nefujiangping / encattagg Goto Github PK

首先非常感谢您共享了代码！结合论文看了代码，有几个问题想请教一下：
1.mutual_attender和integration_attender 部分的代码，一个是公式4一个是公式6，代码核心都是实现Transformer的一个Encoder层，为什么mutual_attender没有和integration_attender 一样使用nn.TransformerEncoderLayer呢？
2.models.py中的TwoLayerLinear是实现公式7吗，但没有看到LogSumExp这个公式的实现部分，还是LogSumExp在哪里实现了我没看见呢？

期待您的回复，感谢！

Thanks for your great work. Since we some people cannot access your paper because it is not free of charge in IEEE, could you provide your results on DocRED / CDR for reference?

Thank you for sharing! I can't find the 'modeling' package at runtime, and I can't install it using pip and conda. Is this package customized or installed from? Hope you can help me, thank you!

Hello, thank you for your sharing. I would like to ask how to set the value of bert_feature during bert training, thank you!