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

关于图片尺寸问题 #485

Open
Chengym2023 opened this issue Nov 13, 2024 · 4 comments
Open

关于图片尺寸问题 #485

Chengym2023 opened this issue Nov 13, 2024 · 4 comments

Comments

@Chengym2023
Copy link

我要生成1024乘1024的图片,用control net导入图片生成pose图,那么导入图片的尺寸有什么要求?节点的参数怎么设置会影响清晰度?

@emogie
Copy link

emogie commented Nov 16, 2024

sorry, I can't help, never learned this language, even if it could be very interesting, but I have no affinity to learn languages, already learning English was difficult for me, since it is not my mother tongue (but I already have my issues to learn my native language). Maybe some other user can help you, If you post your question in English?

@Chengym2023
Copy link
Author

sorry, I can't help, never learned this language, even if it could be very interesting, but I have no affinity to learn languages, already learning English was difficult for me, since it is not my mother tongue (but I already have my issues to learn my native language). Maybe some other user can help you, If you post your question in English?

example: i use sdxl size is 1024by1024,but the pose image size is 512by512 for controlnet ,Does this affect the resolution?How to set controlnet up to make the picture clear

@emogie
Copy link

emogie commented Nov 18, 2024

no it does not affect the output resolution, control net size will only be used for detection, of the image, with higher resolutions you has probably the issue that the pose will not be detected properly for example if you use reolution with 4096x4096, the same with lower resolutions, as the controlnet was trained with a specified resolution. The output depends on your used modell and the choosen output resolution. To get an image sharp you have to play with cfg, choose a bigger size of output resolution, as you need, for, example if you need 1024x1024, try 1152x1152 and rescale than with ultrasharp 4x, or leave it untouched if it is ok for you

@Chengym2023
Copy link
Author

no it does not affect the output resolution, control net size will only be used for detection, of the image, with higher resolutions you has probably the issue that the pose will not be detected properly for example if you use reolution with 4096x4096, the same with lower resolutions, as the controlnet was trained with a specified resolution. The output depends on your used modell and the choosen output resolution. To get an image sharp you have to play with cfg, choose a bigger size of output resolution, as you need, for, example if you need 1024x1024, try 1152x1152 and rescale than with ultrasharp 4x, or leave it untouched if it is ok for you

thank you ,i think i know^-^

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