ComfyUI Conditioning Average ControlNet workflow #comfyui #stablediffusion #controlnet

hi today we will see how to use this conditioning average node and how it is possible to achieve such changes between the different images our workflow is very simple and of course you will find it in the description of the video we have one control net here with depth anything I will attach a video in the link that shows how to install this model of course you have to prepare the model with a pre-processor this is the image based on which we will create the depth map and here you can see the depth map that we will work with of course our seed is fixed so that it does not affect the result and let's see what we have in the prompts notice that I created two positive prompts in both of which I wrote the same prompt only with a change of the season of the year plus a description that fits the season these two prompts are are connected to conditioning average which does exactly what its name implies an average between the two prompts as soon as we set the slider to one we will receive the prompt that is connected to conditioning to as soon as we set the prompt to zero we will receive the prompt that is connected to conditioning from and of course as soon as we are somewhere in the middle we will receive all kinds of combinations of the prompts and in our case different times of the [Music] year let's try to do it now with an image and a different subject and I hope you will see how powerful this tool is to work with I will bypass the control net and write a simple prompt Studio photo of woman blond hair soft light High key sharp I skip the conditioning average for a moment and connect the positive prompt directly to the K sampler and press Q as you can see we got a character that is kind of perfect and quite a few times I want to get characters that are a little more realistic and less Barbies for this I will write in the second prompt words that contradict Beauty and Perfection such as ugly tired blemishes and you can add anything you see fit now before we click on Q I will first copy the image we got to control net so that we get more or less the same character but with a different qualities we will activate the control net and connect the positive to the K sampler let's just see that everything works great and now we have a slider that affects the character of our model as soon as we want a less perfect and clean look we simply reduce the slider so I hope you learned and we will meet in the next lessons of course you are welcome to subscribe to the channel to ask questions comment and like and most importantly have fun bye

Channel: PixelEasel

Views: 1,877

Keywords:

Id: A0an6sz4fcQ

Length: 3min 19sec (199 seconds)

Published: Tue Feb 27 2024