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

请问一般一轮训练的loss value一般为多少 #15

Open
miaoshoukk opened this issue Jun 24, 2024 · 6 comments
Open

请问一般一轮训练的loss value一般为多少 #15

miaoshoukk opened this issue Jun 24, 2024 · 6 comments

Comments

@miaoshoukk
Copy link

请问一般一轮训练的loss value一般为多少

@turturing
Copy link

我loss的值一轮有好几万这正常吗

@DeweiHu
Copy link
Owner

DeweiHu commented Jun 25, 2024

That's reasonable, because I normalize the intensity range to [1, 3]. But if you don't want to use my model checkpoint, you can normalize it to [-1, 1] which is the common setting for diffusion model.

@miaoshoukk
Copy link
Author

我也是20000多下不来

@miaoshoukk
Copy link
Author

这是合理的,因为我将强度范围归一化为 [1, 3]。但是,如果您不想使用我的模型检查点,则可以将其归一化为 [-1, 1],这是扩散模型的常见设置。

请问这是直接在Net里面改吗,我已经在dataloader中改成归一化为 [-1, 1]依然是20000多

@turturing
Copy link

That's reasonable, because I normalize the intensity range to [1, 3]. But if you don't want to use my model checkpoint, you can normalize it to [-1, 1] which is the common setting for diffusion model.

请问是在哪里进行强度归一化

@miaoshoukk
Copy link
Author

这是合理的,因为我将强度范围归一化为 [1, 3]。但是,如果您不想使用我的模型检查点,则可以将其归一化为 [-1, 1],这是扩散模型的常见设置。

请问是在哪里进行强度归一化

应该是在读取数据时进行归一化吧

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