I found better nodes to use with art styles, that is more accurate, so I updated the workflow in the Ep07, check it out here ruclips.net/video/Xsx-u0OMezw/видео.html
I have been using SDXL Prompt Enhancer ... but swapped to your file today. It is very much appreciated. Civitai User 0x00f is going to upload images with an adopted workflow.
Hello, and thank you for the work you provide on your channel. I spend a lot of hours on RUclips to learn COMFY UI and frankly you are an excellent teacher. I repeat myself but again a thousand thanks. I always look forward to your videos! A Frenchman from Paris!
I was out and saw a notification that the new episode was up. As soon as I got home, I watched it right away. I love this channel. Everything I learned about generating images with AI started from here.
There are some tutorial videos in Japan. comfi-ui took only 4 times. And since the “basic” section is followed by a customization section by the video distributor at the beginning, I didn't learn the basics at all. This video series is really easy to understand! In one video, it teaches you step-by-step, from the basics, step-by-step, in more and more depth, theme by theme. This is exactly the series I needed!
just gonna say a big thank you, you have the best tutorials for a beginer, the best ive seen, keep up, you have answers that i didnt ask for but needed.
I guess I'm just binge watching these at this point and stopping only to figure out what I did wrong... XD Thank you again for the detailed and useful tutorials.
I wish there was something like this for character creation. For designers and for those who just create beautiful pictures as a hobby, your videos are very useful.
it's kind of wild that comfyui makes you download all these extra things to do extremely basic functions like using styles! In my opinion the manager and some of these requirements should be included in the initial installation. Nice video though! I'm new to ComfyUI to try out the new 3.5 models. It appears the styles are now working with my Lora but not sure yet if there's anything in my workflow that could be improved. Seems like ComfyUI has so many options it's a bit overwhelming but satisfying when it's working haha.
What you do is fantastic. I can't believe that you make these videos. It is unique and priceless. I have a a question and a note. Is it possible to use more styles in a simpler way? It seems overcomplicated comparing to Fooocus or Forge. It is not easy to understand the difference between LoRA and styles. Ultimately they both are kind of styles. It would be great to watch a video when they should be used. How can we use more than one LoRA? Thank you again for your efforts and you give the people - and me.
check the episode 07, used different method, there the first art style is more important then second, and so on, choose first the art you want to have more influence
You are doing a great job and it was easy to follow. I got only one issue: when I'm looking for a module I have never found the "search", everytime I have to scroll manually, how can get the search?
Great work, I followed you after watching your "Who am I" video. What would be the advantage of this workflow over something like the custom node; sdxl_prompt_styler
I didn't use promptstyler so far and probably are many way to do the same thing. I am just recreating the workflows from forge and a1111 in comfyui, so for me it was more the ability to get my styles file working in comfyui
FYI; If you have another CSV loader installed by PCMonsterx (ComfyUI-CSV-Loader) you need to remove it due a conflict with Pixaroma's recommend CSV loader.
Salut! Multumesc mult pentru aceste videouri 🙏 Se poate un video cu "prompt expansion " ca la Fooocus? Adica introduci ceva simplu si sa genereze ceva mai complex :)
Hello and thank you so much for your video! I have question. If we are merging the styles, how we can control of weight style1 and style2? From your video the weight of styles is equal?
Check the episode 7 that was just published, on that new method the style you put first is more important. There is another method to use the conditioning average node instead of conditioning concat, that way you can have control betterment 2 prompts using values from 0 to 1
I will see in later episodes if can be done, but exactly the same will be difficult unless if trained a lora model or something with that characters so it knows how it look like.
The best way is to have a lora trained for that model, other includes control net, ipadapters and so on, somehow Ai love to generate random stuff 😂 so is make it harder to keep it consistent, also there are extension like reactor. In the future if i make it to work right i will do a video, on some extension i have problems with extra dependencies that needs that didn't install right, so i need to do more research and i try to take the episode in a order, starting simple like letting people generate and under some basic nodes and then slowly adding custom nodes to allow changing like changing a style a face and so on
only need node that show those texts so we can see what styles are. to learn. just like in A1111 i mean forge lol we can move that style directly to prompt and change it. there was to (print) node atleast default
Pixaroma, I think your flaws are on the lack of hashtags or something. I really don't understand why having so great videos, your channel hasn't gone above 100k suscribers, at least...
Can you check episode 7 for an alternative way to use it? Also ep15 show a variation . You can post some screenshots on discord on comfyui channel to see if we can find the cause if ep7 doesn't work.
Hi man. I'm running comyui in a kaggle notebook because my computer's gpu is not nearly enough powerful to run comfyui locally. The problem though is that the only notebook that I found that was actually working doesn't have the comfyui manager. I've been trying to install it for days but I keep getting errors, so I wanted to ask you: is there a way to import this file that you provided withouth using the manager?? Because I'm really going crazy, I have no programming background. Thank you in advance.
I am not sure how that works online, since i use it locally, you could try this to see if this comfyui installation works, that comes with all the nodes and manager too github.com/Tavris1/ComfyUI-Easy-Install , i dont know any other methods, not sure if copying and pasting the manager folder works
Check the link from the RUclips description the one where is about page and subscribe, the old link expired and added a new link but couldn't change on all videos the Link so try that one
Problem : when I use the art styles it all works great. Thanks. But if I want No Styles, that option is not there. So there is always a style active. My solution is to set the styles node to “bypass”. Any ideas on that? I tried to edit the csv file but that screwed up the file….
Try to use maybe the method from episode 8, but if you just downloaded the file and didn't edit it should work, when you edit the fike usually it messed up, if you just edit the file with notepad, don't use Excel, you can change the prompt between quotes. No styles is just a style with empty prompt
@@pixaroma when I downloaded your csv file, somehow the “no style” line was missing. So I downloaded it again, rechecked it, and now it’s working fine 👍🏼 Thanks 👍🏼
the most recent link should be in the channel link section discord.com/invite/gggpkVgBf3 it was set to not expired, but somehow still expired so I changed on some of the older videos but could not change it on all
On.episode 08 probably at the beginning of next week, today i get an update on episode 6 with more accurate styles workflows, so that is 7, and after that i work on flux on 8
![Shoot EPIC MOVIES with this FREE AI Tool! [ComfyUI Tutorial + Free Workflow]](http://i.ytimg.com/vi/gHI6PjTkBF4/mqdefault.jpg)
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
I found better nodes to use with art styles, that is more accurate, so I updated the workflow in the Ep07, check it out here ruclips.net/video/Xsx-u0OMezw/видео.html
I have been using SDXL Prompt Enhancer ... but swapped to your file today. It is very much appreciated. Civitai User 0x00f is going to upload images with an adopted workflow.
Your tutorials are some of the best on youtube And I've watched a lot of them coming from A1111 to comfyUI. They are so easy to follow. Thank you.
I just started with ComfyUI three weeks ago. This video is pure gold!!! Thanks!
Hello, and thank you for the work you provide on your channel. I spend a lot of hours on RUclips to learn COMFY UI and frankly you are an excellent teacher. I repeat myself but again a thousand thanks. I always look forward to your videos!
A Frenchman from Paris!
merci :)
Thank you for doing this series! It's very helpful to an Auto1111 guy who has resisted the Comfy workflow. Much appreciated!
I was out and saw a notification that the new episode was up. As soon as I got home, I watched it right away. I love this channel. Everything I learned about generating images with AI started from here.
Thanks for keeping this tutorial up to date by adding new discoveries in your following videos. You just go the extra mile here.
There are some tutorial videos in Japan.
comfi-ui took only 4 times.
And since the “basic” section is followed by a customization section by the video distributor at the beginning, I didn't learn the basics at all.
This video series is really easy to understand!
In one video, it teaches you step-by-step, from the basics, step-by-step, in more and more depth, theme by theme.
This is exactly the series I needed!
thank you, glad it helps
this is magical dude. I was ambiguous about a ComfyUI implem so far, but you've turned me around. great series, thanks for all the tips!
Glad I could help :)
Thanks!
Thank you so much ☺️
Wow, this was mind-blowing! Above my experience level however good to know what's possible.
just gonna say a big thank you, you have the best tutorials for a beginer, the best ive seen, keep up, you have answers that i didnt ask for but needed.
Thank you! Great tutorial.
So glad I found your channel! Thank you for your time!
Glad I could help ☺️
Thank you for the file. Subbed and liked!
I guess I'm just binge watching these at this point and stopping only to figure out what I did wrong... XD Thank you again for the detailed and useful tutorials.
I wish there was something like this for character creation. For designers and for those who just create beautiful pictures as a hobby, your videos are very useful.
Thank you very much for the tutorial; it’s very comprehensive, and I’m learning a lot. Please keep it up!
Very excited to see this. Can't wait for the flux tutorial. Thanks very much!
WOW MAN! I am literally in tears. thank you, thank you!!!
thanks :) make sure you check ep07 and ep 15 for an update on that :)
Thank you!! I watched your previous 300 styles video and was really looking forward to this video because I use ComfyUI 🙏🙏🙏
Thanks, glad you liked it :) in the episode 7 i updated it so it works even better with styles, even if the prompt is short or one word
it's kind of wild that comfyui makes you download all these extra things to do extremely basic functions like using styles! In my opinion the manager and some of these requirements should be included in the initial installation.
Nice video though! I'm new to ComfyUI to try out the new 3.5 models. It appears the styles are now working with my Lora but not sure yet if there's anything in my workflow that could be improved. Seems like ComfyUI has so many options it's a bit overwhelming but satisfying when it's working haha.
Very good tutorials. I will recommend your channel to others. Keep up the good work.
thank you 🙂
I've been waiting for this one! Thanks so much!
First of all, thank you. This is a really cool workflow video. I'll subscribe and like it. Thank you for watching this great video.
Great video, and I loved everything about it. I can't wait to download these images.
Thank you very much, I really appreciate your sharing with the community.
Great video! Thank you very much for working the Styles into ComfyUI.
Thx for this great work. I'm a huge fan !
great explanation!
great, a lot of valuable information besides the main topic: concat prompts, create a group among others!
easy to understand tutorial, thank you and keep up the spirit. very useful for me. :)
excellent tutorial, thank you for the art style 🙏
Nice tutorial, thank you 🎨 🖼
i learn a lot . . . thanx a lot bro..... easy to understand :) :) and i like the tutorial... nice
Thank you!
Incredible. Thank you so much!
What you do is fantastic. I can't believe that you make these videos. It is unique and priceless.
I have a a question and a note. Is it possible to use more styles in a simpler way? It seems overcomplicated comparing to Fooocus or Forge.
It is not easy to understand the difference between LoRA and styles. Ultimately they both are kind of styles. It would be great to watch a video when they should be used. How can we use more than one LoRA?
Thank you again for your efforts and you give the people - and me.
I will do more videos about styles and diff node more easy next week, you can use power lora loader from rghtree that let you add multiple Lora
@@pixaroma I am not sure that I understand it, but I will see :) Thank you very much.
Enter on discord there i can help you better@@CsokaErno
it's so powerful!
Thank you so much!!!
Thank you for this tutorial
Great tutorial! Very easy to follow. Is there a way to weight the contributions of the art styles?
check the episode 07, used different method, there the first art style is more important then second, and so on, choose first the art you want to have more influence
'prayers were answered'😁
great channel
great tut!
While copy pasting the node hold down the shift key to retain the connection :)
Yeah i know that, I guess i am enjoying making the connections by hand 😂
You are doing a great job and it was easy to follow. I got only one issue: when I'm looking for a module I have never found the "search", everytime I have to scroll manually, how can get the search?
Scroll the mouse wheel down, and it should appear on top filter list where you can type a word to search
love it !
does this work with flux shnell ? i am new to these stuff
I connected it to flux but need some changes to workflo, i will do in next episode the flux and will fo a version of workflow with styles also
@@pixaroma nice thanks
great
Great work, I followed you after watching your "Who am I" video. What would be the advantage of this workflow over something like the custom node; sdxl_prompt_styler
I didn't use promptstyler so far and probably are many way to do the same thing. I am just recreating the workflows from forge and a1111 in comfyui, so for me it was more the ability to get my styles file working in comfyui
FYI; If you have another CSV loader installed by PCMonsterx (ComfyUI-CSV-Loader) you need to remove it due a conflict with Pixaroma's recommend CSV loader.
In the episode 7 I used another one instead of that loader. So that should fix it
Nice!
Salut! Multumesc mult pentru aceste videouri 🙏 Se poate un video cu "prompt expansion " ca la Fooocus? Adica introduci ceva simplu si sa genereze ceva mai complex :)
Vezi ep11 si 13 cu LLM si ollama ☺️
@@pixaroma ok mersic, abia astept sa ajung acolo 😊Sunt la ep 7 si bineinteles , cu liked si subscribed! cheers 👍👍
Hello and thank you so much for your video! I have question. If we are merging the styles, how we can control of weight style1 and style2? From your video the weight of styles is equal?
Check the episode 7 that was just published, on that new method the style you put first is more important. There is another method to use the conditioning average node instead of conditioning concat, that way you can have control betterment 2 prompts using values from 0 to 1
@@pixaroma Thank you so much
Hoho! I see you are part of Olivio Sarikas discord as well!
Yeah i am part of many servers but don't always have to check it all :) but yeah I like Olivio videos as well
Great tutorial! And it would be amazing if Flux could somehow integrate this solution!
I am already working on it, on episode 8, will be about flu, and the style file also works with it
@@pixaroma Very good news! Flux is my favourite and I really miss this style feature! I'm really looking forward to it!
thanks
How about adding a Random promt to select the 2 different styles
Not sure but in 45 minutes i have an update version with better nodes for it in ep7
Love the course so far! I´m getting "Warning: Missing Node Types" although the styles.csv is in the same folder as main. What am I doing wrong?
Can you watch the episode 7? I changed some nodes so it works better, and need to do some settings in config file to work. Thanks
@@pixaroma Hey, yeah I did that too but they won't load there either, just turns out as "empty". Doing all the steps with double//, "path"..
@@danielblom4990 mention me on discord, and show me screenshots, to see what you did, sometimes a missing comma can mess up everything
Hi, i have a draw of my character and i want to redraw it with ai in some another poses, do you have any guides how to do it?
I will see in later episodes if can be done, but exactly the same will be difficult unless if trained a lora model or something with that characters so it knows how it look like.
I have a question How to generate consistency character for story in comfyui please make a video about this,
The best way is to have a lora trained for that model, other includes control net, ipadapters and so on, somehow Ai love to generate random stuff 😂 so is make it harder to keep it consistent, also there are extension like reactor. In the future if i make it to work right i will do a video, on some extension i have problems with extra dependencies that needs that didn't install right, so i need to do more research and i try to take the episode in a order, starting simple like letting people generate and under some basic nodes and then slowly adding custom nodes to allow changing like changing a style a face and so on
Thank you! Great tutorial. I just have issues joining your discord group.
Use the link from RUclips channel header that should work, or from recent videos, make sure it doesn't have a space at the end of the link
only need node that show those texts so we can see what styles are. to learn. just like in A1111 i mean forge lol we can move that style directly to prompt and change it.
there was to (print) node atleast default
I use show any node and can see the text on those. Are just prompts in text format
Pixaroma, I think your flaws are on the lack of hashtags or something. I really don't understand why having so great videos, your channel hasn't gone above 100k suscribers, at least...
I agree
After I installed csv, csv style shows loading error. how can i solve it. Thank.
Can you check episode 7 for an alternative way to use it? Also ep15 show a variation . You can post some screenshots on discord on comfyui channel to see if we can find the cause if ep7 doesn't work.
@@pixaroma thank you for your reply. I will check it!!!
Hi man. I'm running comyui in a kaggle notebook because my computer's gpu is not nearly enough powerful to run comfyui locally. The problem though is that the only notebook that I found that was actually working doesn't have the comfyui manager. I've been trying to install it for days but I keep getting errors, so I wanted to ask you: is there a way to import this file that you provided withouth using the manager?? Because I'm really going crazy, I have no programming background. Thank you in advance.
I am not sure how that works online, since i use it locally, you could try this to see if this comfyui installation works, that comes with all the nodes and manager too github.com/Tavris1/ComfyUI-Easy-Install , i dont know any other methods, not sure if copying and pasting the manager folder works
I try to got he discord link for the workflow but got an error message - Invite Invalid, any idea why?
Check the link from the RUclips description the one where is about page and subscribe, the old link expired and added a new link but couldn't change on all videos the Link so try that one
@@pixaroma thanks
Discord link is not working
Try the the link from RUclips channel header that should work all the time
Problem : when I use the art styles it all works great. Thanks.
But if I want No Styles, that option is not there. So there is always a style active.
My solution is to set the styles node to “bypass”.
Any ideas on that?
I tried to edit the csv file but that screwed up the file….
Try to use maybe the method from episode 8, but if you just downloaded the file and didn't edit it should work, when you edit the fike usually it messed up, if you just edit the file with notepad, don't use Excel, you can change the prompt between quotes. No styles is just a style with empty prompt
@@pixaroma when I downloaded your csv file, somehow the “no style” line was missing. So I downloaded it again, rechecked it, and now it’s working fine 👍🏼
Thanks 👍🏼
@@2008spoonman glad it worked ☺️
the worflow link is dead ..
The one from the header should always work discord.com/invite/gggpkVgBf3
can't get access to your discord channel
Check the link from the channel header or a recent video, let me know if it works
@ yes! got it
Can you update with a valid discord join link? As of 2024-10-09 this one doesn't work anymore.
the most recent link should be in the channel link section discord.com/invite/gggpkVgBf3 it was set to not expired, but somehow still expired so I changed on some of the older videos but could not change it on all
flux when?
On.episode 08 probably at the beginning of next week, today i get an update on episode 6 with more accurate styles workflows, so that is 7, and after that i work on flux on 8
either iam banned from ur discord that iv never joined, or I dont know what it is named.
Is not in public search, is named pixaroma, and I didn't ban anyone yet try this Link discord.com/invite/gggpkVgBf3