Man, your tutorial is great, it has a good rhythm and is very explanatory. You just need to remove the music. These are complex concepts for us, beginners, and the music is too distracting. maybe something more abstract and certainly with a lot less volume works better imho. Thanks!
Awesome tutorial! Is there any way to generate cartoon images without specifying a prompt for each image? I mean, using a generic prompt, but still getting good results?
You can try using general words maybe, like instead of cartoon woman, to say a cartoon person that way can be both a woman and a man. Or if you have different combinations like people and animals maybe use cartoon character instead so it fit any character. For object just use cartoon object or cartoon item. And for other maybe cartoon scene, cartoon illustration
Sometimes I use it, you can check on other tutorials, mostly canny xl, but for this video wanted to do it simple without explaining again how to install extension and models ☺️
@@pixaroma So that's how simple it is to make a photo into a 3D cartoon, right? What version of SD are you using, and what are the minimum computer specifications to run it? is there a tutorial?
@@Vidanimatiopro in the video i used automatic1111 ui, Proto Vision XL. Not sure how much it need, but probably a video card rtx from nvidia and at least 8gb of Vram, more the better. I only have older tutorial on the channel, right now i switched to comfyui
I am facing difficulty while searching for models, even the filter, on civitai is confusing an complicated for me. Perhaps you could create a lesson for beginners to understand that. I remember I was looking for a good model to generate architectural interior and exterior images and I get lost in the internet and I did not know what to download.
You have to keep in mind that even the best models have week points, in the last month i only used juggernaut xl, right now i am at version 9. Interior many can do, but landscape in special vegetation and tree it gets more errors. Try juggernaut in general is good gor everything
hey bro good video, just wondering how you get such fast renderings, mines are taking up to 5 minutes, and I have a fairly good laptop, 32 ram and one of latest nvidia card, please help me :)
Is speed up a little there but depend on the image size on rtx4090 i generate a 1024*1024px image on 4-5 sec, but depends on the video card vram, probably your laptop video card doesn't have enough vram. To get better speed try installing forge UI instead of automatic 1111, it generates faster
@@pixaroma I have the rtx3050 and it takes like 5 minutes, dont know why, cause that one is good enough no? is there any tips to have fast renderings? thankyou bro
@@DiegoGallardo-sn9yh did you looked how much VRAM your video card has? I saw on laptop it has 4-6gb of vram, not 8gb like on desktop, so that might be a cause. But as I said try Forge UI, i have a video on how to install it on my channel, I used to have 3 minutes render on a computer with 6gb of vram on automatic, and switched to forge and now is only one minute, so I think you should get something similar with yours, 1-2 minutes instead of 5, and also the forge does the settings automatic so you dont need any extra arguments. You can stil try --xformers --medvram on your bat file arguments to see if helps before you try anything.
@@pixaroma my Vram is 4000 Mb yeah is not enough i guess. Wait but is Forge UI another rendering software or is it inside stable?? What does the --medvram do?
Stable diffusion is the model, and there are different UI that can run those stable diffusion models, like automatic 1111, forge ui, comfy ui and so on. When you start automatic 1111 you have a .bat file that it runs, there are some lines of text including some with ARGS with a equals sign, after that equal sign you put different arguments depending on the card you have, like ---xformers to speed up, and also how it handle memory if it crash you can ise --medvram or --lowvram , you can read more about here github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Optimizations but if you have forge it does that automatically
@@romangaming111 I dont really use 2d models I am only using mostly general models like juggernaut, but you can look at all models here and maybe is one in the style you like civitai.com/models
It should be possible, I am trying to recreate all the workflows i did with forge and a1111 for comfyui, so when I am able to do it right it will be a video tutorial about how to use it.
@@pixaroma Hello. I'm getting the error "AttributeError: 'NoneType' object has no attribute 'lowvram'". How can I fix this error in the simplest way? :(
@@valorantacemiyimben I used to got none type error on forge when i used width and height that is not divisible with 64 for images. But that none type was a more general error. Other problem might be because of lowvram, do you have enough vram to run the workflows you want? or did you set somewhere that option maybe. This happened to you in forge or comfyui? is hard to tell what cause it since are so many things involved
@@pixaroma I have a GeForce GTX 1060 6G graphics card. The vram value of this graphics card is 6gb. I don't get any errors when using ComfyUI, but I do when I do this:( I am using the dimensions and model you used in the video :(
@@valorantacemiyimben try to use a smaller size image to see if you still get error, like 768px or 512px to make sure is not about the vram, since 6gb on gtx is low, i have a rtx with 6gb and i can not do complex stuff like controlnet with it without crashing, on the video i used rtx4090 that why it works ok for me. So try to see if it work with a small image size, if get the same error then probably is not the vram and is something else
Man, your tutorial is great, it has a good rhythm and is very explanatory. You just need to remove the music. These are complex concepts for us, beginners, and the music is too distracting. maybe something more abstract and certainly with a lot less volume works better imho.
Thanks!
Thanks, will do!
Awesome tutorial! Is there any way to generate cartoon images without specifying a prompt for each image? I mean, using a generic prompt, but still getting good results?
You can try using general words maybe, like instead of cartoon woman, to say a cartoon person that way can be both a woman and a man. Or if you have different combinations like people and animals maybe use cartoon character instead so it fit any character. For object just use cartoon object or cartoon item. And for other maybe cartoon scene, cartoon illustration
@@pixaroma awesome, thank you very much
Very usefull. What could be easier ? To generate the image directly as a cartoon or to convert an existing photo ?
Both are easy, is harder when you are very specific like you really want to look like something, there takes more generation and experimentation
Why not use even controlnet in the process so that you maintain the resemblance?
Sometimes I use it, you can check on other tutorials, mostly canny xl, but for this video wanted to do it simple without explaining again how to install extension and models ☺️
@@pixaroma
So that's how simple it is to make a photo into a 3D cartoon, right? What version of SD are you using, and what are the minimum computer specifications to run it? is there a tutorial?
@@Vidanimatiopro in the video i used automatic1111 ui, Proto Vision XL. Not sure how much it need, but probably a video card rtx from nvidia and at least 8gb of Vram, more the better. I only have older tutorial on the channel, right now i switched to comfyui
I have tried with the models I have but the result was not that good, should I download this Photovision model to get the best results ?
Search for models that are trained to give cartoons, that way is easier to get carton style from a photo
I am facing difficulty while searching for models, even the filter, on civitai is confusing an complicated for me. Perhaps you could create a lesson for beginners to understand that. I remember I was looking for a good model to generate architectural interior and exterior images and I get lost in the internet and I did not know what to download.
You have to keep in mind that even the best models have week points, in the last month i only used juggernaut xl, right now i am at version 9. Interior many can do, but landscape in special vegetation and tree it gets more errors. Try juggernaut in general is good gor everything
hey bro good video, just wondering how you get such fast renderings, mines are taking up to 5 minutes, and I have a fairly good laptop, 32 ram and one of latest nvidia card, please help me :)
Is speed up a little there but depend on the image size on rtx4090 i generate a 1024*1024px image on 4-5 sec, but depends on the video card vram, probably your laptop video card doesn't have enough vram. To get better speed try installing forge UI instead of automatic 1111, it generates faster
@@pixaroma I have the rtx3050 and it takes like 5 minutes, dont know why, cause that one is good enough no? is there any tips to have fast renderings? thankyou bro
@@DiegoGallardo-sn9yh did you looked how much VRAM your video card has? I saw on laptop it has 4-6gb of vram, not 8gb like on desktop, so that might be a cause. But as I said try Forge UI, i have a video on how to install it on my channel, I used to have 3 minutes render on a computer with 6gb of vram on automatic, and switched to forge and now is only one minute, so I think you should get something similar with yours, 1-2 minutes instead of 5, and also the forge does the settings automatic so you dont need any extra arguments. You can stil try --xformers --medvram on your bat file arguments to see if helps before you try anything.
@@pixaroma my Vram is 4000 Mb yeah is not enough i guess. Wait but is Forge UI another rendering software or is it inside stable?? What does the --medvram do?
Stable diffusion is the model, and there are different UI that can run those stable diffusion models, like automatic 1111, forge ui, comfy ui and so on. When you start automatic 1111 you have a .bat file that it runs, there are some lines of text including some with ARGS with a equals sign, after that equal sign you put different arguments depending on the card you have, like ---xformers to speed up, and also how it handle memory if it crash you can ise --medvram or --lowvram , you can read more about here github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Optimizations but if you have forge it does that automatically
How to convert pure 2d cartoon
It might be a little harder, but you can try to find a model that can generate 2d, or is specialized for 2d cartoons
@@pixaroma if i provide you an android application that using Ai filters can you identify which model they are using from that ?
@@romangaming111 you can not find out a model by just looking at a photo unfortunately
@@pixaroma may be just you can use your experience and identify the thing behind that..
@@romangaming111 I dont really use 2d models I am only using mostly general models like juggernaut, but you can look at all models here and maybe is one in the style you like civitai.com/models
hello, can we do this with ComfyUI?
It should be possible, I am trying to recreate all the workflows i did with forge and a1111 for comfyui, so when I am able to do it right it will be a video tutorial about how to use it.
@@pixaroma Hello. I'm getting the error "AttributeError: 'NoneType' object has no attribute 'lowvram'". How can I fix this error in the simplest way? :(
@@valorantacemiyimben I used to got none type error on forge when i used width and height that is not divisible with 64 for images. But that none type was a more general error. Other problem might be because of lowvram, do you have enough vram to run the workflows you want? or did you set somewhere that option maybe. This happened to you in forge or comfyui? is hard to tell what cause it since are so many things involved
@@pixaroma I have a GeForce GTX 1060 6G graphics card. The vram value of this graphics card is 6gb. I don't get any errors when using ComfyUI, but I do when I do this:( I am using the dimensions and model you used in the video :(
@@valorantacemiyimben try to use a smaller size image to see if you still get error, like 768px or 512px to make sure is not about the vram, since 6gb on gtx is low, i have a rtx with 6gb and i can not do complex stuff like controlnet with it without crashing, on the video i used rtx4090 that why it works ok for me. So try to see if it work with a small image size, if get the same error then probably is not the vram and is something else