ComfyUI Basic Workflow Build Explain Text2Img Workflow | Text2img Workflow | ComfyUI

so hey guys welcome back once again and in today's video I'm just gonna discuss about how you can go and create your own workflows on confi UI as you can see over here I was already messing around with the control net workflow that was created by one of my friend Paolo on Discord server let me tell you guys it's easy to go and drag the workflows from the Discord server uh like Paulo is already creating some of the amazing workflow on my Discord server the only thing that you need to do is download this image and drag it let me tell you guys yes these are amazing workflows that are created by Paolo and you can see the results over here so this is difficult to explain right now so what I have decided I will start with the basics text to image in order for you to understand how you can create easy workflows so now guys you can see this is clear over here and now we will start building a basic text to image workflow so the first thing that we need is a checkpoint right because yeah without checkpoint we can't generate any image so what we will do we will try to add checkpoint right click over here or you can double click as well there are two options if you double click then it will load something like this but yes if you are pro in order to if you know the nodes then yes definitely you can use this if you don't know then let me tell you guys just right click and you will find this option over here so today we will go with this option only right click option and it's very easy so the first thing we will add is a checkpoint loader so just right click and add node over here and click onto add note and then we will go to loaders so in loaders you will find checkpoint loader there are two things checkpoint loader and checkpoint loader simple we will go with simple and as you can see over here checkpoint loader is added so the next thing we will do what we need the next thing is we need prompt in order to generate image so in order to add the prompt let's right click over here as well let's add node and what we will do we will go to conditioning you will find it on the conditioning clip text and code we need this thing so what we will do we will add this thing this is basically a text section so we need 2 prompt right like a positive prompt and a negative prompt so the next thing what we will do we will just clone it like we will duplicate it so right click on this section and yes you can duplicate it as well if you clone it yeah you can see there are two sections going on over here so this is basically two sections and yes you can rename it as well like if you right click over here you can go and select the title and let's go with positive and you can rename this as well as I've already renamed it negative okay so that's the second thing we need positive prompt and negative prompt so the next thing that we need is the width and height ratio right for the image so for that you need to go just go on to add node and just go on to latent in latent you will find empty latent image over here so just add this over here and you can see this is empty latent image you can select width and height of your image and batch size over here as well so this is setup now the four things are set up checkpoint positive negative prompt and dividend height ratio and the next thing what we need is that so after this thing the next thing that we need is sampling methods sampling steps CFG scale and all of that thing so in order to get that thing just right click over here add node and in that you will find sampling and in that just click onto and go on to K Samplers so basically this is set over here you can see there is seed going on steps going on CFG scale sampler name and all of this thing going on this is totally set up over here almost we need all of this thing in order to generate a basic image right so the next thing after this thing the next thing that you need to add is the decode vaed code so in order to add the decode just go on to latent over here and add decode over here so in order to decode your image I don't know why it is generating this image but yes you can add this vaed code in order to decode your image and the last thing that we need is the output image right so in order to get the output image just go on to add node and go on to image in that just select save image over here so you will get the output result over here so these are the basic thing that we need to go with the text to image section right now I have not connected any of the things I will explain in detail but first what we will do we will try to add basic things as well if you want Laura then you can add Laura as well so in order to add Laura just go and to add and in add you can go into loaders and you will find Lora loader over here so this is a Lora loader over here you can click on to Laura loader and just press on top of over here so that's the Lora loader and the next thing that you need to add if you want to use vae as well and on Lotus you will definitely find vae loader as well so this is the last thing that that we will use I don't know why these are generating images but yes that's the only thing that you need to do so we will place it over here the reason I will explain you so now guys let's start connecting all of these nodes so let's begin so so the first thing where we'll do is that we will start connecting the model over here so what we will do we'll start connecting the model you will find the model on the case Samplers over here you can adjust this as well wherever you want but yeah we will go with the first thing and yes model is connected over here and the next thing that we will do we will try to connect this clip over here and click on the negative part as well so that's how you can go and clip this as well so basically positive prompt and negative prompt is now connected with the checkpoint loader and now the next thing what we will do we need to connect the positive prompt and negative prompt as well right so what we will do we will try to connect positive prompt to positive on the K sampler and negative on negative sign plus so that's how these are connected and now width and height you need this as well right so what we will do we will try to connect this on to latent image you will see latent into latent image of the K sampler okay I will explain the lower one towards and the next thing that you need to do is that that you just need to connect the latent of the K Samplers on this vaed code so after connecting the output of the K Samplers onto vaed code the next thing is the final thing that you need to do is you just need to add the image on the save image section so basically text to image section is complete and first we will try to generate image and let's see what result do we get now okay guys so now my prompt is ready and now let's try to generate and let's see what result do we get so in order to generate the image you just need to Q prompt over here let's click and let's see what result do we get okay so now we got the error prompt has no properly connected outputs and required input is missing vaed code we forgot vae to connect so what we will do we will try to connect vae on this vaed code so now let's try to generate and let's see what result do we get so now guys as you can see definitely we got the result over here but this is not a great result so that's why this is basically a text to image section and the reason to this I will explain you the reason to this is that some model on confi UI doesn't work without vae so that's why I made these videos with vae loader as well and with Laura loader as well so now what we will do we will try to connect VA e loader as well and Laura loader as well so in order to connect this you just need to connect this vae on this section so we will connect this vae on this section and you will see this vae is disconnected on checkpoint loader and now the next thing what we will do we will try to connect Lora loader as well so what I will do we will try to connect this model you can see this model from over here and yes this model will go to this section and after connecting model to model over here the next thing what we will do we will try to connect over this clip to this as well so basically checkpoint to Laura loader clip and the next thing let's go to this as well and let's connect this clip to this as well so basically now you can see see this is connected Laura is connected and yes Laura is connected to positive prompt and negative prompt as well so that's how you connect and we will simplify this first of all so that's the simple method that you can go and now VA is also is connected and now let's try to generate and let's see what result do we get so now guys as you can see after connecting the Lora loader and vae loader as well so this is the final result that we got don't consider this is a result I don't know this is something bug but yes this is the final result that we got so that's why I made this video with vae loader as well so that you can understand because some of the models don't generate images just like I am using Alpha mix art and arrows also don't generate a result without VA so that's why you need vae as well on text to image section that's how you can generate this and if you want to save this workflow it's very easy to save it as well so just click on to save and yes you can save it I have already saved my workflow over here you can see simple text to image workflow so you can save your workflow and yes then you can directly drag on to confi UI workflows so it will definitely load again so that's how you can go and create and if you want to rename it like if you want to group it then yes you can create the group as well like this you can group and yeah you can just create the group so what group does is that you will be able to move all of this in a single group like it's totally connected all of this are connected so that's how you can create your own workflows so this was a basic text to image workflow with Laura loader and vae loader as well so that's why I made this video in order to like go with the basic and create some amazing workflows just like this this was the control net workflow that was created by one of my friend on Discord server you just need to drag the workflow and you can see yes definitely control net workflow is loaded over here thanks to one of my friend who is creating this video let me thank him once again you can see over here you can definitely get amazing workflows from here this sections on my Discord server you can join and get this workflows guys these are amazing work so that's it for today I hope so guys you like this video and yes that was a text to image basic workflow that you can go and generate and yes that's it guys for today and I will be back with more and more videos tomorrow till then have a good day bye guys love you all guys bye guys

Channel: CHILDISH YT

Views: 7,433

Keywords: stable diffusion, stable diffusion tutorial, comfyui, comfyui guide, ai, ComfyUI Explained, stable, diffusion, Automatic1111 Comfy UI, ComfyUI Hi-Res, Hi-Res Fix Upscaling, Comfy Ui, Guide ComfyUI, ComfyUI Local Hi-Res, Local ComfyUI, ComfyUI Installation, ComfyUI Local Install, ComfyUI Vs Stable Diffusion, Stable Diffusion Vs Comfyui, comfyui basics explained, comfyui nodes, nodes ComfyUI, ai ComfyUI Hi-Res, comfyui tutorial, ComfyUI Hi-Res Fix Upscaling, HiRes Fix Comfyui

Id: obbam4tu3D8

Length: 10min 23sec (623 seconds)

Published: Sun Mar 19 2023