，也没报错哦

想问一下你的环境版本是什么？我用的1.7的torch和11.0的cuda，我也是在这个地方报错。

[vialstar]

I asked via email. They lost all files since the server changing. :(

您好，我想请教一下， pkl文件需要自己生成，格式是什么呢？我看MMKG中的DB15K是h5格式的文件，数据格式是{ID：embedding}，您在复现时是怎么处理的呢

在这里面：https://public.ukp.informatik.tu-darmstadt.de/starsem18-multimodalKB/

万分感谢！想再咨询您一个问题 ，您在训练encoder时，GAT网络计算loss那里的代码“loss = model.loss_func(train_indices, entity_embed, relation_embed)”，您改动了吗。我在电脑上运行到这里的时候报错“RuntimeError: CUDA error: device-side assert triggered”，不知道是我的显存不足，还是代码需要改动。

这句我没改，也没报错哦 你好，我想请问一下这里面的视觉特征转成pkl的时候是存成数组吗？怎么存了以后报文本和图像实体数不匹配啊。