If you have any questions you can post them in Pixaroma Community Group facebook.com/groups/pixaromacrafts/ or Pixaroma Discord Server discord.gg/gggpkVgBf3 When you install manager if says git is not recognized you need to install git first git-scm.com/downloads Get ComfyUI from here: github.com/comfyanonymous/ComfyUI Get the ComfyUI Manager from here: github.com/ltdrdata/ComfyUI-Manager Download SDXL model - example Juggernaut XL - Juggernaut_X_RunDiffusion from here: civitai.com/models/133005?modelVersionId=456194 Download SD V1.5 model - example Juggernaut Reborn from here: civitai.com/models/46422/juggernaut Update: one of the members of our discord channel made an easy to install version with different nodes already installed github.com/Tavris1/ComfyUI-Easy-Install
I take it your a Graphic designer yourself, I see it through how you dedicate your time explaining each and every step. I have watched plenty of tutorials for all these AI tools and most of them are made by programmers/coders who run through it, assuming the entire wide world have knowledge in coding. Thank you so much for proving my point again, when it comes to teaching and understanding the audience better, it can only done by designers. All of your content is absolute gold.
One of the best things I saw in my phone notifications. Excellent, this is very valuable and I hope artists understand the value of this resource and reference. Thank you for taking the time to record this tutorial.❤
Got 5 minutes in and already appreciate you setting up the landscape of comfyui. I already installed it now, trying to digest and learn as a beginner. Seems like I will enjoy your tutorial much more than the last!
Finally, there's a tutorial solved my problem. I tried too many methods to install this.Like installing python ,viusal studio, Rust, git batch and changing function code in python file. its really useful and simplify my working. Much appreciated, it's very clear and concise,and say it again, Thank you!!!
The best video on internet for learn and understand ComfyUI. This is really the best guide to start with it. Everything in this video ist just perfect! I really recommend it! Thanks for it!
Thanks, this ComfyUI tutorial series is the best one I've found so far, Pixaroma simplified everything not like other "Tutorial Videos" tend to make things complex, his channel deserved more view and scribe, oh, and thumb up.
Fantastic tutorial thank you. After a year with a1111 and have a good understanding with it and 80% good results. I've often thought about comfy but the UI has put me off a little, but with your excellent tutorial series your doing this gives me a stronger interest in it. I'm looking forward to seeing more and giving it a try. Once again thank you.
This is by far the best introduction to ComfyUI that I've come across. As usual, I tried to go a bit beyond the tutorial and drag and dropped images from Juggernaut XL on CivitAI. The workflow is a lot more complex than the one presented here, but if you do want to give it a go, just make sure the files referenced (Checkpoint + VAE) are both compatible. You will also need to track down the upscaler file referenced.
Thank you! Wonderful tutorial!!! I am still trying to get my head around what is happening here. I definitely will be sticking with you as my guide in this new adventure.
Awesome video! Clear and easy to follow instructions as always. I am really hoping you are going to create a workflow for Sketch/Image > Design like you did before in Forge, that was awesome, really looking forward to see you replicate that in ComfyUI and show us how to do it or sharing the workflow! I am pretty sure you can build a pretty robust workflow for that with maybe control net/canny + "model training" or reference for the Sketch/Image to turn into something with similar style. Great work! Looking forward to the next episodes!
I am very happy I found you. This first lesson was very helpful as I just goy ComfyUI installed and working. The only thing I am missing is the menu in the lower right corner. At the top of my screen is Manager, which is the same as yours.
Yes comfyui got an update but all the buttons should be on the top left corner, where it say workflow and other menu, i put on discord in the comfyui-faq channel comparison between old and new menuu
Good to see you are replying to the comments even 5 months after posting. I just installed the ComfyUI for the first time, the UI is a bit different. I cannot see the extra options in bottom right. it just shows queue
The interface changed now is on top left bar where is the workflow menu and others, i put a screenshot before and after on discord with all the menu, look for comfyui-faq channel on my discord
Hello, Thanks for these actualized series of Comfy tutorial, I plan to watch them all. I have about two weeks playing with my 4090 in Comfy, i had compared your its/s and you are ahead of me for 10it/s more, how is that possible?? do you mind making a video to go over the hardware and tools you need to run efficiently Comfy? Ty.
Well the versions change and updates but it might be other things to consider, like for example i have 128gb so that might help to load the models in memory, i have ssd so it loads faster, i9 , so not sure how all influence, definitely vram helps the most but the interface change more things are added so maybe some things are working faster or slower , I don't know all the technical stuff 😀
Finally a more recent take at the whole process! I am so grateful you were so organized throughout it. Side note, I might have this fixed by the time you see it, but I haven't been able to even install ComfyUI yet. The unzipping keeps taking forever. It seems to be doing 'something', but I'm 30 mins in. Is that normal?
I don't think it should take so much, the file is big but it should take half hour. Did you try winzip or WinRAR to extract it mayne is because of something else. Try restarting windows, make sure the the zip archive is complete also maybe it didn't download all the way
@@pixaroma This is so frustrating. I tried all of those steps, and tried unzipping with 7zip (it seemed to not do anything after 1.5 hours) and also WinZip. The latter did work, but those .bat files are nowhere to be found. It's as if there's something wrong with those files from their website. Is there an alternative link for download by any chance?
Thanks for doing this. Do you know of a way to use the models folder from automatic 1111 in comfy? I have so many models and don't want to have to download them again just for comfy. Thank you!
Yeah there is a way to do that, try this mitsuhakitsune.com/en/2023/12/20/share-models-between-automatic1111-and-comfyui/ if that doesn't work i will look for exacts steps, im on mobile now, but i saw it being done by many people so there is a way to do it.
yeah i plan to do that in the future, I want to cover all that i learn, but i am starting with simple stuff, with image first because video is in fact a series of images, so we need to master the image creation first :)
I'm having trouble installig VLM nodes. There are some libraries to install. You had a video with how to do the requirements txt or individual modules in the txt file. Which vid is this? Thank you.
those nodes that use llama-cpp-python can give some problems on certain nodes, like in episode 11 some had problem installing the searge node. someone from community made a installer for comfyui and all the nodes I used so far, including the one with searge llm, so maybe it will work if has the rigth dependencies, or you can post on discord maybe someone that has more code knowledge can help, here is the installer i talked about github.com/Tavris1/ComfyUI-Easy-Install
when i try to update ComfyUI-Manager ComfyUI-Manager by cmd ; it message git' is not recognized as an internal or external command, operable program or batch file.
If you have a save node then is saved in the output folder and remain there unless you delete manually, if you have a preview node then is saved in temp folder and is deleted when you restart comfyui
I am not at computer to test it but check the comments on this reddit post like the last comment people replied that the method worked www.reddit.com/r/comfyui/s/Ll6CBHjZtk
you can get it from google drive , click download on top right corner and drag it in your comfyui interface drive.google.com/file/d/1gkcZ6En8bahsK1AmuGcpEGl4yDpTroBe/view?usp=sharing is the same workflow from ep03 that is also available on discord together with all other workflows
A model is often referred to as a checkpoint. The term "checkpoint" refers to a saved state of a model during or after training, which can then be loaded and used for generating images. So when you're loading something like Juggernaut XL, you would typically load it as a checkpoint file (usually with a .ckpt or .safetensors extension), which contains the trained model's weights and data.
Yes, but it might be very slow so not sure it is worth waiting so long for an image to generate instead of running the gpu bat you run the cpu bat when you start comfyui
@@MohdMobasherulIslam can you post some screenshot on the discord server, on the comfyui channel, to see what node is turning red discord.com/invite/gggpkVgBf3 usually that happen when it need a clip text model or is incompatible or missing some dependencies
Under the width and height on the empty latent node you have batch value that is 1, you just change that to 2 or what number of images you want and will generate that number. You can also use Auto Queue. I have talked about those options in some of the episodes at the beginning not sure if ep2 or 3
bro, when I install comfy UI manger then my laptop says 'git' is not recognized as an internal or external command, operable program or batch file. so what i can do
install git first git-scm.com/downloads I already had it installed and forgot to mention in video, just download for your operating system like windows version and install it and try again
Please can you help. I have been running into all sorts of problems with Comfyui manager. Every time I try use the manager to install or update I get this message in red : TensorRT Node for ComfyUI install failed: Bad Request and the command prompt says ...AppData\Roaming\Python\Python311\site-packages\torch\lib\cusolverMg64_11.dll" or one of its dependencies. I have tried every possible sollution on reddit and installed Conda, Python and every other version I can find but nothing helps. I even downloaded every Nvidia update, driver and tool I can find and still nothing. Any idea what is going on?
So the manager works but when you try to install the tensor node it gives you problems? Did you tried the update bat from comfyui update folder? All this complex errors more than i can handle. Worse case just install im other folder comfyui again. Other things chatgpt suggested: Reinstall CUDA and cuDNN: Ensure that you have compatible versions of CUDA and cuDNN installed. Sometimes a mismatch can cause issues. - Update or Reinstall PyTorch: Compatibility issues with PyTorch can also lead to such errors. - Rebuild the Virtual Environment - Manual Installation of Missing DLLs . Sorry i can help you more, but are so many things in place that is hard to know what could cause.
Hi, I would like to install Comfyui on my local disk, but my computer doesn't meet the installation requirements. Can you recommend a laptop that can handle locally? Please suggest the most affordable options. From this description, it's hard for me to choose a laptop that would handle it. Maybe someone has a recommendation?
I cannot recommend a laptop because i don't use any, only desktop. My recommendation is to get one with nvidiat rtx card and a lot of VRAM, 16gb of vram so you can run flux models, 24gb are quite expensive. Maybe you can do some slower with 12gb of vram, but lower than that only if you use sd1.5 model and some sdxl. But for flux I will go with 16 gb of vram
@@pixaroma hello, thank you very much for your answer, it's fantastic that you found time to answer, in that case a better solution would be to build your own computer, hmm, could you give me the rest of the parameters of such a computer so I could calculate what will be cheaper for me, thank you very much in advance
For Image generation what is important is to have a good Nvidia card with a lot of vram that will speedup the generation, 32 gb of system ram and SSD so it loads your model faster, and a processor that can handle the video card you choose, so as you can see the video card is most important for image generation, more vram and newer the video card is the faster you can generate
Forge ui is easier but the official is not updated anymore so i will probably not do tutorials for it angmore, the comfy UI is most active and can do a lot of things once you learn it and i will try to move and recreate all my workflows in forge, takes a little longer to learn but allows me more control
not sure that is enough and even if you manage to run like a sd1.5 model, probably will be too slow. I mean even on rtx2090 6gb of vram it struggle with sdxl sometimes, all this AI need good video card with more vram.
No, all the time i used things extra to install something at some point break when i update or do something, so i prefer to have them installed manually, but is good to use if you install a lot of interfaces
so i had comfy installed on pinokio or whatever its called, but then that program stopped working or some reason, anyways installed it following this tutorial and there is no option for installing models and ect like i did when i installed it the other way...? something isnt right.
Maybe Pinocchio had some options for that, that is used to install different user interfaces front what i know. But if you installed the comfyui manager you have a button for installing models also. But I usually include links from where i get the models to download directly because on the model page many times you can get more info on what setting you can use with that model.
@the_RCB_films is it in the models and then checkpoints folder? Some flux models and gguf go to unet instead, but all other should go to checkpoints folder
depends on what workflow are you using, what settings, if it appears all the time, or only sometimes, post some screenshots of workflow and the command window when that happen on discord to take a look
I want to make sure that you are aware that juggernaut X is a new base model and it needs time It's not a new base model, it's simply using SDXL base as jumping off point again. It doesn't continue on from Juggernaut 9's training, it went back to the start. with much less images in the training dataset (which results into it being extremely narrow).
@@makadi86 I mean I didnt include that information in the video, because it was too much info for an introduction video, you are free to give any advice :)
@@makadi86 I am sure are many models that can do different things, depends on what you need it for, I dont think there is the best model that can do anything, you choose model for what you need, there photorealistic models, cartoon models, all kind of models, we should use the models that works for what we need. I use juggernaut just because it can do a little bit from all, like a general model, so that works for me and what I want to explain in the video :)
for me was quite similar, I use comfyui because i can do a lot of things in it that forge doesnt let me, building my own workflow instead of working with a fix interface that maybe dont have the things i need.
A1111 is the automatic1111 name for one of the many interfaces for stable diffusion, if you search automatic 1111 web ui you can find more info, simila to forge ui, and comfy UI that is just a different interface that can run stable diffusion models
I used a1111 for a while if you look at older tutorials, then switched to forge that looked similar and was faster, now i am switch to comfy ui because forge is not updated anymore and a1111 is slower, so i am using comfy ui for a while until a better one appears
@@pixaroma Three questions, then: - How well does it fit with all the models and Lora? - how much faster is it than the 1111? - and how is this interface more convenient than promt? I understand that you can get used to it, but I don't understand what the Forge or 1111 interface is bad to this? We're still working with text - promt.
@@Fayrus_Fuma I didnt say the forge interface is bad, I love the forge and a1111 interface :) just a1111 speed is not as fast as forge yet so hope it get fixed, and the forge interface is not officially updated anymore. Comfy ui is more convenient because you can do what you need to do exactly how you want, you dont depend on a fix interface, you compare now just a simple text to image interface, but that is not what it can do. For example if i want i can do a workflow that generated generate a prompt using chatgpt or other free model, then take that prompt and make an image and then take that image is then upscaled, then it can use a mask to inpaint something, and then take that image and make a video, then take that video and add a style and everything you can thing of. All that with a single press of the button once the workflow is done. Can you doo that in forge or a1111 you can not. Is it a1111 easier to use than comfyui ? yes it is. Do i prefer the a1111 and forge interface over this? yes. Can comfyui do more then a1111? yes can do, much more. Supports stable audio, cand do chat agents, all kind, video, I am only at the beginning and saw so many things that can do. I didnt test the speed different but forge and comfyui is faster then a1111
It’s strange (perhaps I missed something), that you’re only able to create close 512x512 images; like why? I mean I get that it’s trained that way, but why? Who wants images limited to such a small size?
That is more advanced, i will see if i can do it on a later episode after i finish with basics episodes, then it probably needs control net, ip adapter so only then I can do that
I installed ComfyUI via Pinokio and downloaded Juggernaut model. I moved the GIANT safetensor file into the correct folder but now I see that it is duplicated in several places under Pinokio folder. Why is that? Is that a way to avoid this duplication?
![Shoot EPIC MOVIES with this FREE AI Tool! [ComfyUI Tutorial + Free Workflow]](/img/1.gif)
If you have any questions you can post them in Pixaroma Community Group facebook.com/groups/pixaromacrafts/
or Pixaroma Discord Server discord.gg/gggpkVgBf3
When you install manager if says git is not recognized you need to install git first git-scm.com/downloads
Get ComfyUI from here:
github.com/comfyanonymous/ComfyUI
Get the ComfyUI Manager from here:
github.com/ltdrdata/ComfyUI-Manager
Download SDXL model - example Juggernaut XL - Juggernaut_X_RunDiffusion from here:
civitai.com/models/133005?modelVersionId=456194
Download SD V1.5 model - example Juggernaut Reborn from here:
civitai.com/models/46422/juggernaut
Update: one of the members of our discord channel made an easy to install version with different nodes already installed github.com/Tavris1/ComfyUI-Easy-Install
do you have a civitai referral code?
@@TheSpendog I don't have
I take it your a Graphic designer yourself, I see it through how you dedicate your time explaining each and every step. I have watched plenty of tutorials for all these AI tools and most of them are made by programmers/coders who run through it, assuming the entire wide world have knowledge in coding. Thank you so much for proving my point again, when it comes to teaching and understanding the audience better, it can only done by designers. All of your content is absolute gold.
Thank you 😊
i do have some knowledge in coding & i still cant even begin to unravel this 🐂💩...
@@libbychang413 lol 😂
@@libbychang413 same ahah 😅😅
Thanks for your effort. This series is just so friendly for a beginner.
You explained it as better as it can be. As a first time user I'm telling you "Thank you".
Awesome tutorial man, Was super hesitant to make the move but i finally did, Can't wait for your new episodes
Not as scary as i thought it would be :D Thanks so much for this!
One of the best things I saw in my phone notifications. Excellent, this is very valuable and I hope artists understand the value of this resource and reference. Thank you for taking the time to record this tutorial.❤
Damn, this video makes Comfy UI way easier to get. No more freaking out!
Can't wait for Ep02 😊
I love how easy you make this look. Thank you for this tutorial.
Got 5 minutes in and already appreciate you setting up the landscape of comfyui. I already installed it now, trying to digest and learn as a beginner. Seems like I will enjoy your tutorial much more than the last!
THANK YOU! for this series of tutorials. You are a godsend.
thank you! you have no idea how useful this series has been
Finally, there's a tutorial solved my problem. I tried too many methods to install this.Like installing python ,viusal studio, Rust, git batch and changing function code in python file. its really useful and simplify my working. Much appreciated, it's very clear and concise,and say it again, Thank you!!!
Glad it helped ☺️
Loved all your FORGE videos and now you are doing tutorials on COMFY I may yet give it a go. Thanks for the clear and concise vids, much appreciated.
Thank you for this, your training style and flow is super helpful to keep me on track.
The best video on internet for learn and understand ComfyUI.
This is really the best guide to start with it.
Everything in this video ist just perfect!
I really recommend it!
Thanks for it!
Great video! Your series is gonna help bring me up to date with this program super fast and I heartily thank you for it
Thanks, this ComfyUI tutorial series is the best one I've found so far, Pixaroma simplified everything not like other "Tutorial Videos" tend to make things complex, his channel deserved more view and scribe, oh, and thumb up.
I just found out about this whole community and having a stable diffusion locally and could not been more excited :>
thanks for putting this together man, you won yourself a subscriber!
Thanks ☺️
Fantastic tutorial thank you. After a year with a1111 and have a good understanding with it and 80% good results. I've often thought about comfy but the UI has put me off a little, but with your excellent tutorial series your doing this gives me a stronger interest in it. I'm looking forward to seeing more and giving it a try. Once again thank you.
Thank you for the lessons, the functionality is explained in great detail, you are an excellent teacher.
Great tutorial! Looking forward for the next one!! I am starting this journey with you!
Awesome! Thank you! ☺️ Can't wait to learn also more about so I can share
Great video man, Excited to see the next episodes. Keep the great work
This is by far the best introduction to ComfyUI that I've come across. As usual, I tried to go a bit beyond the tutorial and drag and dropped images from Juggernaut XL on CivitAI. The workflow is a lot more complex than the one presented here, but if you do want to give it a go, just make sure the files referenced (Checkpoint + VAE) are both compatible. You will also need to track down the upscaler file referenced.
Thank you! Wonderful tutorial!!! I am still trying to get my head around what is happening here. I definitely will be sticking with you as my guide in this new adventure.
Amazing and straight to the point tutorial. Thanks man.
What a great course to start with.. very clear and inspiring for futur work to do
Thank you for making this. Very clear steps!
Thank you so much for this clear and cool presentation. I am very new to this and it helped a lot.
Glad it helped ☺️
Awesome video! Clear and easy to follow instructions as always. I am really hoping you are going to create a workflow for Sketch/Image > Design like you did before in Forge, that was awesome, really looking forward to see you replicate that in ComfyUI and show us how to do it or sharing the workflow! I am pretty sure you can build a pretty robust workflow for that with maybe control net/canny + "model training" or reference for the Sketch/Image to turn into something with similar style. Great work! Looking forward to the next episodes!
Yeah, after I am doing all the basics that are need to know, I will go step by step on more stuff and practical projects, just like I did on forge.
This was very helpful, thank you for the explaination.
really good video! clear instructions !
Thanks so much for this, VERY helpful!
I am very happy I found you. This first lesson was very helpful as I just goy ComfyUI installed and working. The only thing I am missing is the menu in the lower right corner. At the top of my screen is Manager, which is the same as yours.
Yes comfyui got an update but all the buttons should be on the top left corner, where it say workflow and other menu, i put on discord in the comfyui-faq channel comparison between old and new menuu
Lovely video! waiting for more
Good to see you are replying to the comments even 5 months after posting.
I just installed the ComfyUI for the first time, the UI is a bit different.
I cannot see the extra options in bottom right. it just shows queue
The interface changed now is on top left bar where is the workflow menu and others, i put a screenshot before and after on discord with all the menu, look for comfyui-faq channel on my discord
Your clear and precise. Thumbs up, subscribe. and informative. great.
Thank you ☺️
i was searching 2h how to add this manager omg. your video amazing
This is great! Thank you.
Hello, Thanks for these actualized series of Comfy tutorial, I plan to watch them all. I have about two weeks playing with my 4090 in Comfy, i had compared your its/s and you are ahead of me for 10it/s more, how is that possible?? do you mind making a video to go over the hardware and tools you need to run efficiently Comfy? Ty.
Well the versions change and updates but it might be other things to consider, like for example i have 128gb so that might help to load the models in memory, i have ssd so it loads faster, i9 , so not sure how all influence, definitely vram helps the most but the interface change more things are added so maybe some things are working faster or slower , I don't know all the technical stuff 😀
Nice man really love it 🤘
thanks mate, love the video please do more
Thanks, will do 😊
Thank you very useful tutorials
Finally a more recent take at the whole process! I am so grateful you were so organized throughout it. Side note, I might have this fixed by the time you see it, but I haven't been able to even install ComfyUI yet. The unzipping keeps taking forever. It seems to be doing 'something', but I'm 30 mins in. Is that normal?
I don't think it should take so much, the file is big but it should take half hour. Did you try winzip or WinRAR to extract it mayne is because of something else. Try restarting windows, make sure the the zip archive is complete also maybe it didn't download all the way
@@pixaroma This is so frustrating. I tried all of those steps, and tried unzipping with 7zip (it seemed to not do anything after 1.5 hours) and also WinZip. The latter did work, but those .bat files are nowhere to be found. It's as if there's something wrong with those files from their website. Is there an alternative link for download by any chance?
@@pixaroma Finally, WinRAR was the only one that was able to extract it after hours waiting. Thanks!
very nice 💚💚
Thank you 😊
Thank you, much appreciated!
Thank you. Very nice tutorial
Perfect You are the best
Perfect 🎉 thank you
Thanks for doing this. Do you know of a way to use the models folder from automatic 1111 in comfy? I have so many models and don't want to have to download them again just for comfy. Thank you!
Yeah there is a way to do that, try this mitsuhakitsune.com/en/2023/12/20/share-models-between-automatic1111-and-comfyui/ if that doesn't work i will look for exacts steps, im on mobile now, but i saw it being done by many people so there is a way to do it.
@@pixaroma Thank you. Will give it a try. Thanks again for your Comfy series. I enjoy your channel.
@@ralphmccloudvideo careful on the slash and back slash when you put the path, need to be like in example / is not the same with \
@@pixaroma Thanks for the "heads up" I'll look at it as soon as I finish up client work this week :) Thanks again!
starts well. there is a sub and a like. greetings from PL
Thanks ☺️
Thank you
Thank you so muchh !!
will you do a text2vid tutorial in the future?
yeah i plan to do that in the future, I want to cover all that i learn, but i am starting with simple stuff, with image first because video is in fact a series of images, so we need to master the image creation first :)
Simply I was lost until I found this
I'm having trouble installig VLM nodes. There are some libraries to install. You had a video with how to do the requirements txt or individual modules in the txt file. Which vid is this? Thank you.
those nodes that use llama-cpp-python can give some problems on certain nodes, like in episode 11 some had problem installing the searge node. someone from community made a installer for comfyui and all the nodes I used so far, including the one with searge llm, so maybe it will work if has the rigth dependencies, or you can post on discord maybe someone that has more code knowledge can help, here is the installer i talked about github.com/Tavris1/ComfyUI-Easy-Install
very useful!
thank you so much
when i try to update ComfyUI-Manager ComfyUI-Manager by cmd ; it message git' is not recognized as an internal or external command,
operable program or batch file.
Download and install git first git-scm.com/downloads
Would you make a tutorial about loras?
I will do later so is in order, probably after image to image, control net, upscaling and stuff
@@pixaroma thats great!
What happens to unused image creations in Comfy ui? Are they just deleted after so many, or are they accumulating somewhere in the pc?
If you have a save node then is saved in the output folder and remain there unless you delete manually, if you have a preview node then is saved in temp folder and is deleted when you restart comfyui
Is there a way to have the output folder be on a nas instead of the default comfyui\output folder?
I am not at computer to test it but check the comments on this reddit post like the last comment people replied that the method worked www.reddit.com/r/comfyui/s/Ll6CBHjZtk
Thankyou buddy.
Can you send the link of workflow to download.
you can get it from google drive , click download on top right corner and drag it in your comfyui interface drive.google.com/file/d/1gkcZ6En8bahsK1AmuGcpEGl4yDpTroBe/view?usp=sharing is the same workflow from ep03 that is also available on discord together with all other workflows
The word model now is used everywhere. What the model in this context is referring to?
A model is often referred to as a checkpoint. The term "checkpoint" refers to a saved state of a model during or after training, which can then be loaded and used for generating images. So when you're loading something like Juggernaut XL, you would typically load it as a checkpoint file (usually with a .ckpt or .safetensors extension), which contains the trained model's weights and data.
@@pixaroma thank you. So here the model is referring to an AI model.
@@johnovercash1798 yes
Thanks, Can i run this on CPU because i don't have GPU
Yes, but it might be very slow so not sure it is worth waiting so long for an image to generate instead of running the gpu bat you run the cpu bat when you start comfyui
i cant get the right bottom thing that box which has save load refresh thing how to get that
That is in the old interface in the new one you have menu on top left where you have load workflow and refresh
I don't have the save load and refresh how do I get that.
On the new interface there are on the bar, for example i have activated top bar and there i have load workflow, export, refresh and all those
Any clue on why I am getting: "'NoneType' object has no attribute 'tokenize'" error message after hitting queue prompt?
Is it updated to the latest version? I think someone mentioned a similar error on discord
@@pixaroma I just updated the comfyui, and still getting the same error message.
@@MohdMobasherulIslam can you post some screenshot on the discord server, on the comfyui channel, to see what node is turning red discord.com/invite/gggpkVgBf3 usually that happen when it need a clip text model or is incompatible or missing some dependencies
How do I generate multiple images at once? I could do batch count in A1111 but I can’t find anywhere on RUclips that says how to do it😢
Under the width and height on the empty latent node you have batch value that is 1, you just change that to 2 or what number of images you want and will generate that number. You can also use Auto Queue. I have talked about those options in some of the episodes at the beginning not sure if ep2 or 3
What will you do with your Forge? Will you keep it?
I still have it, until i remake all the workflows from forge in comfyui
bro, when I install comfy UI manger then my laptop says 'git' is not recognized as an internal or external command,
operable program or batch file. so what i can do
install git first git-scm.com/downloads I already had it installed and forgot to mention in video, just download for your operating system like windows version and install it and try again
Please can you help. I have been running into all sorts of problems with Comfyui manager. Every time I try use the manager to install or update I get this message in red : TensorRT Node for ComfyUI install failed: Bad Request and the command prompt says ...AppData\Roaming\Python\Python311\site-packages\torch\lib\cusolverMg64_11.dll" or one of its dependencies.
I have tried every possible sollution on reddit and installed Conda, Python and every other version I can find but nothing helps. I even downloaded every Nvidia update, driver and tool I can find and still nothing. Any idea what is going on?
So the manager works but when you try to install the tensor node it gives you problems? Did you tried the update bat from comfyui update folder? All this complex errors more than i can handle. Worse case just install im other folder comfyui again. Other things chatgpt suggested: Reinstall CUDA and cuDNN: Ensure that you have compatible versions of CUDA and cuDNN installed. Sometimes a mismatch can cause issues. - Update or Reinstall PyTorch: Compatibility issues with PyTorch can also lead to such errors. - Rebuild the Virtual Environment - Manual Installation of Missing DLLs . Sorry i can help you more, but are so many things in place that is hard to know what could cause.
Hi, I would like to install Comfyui on my local disk, but my computer doesn't meet the installation requirements. Can you recommend a laptop that can handle locally? Please suggest the most affordable options. From this description, it's hard for me to choose a laptop that would handle it. Maybe someone has a recommendation?
I cannot recommend a laptop because i don't use any, only desktop. My recommendation is to get one with nvidiat rtx card and a lot of VRAM, 16gb of vram so you can run flux models, 24gb are quite expensive. Maybe you can do some slower with 12gb of vram, but lower than that only if you use sd1.5 model and some sdxl. But for flux I will go with 16 gb of vram
@@pixaroma hello, thank you very much for your answer, it's fantastic that you found time to answer, in that case a better solution would be to build your own computer, hmm, could you give me the rest of the parameters of such a computer so I could calculate what will be cheaper for me, thank you very much in advance
@@pixaroma or maybe an option would be a macbook m2 and adding a hard drive to it
For Image generation what is important is to have a good Nvidia card with a lot of vram that will speedup the generation, 32 gb of system ram and SSD so it loads your model faster, and a processor that can handle the video card you choose, so as you can see the video card is most important for image generation, more vram and newer the video card is the faster you can generate
Some had problem running on mac some of the nodes or configuration, the best worked on nvidia with windows
im just getting started, what would you suggest me to begin with? ComfyUI or Forge UI?
Forge ui is easier but the official is not updated anymore so i will probably not do tutorials for it angmore, the comfy UI is most active and can do a lot of things once you learn it and i will try to move and recreate all my workflows in forge, takes a little longer to learn but allows me more control
@@pixaroma thank you brother! Starting to learn 🏆
can it run on nivdia 920M 2GB ?
not sure that is enough and even if you manage to run like a sd1.5 model, probably will be too slow. I mean even on rtx2090 6gb of vram it struggle with sdxl sometimes, all this AI need good video card with more vram.
The comfyui direct download link isn't working on my pc 😢
Try a different browser, i just tried on mobile and download started so the links works, try on different browser and see if download starts
@@pixaroma you're right 😊🙏 the link actually works fine..
It was my Internet connection that had a problem.
Thanks for the quick response 🙏✨
Im unable to find:: Sampler: DPM++ 2M SDE
it appears in the list in ksampler and the pp is for plus plus, look for: dpm_pp_2m_sde
thank you , can you show us how to animate images on ComfyUI
I will see what i can do in later episodes, videos are composed with images so we need to master the images first before making them move :)
@@pixaroma THANKS
You don't use Stability Matrix?
No, all the time i used things extra to install something at some point break when i update or do something, so i prefer to have them installed manually, but is good to use if you install a lot of interfaces
@pixaroma I agree, that's why I took a break from SD but will follow your tutorials to catch up. Thanks bro
I got it to do a 2k image with juggernaut reborn and it was a topdown view of a battle scene, kind of looks like a screen shot of an old RTS game lol
'jiggabyte' ;) Good video, thanks for creating this!
so i had comfy installed on pinokio or whatever its called, but then that program stopped working or some reason, anyways installed it following this tutorial and there is no option for installing models and ect like i did when i installed it the other way...? something isnt right.
nvm
Maybe Pinocchio had some options for that, that is used to install different user interfaces front what i know. But if you installed the comfyui manager you have a button for installing models also. But I usually include links from where i get the models to download directly because on the model page many times you can get more info on what setting you can use with that model.
@@pixaroma i need a checkpoint installed? it doesnt detect them when in the model folder?
@the_RCB_films is it in the models and then checkpoints folder? Some flux models and gguf go to unet instead, but all other should go to checkpoints folder
@@pixaroma oh, and also i got stupid for a sec, im tryna install flux.
what causes output image to be grey or black
depends on what workflow are you using, what settings, if it appears all the time, or only sometimes, post some screenshots of workflow and the command window when that happen on discord to take a look
I want to make sure that you are aware that juggernaut X is a new base model and it needs time It's not a new base model, it's simply using SDXL base as jumping off point again. It doesn't continue on from Juggernaut 9's training, it went back to the start. with much less images in the training dataset (which results into it being extremely narrow).
Yeah I saw in their description just is too much information for beginners that don't actually need :)
@@pixaroma I didn't understand you. Do you mean that you don't need such information from me! and I should avoid that?
@@pixaroma I just wanted to say that juggernaut X is not the best.
@@makadi86 I mean I didnt include that information in the video, because it was too much info for an introduction video, you are free to give any advice :)
@@makadi86 I am sure are many models that can do different things, depends on what you need it for, I dont think there is the best model that can do anything, you choose model for what you need, there photorealistic models, cartoon models, all kind of models, we should use the models that works for what we need. I use juggernaut just because it can do a little bit from all, like a general model, so that works for me and what I want to explain in the video :)
thanks for all your hard works, they are amazing, do you have a telegram channel for updates and news? i would love to join.
I don't, only the discord channel or Facebook group, see the pinned comment, thanks ☺️
thanks
Is comyfyui faster than forgeui?
for me was quite similar, I use comfyui because i can do a lot of things in it that forge doesnt let me, building my own workflow instead of working with a fix interface that maybe dont have the things i need.
Hello and thank you, I do Exactly step by step but my generated result is totally different and not interesting at all
Can you post on discord screenshot with workflow and the results? Look for comfyui channel
Its great that you can save setting but its still a weird ui
What's it for if there's an 1111? I don't get it. 🤨
A1111 is the automatic1111 name for one of the many interfaces for stable diffusion, if you search automatic 1111 web ui you can find more info, simila to forge ui, and comfy UI that is just a different interface that can run stable diffusion models
@@pixaroma I still don't get it, is the point of going from 1111 to this?
I used a1111 for a while if you look at older tutorials, then switched to forge that looked similar and was faster, now i am switch to comfy ui because forge is not updated anymore and a1111 is slower, so i am using comfy ui for a while until a better one appears
@@pixaroma Three questions, then:
- How well does it fit with all the models and Lora?
- how much faster is it than the 1111?
- and how is this interface more convenient than promt? I understand that you can get used to it, but I don't understand what the Forge or 1111 interface is bad to this? We're still working with text - promt.
@@Fayrus_Fuma I didnt say the forge interface is bad, I love the forge and a1111 interface :) just a1111 speed is not as fast as forge yet so hope it get fixed, and the forge interface is not officially updated anymore. Comfy ui is more convenient because you can do what you need to do exactly how you want, you dont depend on a fix interface, you compare now just a simple text to image interface, but that is not what it can do. For example if i want i can do a workflow that generated generate a prompt using chatgpt or other free model, then take that prompt and make an image and then take that image is then upscaled, then it can use a mask to inpaint something, and then take that image and make a video, then take that video and add a style and everything you can thing of. All that with a single press of the button once the workflow is done. Can you doo that in forge or a1111 you can not. Is it a1111 easier to use than comfyui ? yes it is. Do i prefer the a1111 and forge interface over this? yes. Can comfyui do more then a1111? yes can do, much more. Supports stable audio, cand do chat agents, all kind, video, I am only at the beginning and saw so many things that can do. I didnt test the speed different but forge and comfyui is faster then a1111
It’s strange (perhaps I missed something), that you’re only able to create close 512x512 images; like why? I mean I get that it’s trained that way, but why? Who wants images limited to such a small size?
With sdxl you can do bigger, and you also have upscalers, i do with flux 1024, and then upscale x2 and then x2 again so i get 4k images.
How to faceswap
That is more advanced that need control net and maybe custom nodes, I will see if i can make a video on later episodes after i finish with the basics
That is more advanced, i will see if i can do it on a later episode after i finish with basics episodes, then it probably needs control net, ip adapter so only then I can do that
installing the manager did not work
Did you get a git not recognized error? Try to install git first git-scm.com/downloads
I feel like when I import someone's workflow I'm needing to download hundreds to thousands of add nodes and models.
I tried.to keep my workflows to minimum
I installed ComfyUI via Pinokio and downloaded Juggernaut model. I moved the GIANT safetensor file into the correct folder but now I see that it is duplicated in several places under Pinokio folder. Why is that? Is that a way to avoid this duplication?
Sorry I don't know how pinokio handles the model, they have a discord channel maybe can ask there
2 months later, 2:19 was not necessary
😁
Nope. To much grind to get a descent image. I prefer Forge, Flux_webUI and Fooocus.
use the UI that what works for you :)
nnah
I said the same first time, I postponed learning with a few months, until forge stopped updating their UI and I had to switch, and never looked back