Thank you very much for showcasing everything ! And thank you for showing Adetailer, I really never knew about it now I don't need to inpaint every image after generation !
Hey ! thanks for the great tuto ! How to generate only depth map or pose pictures with CN, and use each of them as a mask for batch apply after ? I want to use it as mask for an image sequence for doing videos.
You are welcome, in Img2Img if you apply controlnet in Batch image processing, it will automatically create the mask for the current input image, as i display in this video ua-cam.com/video/k4ZWJD6W8d0/v-deo.html so you can use controlnet for openpose/depth map combination for instance and it will apply that .... this will not however save the openpose or depthmap pictures indepenently .... if you want them independent for other purposes one can possibly use ComfyUI and save the preprocessor mask output directly using save image node ... or in A1111 one may use for instance Rembg script in Extras tab and check Return Mask option (this requires Rembg extension installation), similarly there should extension for openpose such as openpose editor which may allow you to generate poses and save them as images if that is what you mean.
@@AI-HowTo thanks for sharing the tip ! I will do some experiments and hope it’s functioning ! I will try the extensions and give a shot to ComfyUI, btw. Maybe it will be better. I want to create a mask as pre production to light the process;
Thank you so much for the information. Could you please provide information about how to add IP2P in ControlNet options , and please provide link to model for sdxl. I am not able to find it.
Yes, absolutely, as both use the same Controlnet models, and the same principles, the application methodology changes... what matters is to understand what these models do then discover how they can be used in the target tool such as A1111/ComfyUI
That is a truly awesome guide; I have just subscribed. If I may ask, what should I do if I want to exclude a specific part from the Canny model? Here's my scenario: I have a custom model, and I want to mimic images, such as the pose and environment. To achieve this, I've created two instances in ControlNet 1- OpenPose 2- Canny. The issue I'm encountering is that Canny is replicating the overall shape of my head from the source image. How can I exclude the head part from Canny? I hope you understand my question, and a big thanks in advance.
thanks, canny can be very strict, this is why we can use depth maps for example to reduce this strickness and get the overall shape of the body, or change the edge detection settings to make it weaker .... not sure however if we can exclude a direct part of the canny processor automatically in A1111, I dont this this feature exsits at the time being, more complicated workflows can be performed in ComfyUI but that requires more work.... if you are working on a single image then you can use inpaint afterwards to fix it, but for automation ComfyUI is the way to go in complicated scenarios.
I got them in the past from civitai, not sure from which source, so I give you a better link civitai.com/tag/poses which has over 700 collections of different poses/collections with their previews that you can download any .
some very good pose collections are: civitai.com/api/download/models/33869 civitai.com/api/download/models/22375 civitai.com/api/download/models/21847 civitai.com/api/download/models/22265 civitai.com/api/download/models/93697 civitai.com/api/download/models/59119 etc.
Thanks so much for this...really makes sense & clear the confusion of its powerful function of each buttons & perform actions & render...hope to see more explanations... just subscribe.
I am not sure if you can! if you check this video studio.ua-cam.com/users/videoPdkBmDoUIy8/edit for downgrading A1111 there was a release history with hascode allowing us to go back to each previous release, but in Controlnet there is not on github, so probably you cannt downgrade since info doesnt exist online, there might be a way to rollback, but i am not sure how .... there is only an older archieved version which is v1 github.com/lllyasviel/webui-controlnet-v1-archived which can be installed by uninstalling current contorlnet extension, then installing this from the link i just provided.
Really handy way of showing off the features of controlnet.. I was a bit lost as to what it does. Thanks!
Glad it was helpful
Thank you very much! Keep up the good work! Finally someone who is explaining all these ControlNets. The official info about them are very sparse.
Much appreciated! hopefully it is useful and clear enough video.
Thank you very much for showcasing everything !
And thank you for showing Adetailer, I really never knew about it now I don't need to inpaint every image after generation !
You are welcome, after detailer should indeed by integral part of A1111.
Hey ! thanks for the great tuto ! How to generate only depth map or pose pictures with CN, and use each of them as a mask for batch apply after ? I want to use it as mask for an image sequence for doing videos.
You are welcome, in Img2Img if you apply controlnet in Batch image processing, it will automatically create the mask for the current input image, as i display in this video ua-cam.com/video/k4ZWJD6W8d0/v-deo.html so you can use controlnet for openpose/depth map combination for instance and it will apply that .... this will not however save the openpose or depthmap pictures indepenently .... if you want them independent for other purposes one can possibly use ComfyUI and save the preprocessor mask output directly using save image node ... or in A1111 one may use for instance Rembg script in Extras tab and check Return Mask option (this requires Rembg extension installation), similarly there should extension for openpose such as openpose editor which may allow you to generate poses and save them as images if that is what you mean.
@@AI-HowTo thanks for sharing the tip ! I will do some experiments and hope it’s functioning ! I will try the extensions and give a shot to ComfyUI, btw. Maybe it will be better. I want to create a mask as pre production to light the process;
Is there any possibility to use batch with openpose, making each image with a different seed?
Thank you so much for the information. Could you please provide information about how to add IP2P in ControlNet options , and please provide link to model for sdxl. I am not able to find it.
You are welcome, unfortunately i am not planning on doing new videos now, not all contorlnets have SDXL model, not sure about ip2p.
Great tutorial! Can this be achieved with Comfyui ?
Yes, absolutely, as both use the same Controlnet models, and the same principles, the application methodology changes... what matters is to understand what these models do then discover how they can be used in the target tool such as A1111/ComfyUI
That is a truly awesome guide; I have just subscribed.
If I may ask, what should I do if I want to exclude a specific part from the Canny model?
Here's my scenario:
I have a custom model, and I want to mimic images, such as the pose and environment.
To achieve this, I've created two instances in ControlNet
1- OpenPose
2- Canny.
The issue I'm encountering is that Canny is replicating the overall shape of my head from the source image.
How can I exclude the head part from Canny? I hope you understand my question, and a big thanks in advance.
thanks, canny can be very strict, this is why we can use depth maps for example to reduce this strickness and get the overall shape of the body, or change the edge detection settings to make it weaker .... not sure however if we can exclude a direct part of the canny processor automatically in A1111, I dont this this feature exsits at the time being, more complicated workflows can be performed in ComfyUI but that requires more work.... if you are working on a single image then you can use inpaint afterwards to fix it, but for automation ComfyUI is the way to go in complicated scenarios.
definitely gonna use inpaint afterwards, thatnks in advance @@AI-HowTo
Great content! Thank you for sharing and keep the great work. Just found your channel, really good 👍🏻
Thank you for your supporting words, best of luck to you as well.
great explanations thansk !
You are welcome!
Great one! Could I maybe have you pose folder?
I got them in the past from civitai, not sure from which source, so I give you a better link civitai.com/tag/poses which has over 700 collections of different poses/collections with their previews that you can download any .
some very good pose collections are: civitai.com/api/download/models/33869
civitai.com/api/download/models/22375
civitai.com/api/download/models/21847
civitai.com/api/download/models/22265
civitai.com/api/download/models/93697
civitai.com/api/download/models/59119
etc.
you're the best! keep going@@AI-HowTo
Noob here. What SD interface are you using?
A1111, Automatic 1111, You may consider watching this ua-cam.com/video/RtjDswbSEEY/v-deo.html if you are new to this free tool.
Amazing, thanks
Thanks so much for this...really makes sense & clear the confusion of its powerful function of each buttons & perform actions & render...hope to see more explanations... just subscribe.
Glad it was helpful
and welcome
i need to downgrade CN, how do i do that please?
I am not sure if you can! if you check this video studio.ua-cam.com/users/videoPdkBmDoUIy8/edit for downgrading A1111 there was a release history with hascode allowing us to go back to each previous release, but in Controlnet there is not on github, so probably you cannt downgrade since info doesnt exist online, there might be a way to rollback, but i am not sure how .... there is only an older archieved version which is v1 github.com/lllyasviel/webui-controlnet-v1-archived which can be installed by uninstalling current contorlnet extension, then installing this from the link i just provided.