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

请问作者大大,encoder如何训练? #996

Open
onedotone-wei opened this issue Apr 30, 2024 · 2 comments
Open

请问作者大大,encoder如何训练? #996

onedotone-wei opened this issue Apr 30, 2024 · 2 comments

Comments

@onedotone-wei
Copy link

看了知乎链接的教程,尝试训练encoder
Screenshot 2024-04-29 184437
练了半天,这结果似乎没什么变化
wei_umap_038600
wei_umap_038700
数据自建的,有2个多G
问题1:这种情况是正常的吗?如果不正常是什么原因造成的?
问题2:根据知乎上的说法“实测了一次 训练synthesizer时,4000左右step就能attention收敛,22k step的时候loss就到0.35了,可以很快进行finetune,算是超越预期。”,训练synthesizer时,如何把encoder加入?

@babysor
Copy link
Owner

babysor commented Apr 30, 2024

正常。encoder的训练要求要高很多,数据量要大、step要多很多,建议只做微调。

由于结构问题,encoder和synth是分开训练的

@onedotone-wei
Copy link
Author

感谢回复,我再跑一段时间看看

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