Mastering ComfyUI: Custom Nodes and Character Placement in Your Images! - TUTORIAL

uh hello everyone and welcome back to Dreaming AI Noob Theory and today we are going to learn about some additional nodes that can be used in comfy UI along with an introduction to installing custom nodes custom nodes are one of the things I love most about comfy UI developers all around the world who are interested in extending comfy-wise functionalities can add or modify features that are already present in a very simple and Powerful way for the majority of cases you will find these nodes on GitHub the most famous ones are undoubtedly comfy UI manager deficiency notes for comfy UI was nodesuite come the UI nested node Builder and as far as I'm concerned also come for UI custom Scripts each of these add-ons adds specific features that make an already comprehensive interface even more versatile I suggest you read the specifications of each node for more details and choose only those that help you with what you want to achieve avoiding making comfy UI 2 cluttered I've included all the information in the description the installation process is practically the same every time and for convenience I recommend using get which we should have previously downloaded and installed from the website I'll provide in the description navigate to the custom nodes folder in comfy UI type CMD in the address bar and press enter now enter the command get clone followed by the GitHub repository address which you can find by clicking on code and selecting https so in this case the command will be like this press enter and the plugin will be automatically downloaded into that folder upon launching conf UI it will be loaded automatically now let's move on to today's example I want to create an image that features a female character on the right side against a cyberpunk city background the challenge in this example is the positioning of the character as they are typically placed in the center or at a random position um so let's open the example from the last time and modify it to achieve our goal since I want to do a bit of cleanup I'll enable the option related to the custom node conf UI custom scripts to make the connections between noted streamlines looks better right I also want to change the model since we're at it now first of all I need to separate a prompt I'll use to create the background from the one I'll use to create the character unlike the previous example we'll see how to extract the text and put it in a separate node so we can use it with multiple clip text and code nodes without having to rewrite it and making our project much more Dynamic right click on the clip node and select convert text to input this way we can pass the text from another node the node we'll use is the Primitive node this node adapts to the fields it's connected to so in this case if we connect it to a text field it will show a field to input the text do the same for both the negative node which in this example is only one and the positive notes I'll remove some tags from the character Park related to the background and add them to the other prompt now let's use another custom node called string function this node allows us to combine two text strings to have one that includes both similar to what we did with clip texting code let's turn the text a and text B fields into input nodes this way and now connect both the text for the character and the one for the background to this node connect the output to another clip text in code next we'll use another specialized node that allows us to select the area affected by our props it's called conditioning set area and we'll use two of them since we want to divide our image into right and left as you can see this node allows us to select the size of the area affected by the connected prompt as well as where it starts with the X and Y values so in our example our our output image has dimensions of 1280x704 to split it in half we'll set the width to 640 for both conditioning nodes and for the one identifying the character we'll also adjust the X so that the prop starts affecting from the 640th pixel onwards in reality I I've learned that we shouldn't be too precise with these kinds of operations because we risk creating images with left and right sides that don't match well therefore I'll reduce the height a bit on both nodes um now let's use the last new node for today which is conditioning combine this allows us to merge various conditions so let's first combine the two conditioning set area nodes and then combine the output of this with the output of the clip texting code that only has the background prompt as input it might seem complicated but in reality it's not at all we're simply killing conf UI that we want only the background in the background area and the character along with the background in the character area now connect the output of this final combine to the positive prompt of our K sampler okay also changing some settings of the K sampler that seems to work better for this kind of things but feel free to do your own tests and let's see how our first controlled area image looks not bad huh I hope this tutorial helped you learn a little bit more of the comfy UI interface if you have any questions please let me know in the comments below I'll be happy to help you out as much as I can please consider subscribing and leave a like if you want to help this channel and as always keep dreaming

Channel: DreamingAI

Views: 12,522

Keywords: generate images, images positioning, ComfyUI, basic, texttoimage, AI, stable diffusion, artificial intellingence, dreamingai, ai news, best free ai, best ai model, dreamingai tutorials, WAS Node Suite, Efficiency Nodes, ComfyUI-Custom-Scripts, ComfyUI-Manager

Id: AjwfswzLmxU

Length: 7min 14sec (434 seconds)

Published: Sat Aug 19 2023