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

为什么OWOD任务第一步文本特征提取,每次提取的feature都不一样 #12

Open
wl654655902 opened this issue Jan 17, 2025 · 1 comment

Comments

@wl654655902
Copy link

第一步我运行多次,将object.npy读取出来发现前后两次的数值不一样,同样task的npy也不一样。
如果我任务task1是人车,task2是烟火,执行第一步,第二步,第三步和第四步。
当我交换task顺序,task1是烟火,task2是人车时,我执行第一步,可以不执行第二步和第三步吗,直接使用交换任务顺序之前得到的object_tuned.npy,直接执行第四步,我看跳过第二步和第三步,直接执行第四步也能收敛。

第一步执行多次为什么提取的特征不一样,会对后面步骤有影响吗,不想每次都执行第二步和第三步,第二步训练费时

@leonnil
Copy link
Contributor

leonnil commented Jan 17, 2025

感谢您的关注。第二步和第三步应该是和open-world task无关的,只需要运行一次即可。您可以在提取特征代码里加入model.eval()试试保证提取特征的稳定性。我们后续也对代码做相应的更新,谢谢!

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

2 participants