Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

句子对任务的RoBERTa-tiny-pair的ckpt文件的问题 #4

Open
drzqb opened this issue Mar 10, 2020 · 8 comments
Open

句子对任务的RoBERTa-tiny-pair的ckpt文件的问题 #4

drzqb opened this issue Mar 10, 2020 · 8 comments

Comments

@drzqb
Copy link

drzqb commented Mar 10, 2020

句子对任务的RoBERTa-tiny-pair的ckpt文件里面为什么没有pool层出口处的(312,2)的张量权重呢,就是"cls/seq_relationship"下的“output_weights”和”output_bias“”?,没有这个怎么得到相似与否的概率值呢?难道这个相似度计算是由pool出口的向量用余弦相似度计算的?

@brightmart
Copy link
Member

你可以再下游任务训练一下,就可以了。

@brightmart
Copy link
Member

你可以下游任务训练吗?

@drzqb
Copy link
Author

drzqb commented Mar 10, 2020 via email

@DukeEnglish
Copy link
Contributor

DukeEnglish commented Mar 11, 2020 via email

@drzqb
Copy link
Author

drzqb commented Mar 11, 2020 via email

@brightmart
Copy link
Member

添加了新模型,这两个新模型下面都 包含全部权重。你看看

@drzqb
Copy link
Author

drzqb commented Mar 11, 2020 via email

@drzqb
Copy link
Author

drzqb commented Mar 11, 2020

测试了一下,用tiny3L312,结果挺奇怪的,不管是完全相同的两个句子的相似度还是完全不同意思的两个句子的相似度都是大约0.5,有点随机初始化权重的感觉。有哪位大佬测试过吗?请教学习

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants