search

zjukg / Structure-CLIP

[Paper][AAAI2024]Structure-CLIP: Towards Scene Graph Knowledge to Enhance Multi-modal Structured Representations
https://arxiv.org/abs/2305.06152
104 stars 2 forks source link

myTransformer bert_Transformer #11

Closed zynlp1999 closed 4 months ago

zynlp1999 commented 5 months ago

Hello, I am very interested in your article. To repeat your comparison experiment, triple_Transformer myTransformer bert_Transformer. There are many problems when changing myTransformer bert_Transformer, how should I operate it? Thank you for your reading.

BigHyf commented 4 months ago

Hello, I am very interested in your article. To repeat your comparison experiment, triple_Transformer myTransformer bert_Transformer. There are many problems when changing myTransformer bert_Transformer, how should I operate it? Thank you for your reading.

你好，感谢你关注我们的工作，我们论文中用到的应该只有triple transformer，也就是主程序里引用的部分，triple transformer和my transformer是之前的一些尝试，在最终的实验代码里面并没有用到。

zynlp1999 commented 4 months ago

好的好的，非常感谢您的回复，想再问一下，如果我想跑没有语义负样本的可以直接把neg_loss的比例设置为0吗，这样就算没有语义负样本的了吗？非常感谢您

BigHyf commented 4 months ago

好的好的，非常感谢您的回复，想再问一下，如果我想跑没有语义负样本的可以直接把neg_loss的比例设置为0吗，这样就算没有语义负样本的了吗？非常感谢您

你可以试一试，但我印象里如果没有语义负样本的话在下游数据集上的效果挺差的。我们这片工作的核心就是在语义负样本这一块。

zynlp1999 commented 4 months ago

好的好的谢谢您 那想问下您的代码中是怎么实现没有语义负样本呢的实验呢