Skip to content
This repository has been archived by the owner on Nov 11, 2023. It is now read-only.

RuntimeError: output with shape [1, 256] doesn't match the broadcast shape [200, 256] #53

Closed
thekingofcircus opened this issue Mar 18, 2023 · 15 comments

Comments

@thekingofcircus
Copy link

在Featurize平台上租卡跑的时候(linux系统),在训练时报错RuntimeError: output with shape [1, 256] doesn't match the broadcast shape [200, 256],试了一下午了,怎么办捏

@HuanLinOTO
Copy link
Contributor

没给配置 没给上下文 你说我能不能帮你找问题


Generated by koishi-plugin-github.

@NaruseMioShirakana
Copy link
Contributor

报错不发上下文,我们能做的只有给你算上一卦()

@Miuzarte
Copy link
Contributor

9V6T1GIT{{PD@UP~UYFCT_4

@HuanLinOTO
Copy link
Contributor

把报错贴到 这儿 不要往 issue 里盖高楼


Generated by koishi-plugin-github.

@thekingofcircus
Copy link
Author

用的是3090的卡,报错如下:
https://pastebin.com/1v5fNbA1

@HuanLinOTO
Copy link
Contributor

上下文仍然不足 请提供运行的命令 配置

@thekingofcircus
Copy link
Author

@HuanLinOTO
Copy link
Contributor

你配置呢

@HuanLinOTO
Copy link
Contributor

尝试不使用底模


Generated by koishi-plugin-github.

@thekingofcircus
Copy link
Author

好的,我试试,红豆泥阿里嘎多

@NaruseMioShirakana
Copy link
Contributor

这个问题也许是由于多角色引起的,模型的 gin_channel 为 256 而 n_speaker 为 200。
试着将模型配置文件之中的n_speaker改为1,重启训练看看还有没有这个问题

@thekingofcircus
Copy link
Author

去掉底模之后成功运行了,多角色没有尝试

@a-cold-bird
Copy link

没有底模会影响模型质量,但是加上底模就报错(悲

@a-cold-bird
Copy link

a-cold-bird commented Apr 12, 2023 via email

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

No branches or pull requests

6 participants