Video

@odrammurks1497 "Wow, I guess my video really blew your mind...into pieces! 😂 Don't worry, no brains were harmed in the making of this content. But if you ever need help putting your mind back together, feel free to hit me up!"

@mr.entezaee Hmm, I tried to make the workflow as user-friendly as possible. You can leave out or delete functions. What do you have in mind? It would be possible to simplify the first part of the workflow up to the sampler. The other modules can then be easily removed or replaced by anyone. Is that what you mean?

@@murphylanga It may seem a bit busy for regular and novice users like me. I don't know, maybe there is a simpler workflow to take us from a low quality photo to a very high quality photo. I have seen a few in the past. But I am always looking for new methods. And yes, your method was new.Thankful

@mr.entezaee I have simplified the workflow for you. If you want, you can also delete the upscaler groups. Just remove the nodes in the individual groups. You can download it here: civitai.com/models/525976?modelVersionId=584454

Thanks for the info, ClipChamp is too slow for me, I only do it on the side. If the music bothers you, turn the sound off.

@Pashkee123 Thank you very much, I was very pleased with your comment

Thanks

I'm delighted you had fun. Have a great vacation. I'll probably slow down a bit too. But, so that the end of the vacation isn't so hard, the next thing I want to put in is the control net. It might just be finished when you get back.

@@murphylanga I know, the thought of going back to work :( Your workflow works very well now, I look forward to the controlnet. Its last bit of control needed :) You sound like you need a rest too

@Dreddoom :-) no I don´t.... :-) that's probably just my bad English...and I have a few more ideas, but they won't be implemented until after the controlnet

I always like more tools to create the exact image I'm looking for. Sounds good 👍

Ne, nicht wirklich, sind alles einzelne Module. Wenn man das Prinzip mal verstanden hat, ist ziemlich einfach. Da Gute ist durch die RGthree Nodes kann ich alles abschalten und gewinne Performance. Ich habe noch einen ganz großen, der ist aber niemand zuzumuten. Von dem baue ich Stück für Stück als austauschbaren Modul um... Leider habe ich nicht die Zeit die ich gerne hätte.

@@murphylanga Ich meinte mit der Pflege auch eher, wenn sich bei einem Update von ComfyUI und den Nodes mal wieder alles ändert. Das sind ja alles noch keine Stable Releases. Da wird mal ne Variable geändert oder etwas hinzugefügt ohne es mit einem sinnvollen Default zu unterlegen usw. Ich finde deinen Workflow mehr als beeindruckend. Könntest du dir eigentlich Vorstellen Workflows als Auftragsarbeit zu erstellen?

@jurgenmayr8164 Mit der Zeit weiß man welche Nodes "wacklig" sind, oder sich einfach ohne Doku ändern. Aber es stimmt schon, es kann Überraschungen geben. Diese Woche zum Beispiel hat ein Entwickler ein Update veröffentlicht, das den kompletten Workflow lahmgelegt hat. Ein Anwender hat mich angeschrieben, den vorab eine Version geschickt habe, dass nichts mehr funktioniert. Ich hatte schon etwas gebraucht, um den Fehler zu finden. Meistens, ist Comfy aber bei der Fehlersuche behilflich. Ich habe den Entwickler kurz angeschrieben und 15min später war das Problem gefixt und ein neues Update veröffentlicht. Einige glücklichen Zufälle, sodass fast niemand das überhaupt mitbekommen hat. Vielleicht ist auch die Erfahrung aus meinem "richtigen" Job hilfreich, da beschäftige ich mich ab und an mit parametrischer 3D-CAD Methodenentwicklung und Automatisierung und schule Kollegen. Es gibt immer mehrere Lösungsmöglichkeiten und Bedenken habe ich nur, wenn sich etwas Grundsätzliches im Zusammenspiel mit Comfy und den Rgthree nodes passiert. Zu deiner 2. Frage. Ich hatte bei OpenArt letztes Jahr einen Anerkennungspreis gewonnen. Und wurde auch angefragt, ob ich Interesse hätte etwas für sie zu machen. Irgendwie ist das im Sand verlaufen, Spamfilter oder keine Ahnung. Ich habe das aber auch nicht weiter verfolgt. Grundsätzlich kann ich mir das schon vorstellen. Aber meine Zeit ist begrenzt.

Hm, operating this workflow is easier than operating Automatic1111. But of course ComfUI is more complex than Automatic1111. But that wasn't your question. I could write because there are many more options. I have Automatic1111 and also Foocus installed on my computer. But I hardly ever use it. Because, and this is my reason, Comfy is much, much faster for me than the other tools.

Hello, I can understand that very well. I also have lots of little ones :-), but at some point I lost track of which ones I need. And then it's too much work for me to keep them all up to date. I once packed more into a workflow, but that wasn't practical. I tried to make everything switchable and interchangeable in “modules”. So there is one with a focus on IP-Adapeter, one with a focus on Styler, one with Inpaint/Outpaint... Everyone has to decide for themselves. You can also pick out individual modules.

@@murphylanga Understood!

Yes, you are right, but I would like to write the information of the Lora models in the generated images. So far this only works with these special Lora loaders. This means that this information is also automatically linked when uploading to e.g. CivitAI. But, you can easily rebuild this, the workflow is modular. At the top of the upscale groups there is what I consider to be the best compact Lora loader from the Rgthree nodes. You can install it, for example. The developer may even customize this node. Then I'll convert that too.😉

Thank you for your positive feedback! I'm glad you find the image to image options in workflows valuable. It's great that you're looking forward to trying out this workflow. I'll continue to incorporate image to image options in future workflows. We've realized that sometimes, a slower pace can actually lead to better results. In the world of content creation, less can sometimes be more, and we're learning to embrace that philosophy. ;-)

@@murphylanga Hi murphy, the workflow looks great. I'm having trouble with the checkpoint loader preview. it fails to load. I've updated it in comfy and manually, and reinstalled comfy several times. Its happening in all the workflows that have it. I got around it by putting a standard checkpoint loader on its own. If I right click on the checkpoint it loads the info. Do you know if there's a problem with it? Thanks

@Dreddoom Hello, the “Checkpoint Loader Preview” node is only responsible for the preview of the checkpoints (images from CivitAI or your own) You have to land the checkpoint in the node below: “Checkpoint Selector” I hope you can solve your problem with this information

@@murphylanga Thanks Murphy, but it fails to load and is just red. It doesn't matter. Thanks for helping. The workflow works great though. keep up the good work

@Dreddoom Ah ok, it can only be related to Pythongossssss/Comfyui Custom Scripts. Have you ever tried to uninstall them and install them again? I don't have the mistake, I checked that in different versions. And also all completely updated ... However, I discovered this mistake: failed: comfyui nodes for inference.core. You can put the filter on import failed in the manager under Custom Models and see what is. I deleted this node with me, I can't remember what it would need for.

Thank you! I'm glad you enjoyed the video and found the results impressive. I'm excited for you to try out the new version with the checkpoints and loras we discussed, especially for impressionist-inspired results. Remember, experimentation is key, so feel free to play around with different settings and models to discover your own unique style. By the way, I came across this checkpoint that might interest you: civitai.com/models/240154/painters-checkpoint-oil-paint-oil-painting-art-style?modelVersionId=289591. It seems to work well with various art styles and might complement the Loras we discussed, especially for oil painting-like effects. Give it a try if you haven't already!

Hello @hatonafox5170 , thank you. Thanks also for your question. First, try to downgrade Denoise very low in the Ultimate Upscale group. 0.01 is possible and changes virtually nothing to the image. This is a good way to work for paintings. A second option is to use other upscale models with the Ultimate SD upscalers and the Highresfix. For example, I wouldn't use Ultrasharp, it's good for photos. Try the 4x_NMKD-Siax_200k or 4x_foolhardy_Remacri which can also be downloaded via the manager. There are plenty of others too. Please write to me what you found out. I can also search a little deeper. Best regards and have fun Murphy

Thank you for your suggestions. I will give it a try and let you know how it goes. Looking forward to your next videos and the addition of ControlNet to the workflow!

@@murphylanga So I gave your suggestions a try. I'd say the results were improved but it still doesn't look believable that the images produced were painted. I guess I'd describe the results as being simply a higher resolution version of a slightly blurry or unrefined image. Midjourney actually does a good job at creating oil paintings with impasto effects or clearly defined brush strokes that feel even more real after upscale (most of the time) but I can't get anywhere close to that using stable diffusion. I'm relatively new to this so this might just be a skill issue on my part but I've seen this question a lot and none of the answers I've seen produce realistic results. Any tutorials or workflows you could share would be a big help, I'm sure to many people who like this art style.

I can imagine that what they have in mind is not so easy. Achieving the quality of Midjourney is certainly not easy. Nevertheless, I believe it should be possible somehow. Most people want photorealistic images, so what you and I would like to do is not so common. It might make sense to put a special Lora model in front of an upscaler again. Try to add an artist with your desired painting style to the prompt and possibly enhance the prompt. Maybe an image to image (is like a latent upscale) will also help. There will be a new video tomorrow in which we will go into this in more detail. Or a suitable upscale model. Here is a page where you might find what you are looking for. But I don't really know the site and can't say whether the models there are all “safe”. openmodeldb.info/ At the moment I want to realize some other things first. But I am also interested in what you have in mind. Best regards and thanks for the information.

@@murphylanga thank you for the additional information and feedback. I’ll give those ideas a try as well. Looking forward to your next video.

Thank you for your kind words and for appreciating my painterly image generations through text2img! I'm glad you enjoy my work. Regarding your inquiry, I've already published workflows that can help convert realistic portrait photos into painterly images without distorting facial features. You can find one on CivitAI, and on OpenArt, you can check out the Ultimate V9 Workflow. However, please note that the Ultimate V9 Workflow is quite complex and may be challenging to use. Additionally, I've also contributed to a workflow on OpenArt that serves a similar purpose. You can find it here: openart.ai/workflows/tamarin_extra-large_98/sdxl-basic-image2image-face-optimiztion/5ki2QaHfXtbOyVFIQjXP I'm currently working on a new workflow in the Automatic Modular series, which will be easier to use. It's almost complete and is scheduled for release next week. However, please note that it won't include ControlNet support upon release. I appreciate your patience, and I hope you continue to enjoy using these workflows!

@@murphylangaThanks for your reply. I've watched almost all your youtube clips to find out how to convert portrait pictures to painterly pictures without distorting facial feature. I've downloaded V9 workflow but I have even no idea how to start. Your workflows are too complicated and too large to average users. However if you have some time, please try to teach on this topic ( img2imgToPainterlyStyle) and please smaller your workflow if possible. I wonder if comfyui can really convert a people portrait to how much beautiful painting photo.

Hello, next week I will provide something that is very easy to use apart from installing the @suthamchindaudom6410 comfyui-reactor-node. It's not perfect yet, but I think you'll like it. Not the next video, but the one after that ;-) If you were able to open the Ultimate V9 without errors, then you have already solved the problem with installing the comfyui-reactor-node. For further optimization, I first have to build the modules for Controlnet and Face IP adapter. However, I think that you can have the facial features best with the comfyui-reactor-node. A mixture of everything will then be able to optimize it. But I can't get to that at the moment.

@@murphylanga I appreciate your contributions. When your new workflow finished , please let me know. Thanks

Hello @suthamchindaudom6410 , a new workflow has been published. The faceswap module is still missing to better achieve your goal. But the image to image function is already built in. Best regards

Hello, this workflow will not run smoothly with 4GB, and if it does, it will be extremely slow. However, ComfyUI should still function properly. In that case, you should use SD1.5 models instead of SDXL models for optimal performance. I create these workflows primarily for my own use, but I make them available to anyone who wants to use them. However, I cannot guarantee that they will function on other systems or environments. If I were to provide such a guarantee, it would require significantly more time and investment from me, which I currently cannot afford. I hope you understand my position. Sorry that I don't have better news for you. Best regards!

😊 So in my upcoming video, I'll be diving into the enigmatic world of Samplers and Schedulers in Image to Image🤯. I must confess, I'm no expert on these topics either, but hey, we'll figure it out together, right? 😉 And don't worry, I'll do my best to keep it fun and entertaining-even if it means throwing in some dad jokes along the way!!

@@murphylanga look forward to it 😉

Oh my gosh, you're making me blush! 😊 It's comments like yours that make all the hours of rendering and editing worth it. And hey, if you're laughing even through translation, my humor game must be on point! 🤣 Just remember, even I don't always know what I'm translating-sometimes it's a surprise to me too!

@laurencechen1111 , Hello, yes something has changed in the latest Comfy update. The easiest thing to do is, to delete the last node in the Model Patch group, which is now displayed in red. The function was only installed for testing anyway. I have now also uploaded an update for all workflows. I will perhaps publish an upgrade of the workflow today with some additional options. A node is also added to incorporate the pony models that are popular with some.

@mykaeld76 "We're thrilled to hear that you're enjoying ComfyUI and finding it even more fun and useful! Thank you for your kind words and happy creating!"

@Dreddoom "Hey, absolutely! We'll prepare a video explaining our model management setup-it's more about adjusting a few files than a complex process. Stay tuned for the details! Cheers!"

@@murphylanga thanks, that sounds great. always learning.

@Dreddoom I've saved you a guide with annotated texts under www.youtube.com/@murphylanga/community Take a look and see if it helps.

Thanks Murphy, I'll take a look

@marcelo.pereira That's fantastic to hear! Thank you for trying it out, and we're thrilled that you're pleased with the results! Your support means a lot to us!

Hello @bookssumsThank you so much for your wonderful feedback! We are delighted to hear that you enjoyed the video and found the content valuable. By the way, the workflows are available for free download

Hello @Dreddoom; Thank you so much for your wonderful feedback! We are delighted to hear that you enjoyed the video and found the content valuable. By the way, the workflows are available for free download.

Thank you for reaching out and considering me for assistance with your custom workflow. Unfortunately, I'm currently unable to accommodate individual workflow requests. However, I'm confident that with a bit of experimentation and utilizing available resources, you'll be able to find a solution. Best of luck with your project!

Hello, @hatuey6326 Normally the paths should be generated automatically. But if you don't have a D: drive, that could be the reason for your problem. Check the following: "9" on the keyboard: -> "Finish Output Settings" group. You can set the path to be used for saving there. "Set SD Prompt Path". Normally the path should be created automatically if it does not exist. You can also switch there if you want to use the TTN spokes, you have to enter the path directly in the black node. Or switch off both to test. "x" on the keyboard: -> "Settings / individual Steps" group. If you turn something on there, you have to enter the path under "Output Path". Look for the group in which a QR code is created, which is always in a different place, depending on the version. You also have to enter the storage path there. -> "Output QR code". If you want to find errors, use the "Follow execution" with the right mouse button. This is often helpful. I hope I could help you.

Hello new follower, delighted to have you! We were so busy creating a new video today that we didn't have a chance to check the comments. We apologize for that oversight. A warm welcome to you, and many greetings. Barring any unforeseen circumstances, we'll be introducing a new training video tomorrow

@@murphylanga As I'm doing faceswaps, I found that Reactor or more recently Fast Swap does this very well on 'realistic' photos. IP Adapter to copy a style but combining the two is a real challenge without going through Loras training. I hope to achieve this with your workflow and the developments of ComfyUI

@hatuey6326 Lora Training is also an option, but it costs time and computing power. My laptop with an RTX3070 is too weak for SDXL. I did it once, but that was over 30 hours of computing time. For 1.5 it works quite well in 20 minutes. But I no longer use it. You can achieve good results with Face Reactor. But you should improve them with e.g. Ultimate SD Upscaler or highres fix. What you have in mind also works. Small spoiler: I will probably put a video online today that shows how you can transfer parts from the Ultimat workflow into your own or a “starter workflow” provided by me. Translated with DeepL.com (free version)

"Hey there! Thanks for sticking with me through the craziness! I totally get where you're coming from, and I appreciate your patience. I'll definitely work on creating shorter workflows for simpler tasks to make it easier for everyone to follow along. Thanks again for your support! 😊🤖"

@@murphylanga Thanks again considering, you have already opened my mind alot :)

"Thanks a bunch! Your support means a lot. I'm thrilled you're enjoying the content! 😊 Stay tuned for more-I've got some exciting plans in the works and will release more when time permits!"

"Thanks a lot! I'm really glad you're enjoying the workflow I showcased in the video. I'll keep working on it to make it even better in the future. Stay tuned for more updates!"

@BolteDev Oops, embarrassing... Thanks for the tip, I corrected what was possible with CZcams. A completely new version because of an, albeit embarrassing, spelling mistake is too much effort for me.

@@murphylanga We all live and learn

@Rico-nj3vl Hello, I hope you have fun with it. In some areas it is important to switch different nodes or groups and also make the appropriate selection or to hang the noodle correctly. It would be possible to automate this, but I don't want to invest the time for it at the moment. The most important thing is that I'm thinking about reading notes in the workflow to understand the dependencies. Now have fun ;-)

Thank you for your comment, I was very happy when I read it. You recognized what motivated me to build this workflow. And it's still not quite where I want it. Your comment encourages me to keep going. If you notice something that isn't right, just write to me. Have fun exploring the workflow further.

Hallo, so ein Monster ist es gar nicht. Das schwierigste ist es, die Custom Nodes alle richtig zu installieren. Ich habe noch eine Variante veröffentlicht, in der ich Face Swap und den Übersetzter entfernt habe. (Das müssten alle CustomNodes über den Manager installierbar sein. Möglicherweise muss die Install.bat bei den WAS-Nodes noch manuell ausgeführt werden, um das Blip-Modell zu erhalten. (Oder ich habe das schon geändert, ich weiß es gerade nicht) Trau dich, trau dich, das Monster frisst dich nicht :-) Vorher ne Sicherung vom Python_embedded Ordner, ist immer ne Gute Idee, falls du die Reactor node installieren willst. Da gibt es ein extra Video von mir dazu. Manche jammern wegen der Änderungen des IP-Adapteres V2. Wer mit den alten weiterarbeiten will, wird in Zukunft keinen Spaß mehr haben, deshalb besser die Workflows umbauen. Ich habe das für den Ultimate V7 gemacht. Alter und neuer IP-Adapter in einer Comfy Installation wird nicht funktionieren.

@@murphylanga Vielen Dank Also ich habe die Installation gut und einfach über die Bühne gebracht und auch schon ein paar Tests gemacht. Ist doch alles noch etwas verwirrend und muss da noch etwas mehr Zeit investieren

a perfect couple 🤣

Hello, yes, that's interesting. Most of it could be automated in Comfy. In general, I do intend to implement something similar. However, there are a few things holding me back. Currently, my health doesn't allow me to sit at the computer for long periods. My laptop, equipped with an RTX3070, is too weak to handle such video workflows. I have to resort to low VRAM options, which makes everything unbearably slow. I still have plenty of other ideas for image editing. I don't know when, but I plan to invest in new hardware once I'm fully recovered. However, the timing remains uncertain. Sorry, I don't have better news for you.

@@murphylanga damn hope u get well soon and hope to hear from u soon :)

Hi, thank you :-) I'm feeling better again. I've done a bit of crafting again. When I'm really fit again, I should get something with more VRAM. I was able to set the SUPIR Upscale so that the 8GB VRAM works. You can watch this in my new video.

If I knew what you mean. I create workflows, take pictures with them, let an AI speak the texts and cut a video from them. Do you mean this?

@Nojlink Hello, I'm glad you like it. There is a problem with the model names if you have them in subdirectories; the SD Promt Saver Node cannot resolve this. You can also save with the other variant next to it. But if you want to save the information in the "Automatic1111 Format" e.g. for CivitAI in the image, SD Promt Saver Node offers better options. The way I solved it was that when I started Comfy I created the paths for the models (and rewritten even more) in small batch files. The English thing was intentional, but it was only supposed to be a short sentence, so I accidentally mixed it up a bit. I noticed that too, but didn't want to re-upload the video. Thanks for the tip

@Nojlink ich habe etwas eingebaut das das Problem beheben müsste, der workflow ist im Bild eingebettet: civitai.com/images/8230652