Thank you for a great video. Just to save you all some time in the future, you don't actually have to restart the application to load new models. Just click the "Refresh all files" button on the Model tab.
very best of the tutorial i could find fr a newcomer, job well done. What i am missing, from 3:45 to 4:45, is simpl explanantion, why are you doing what you are doing. For better understanding, me as a viwer i just blindly follow your lead because you got me hooked, but i have no idea what am i doing. Add a simple sentence this setting will do x if you change y and let us play with it.
How fast is the Turbo compared to the default? For me the default generates an image in 5-10 minutes, and do the inpainting and upscaling in about an hour.
What I don't understand is why the results in A1111, Forge-UI and Fooocus are so different. Firstly, Forge-UI is very fast, secondly there is A1111. On my 3050Ti/4GB Forge-UI generates images in 4-6 it/s, in Fooocus in 3-6 s/it! Especially in inpainting, changing the background, Fooocus does an excellent job, while the other two don't, it seems due to the way of interpreting the mask. Fooocus interprets the mask well and handles the edges well while the other two end up with a blurry edge.
Thank you for the video. I'd love you to start by saying, why you recommend this Plugin. I was watching the whole video and kept wondering, what does it do?
I've got a question. If I download a pony model should I pick that as my diffusion method and use it in the screen to select a checkpoint? Or is default fine
Great video. I use comfy ui and is there a way to use my ComfyUI checkpoint folder instead of the Fooocus one. I don't want to do a copy of the checkpoints. It's big files. Thanks for your help
Thank you. I asked this Question on Github I have got the best answer. Great Video. I have one more question how do you usually upscale your final image. Are you changing the model to non turbo SDXL and choose generation to Quality. What do you recommend.
One thing I would like at some point is being able to use any as base model being forced to use some sort of SDXL is a bit of a downside as like many other models and what not they have pro's and cons
@@MindRenders You can sort of "force" it via setting it to swap at 0.1 though it for sure doesn't like doing that in most cases, now this is something I've not figured out in my last hour of looking :D is there some "update" script for Fooocus?
So what causes the images to have that burnt look or like someone added a layer a put it on overlay with some noise to it? I have the setting you used and they are not coming out quite as clear as yours and I cant figure out what is doing it. I do have a fairly good computer with an older rtx2070 super and some i9 CPU.
GOODJOB 2 questions please: 1. Dreamshape xl turbo works with LORA and if it does which one yoi recommend. 2. If i copy paste your promote and lets say ur useing LORA and i dont have it what will it do? Thnx 3. Which top 3 turbo u recommend for high quality 1024x1024 that can perfectley work with Lora if possible
Good questions and thank you! 1. I only use the locon_perfecteyes_v1_from_v1_64_32 (1).safetensors so far 2. I only do testing without loras. So I only share my raw parameters unless the settings are baked into Fooocus.
Any attempt to change debug mode list are met with a white "forbidden" icon. This applies to all of the list such as guidance, sampler, scheduler etc. The debug mode is checked but I still am unable to change anything. Any ideas how to disable the disabler????
I haven't come across this before. Have you asked over on the issue tracker here? github.com/lllyasviel/Fooocus/issues You can also search and see if the issue is already resolved there.
As a new AI user I'm still trying to understand why my images are blurry. I've been creating simple forest scenes (leaves, trees, plants etc) and much of them are... blurry. Nothing I do seems to sharpen them. Will "DreamshaperXL" help solve my problem, or, is this intended just for creating faces? Thanks.
What's your favorite turbo model and why?
Do you prefer now TurbovisionSuperFastXL? And of so, why. Your following video swapped to this one.
They are both great. I like TurboVision for vibrant art. DreamShaper for realism.@@karlpj1
ultraspiceXLTURBO... it's spicy!
Thank you for a great video. Just to save you all some time in the future, you don't actually have to restart the application to load new models. Just click the "Refresh all files" button on the Model tab.
very best of the tutorial i could find fr a newcomer, job well done.
What i am missing, from 3:45 to 4:45, is simpl explanantion, why are you doing what you are doing. For better understanding, me as a viwer i just blindly follow your lead because you got me hooked, but i have no idea what am i doing. Add a simple sentence this setting will do x if you change y and let us play with it.
thanks for the settings to use the turbo model in fooocus. I had not managed to get any good results using the "extreme speed" radio button!!
So happy to help! Thanks for watching :)
3:43 Bro that Random is not for Negative prompt, it is for Seed 😂
Thank you. It's a little different question, but how can I use the eular a sampler in fooocus?
How fast is the Turbo compared to the default? For me the default generates an image in 5-10 minutes, and do the inpainting and upscaling in about an hour.
What I don't understand is why the results in A1111, Forge-UI and Fooocus are so different. Firstly, Forge-UI is very fast, secondly there is A1111. On my 3050Ti/4GB Forge-UI generates images in 4-6 it/s, in Fooocus in 3-6 s/it! Especially in inpainting, changing the background, Fooocus does an excellent job, while the other two don't, it seems due to the way of interpreting the mask. Fooocus interprets the mask well and handles the edges well while the other two end up with a blurry edge.
Thank you for the video. I'd love you to start by saying, why you recommend this Plugin. I was watching the whole video and kept wondering, what does it do?
Just asking isn't it the same as using the "Extreme Speed" setting in Fooocus?
I've got a question. If I download a pony model should I pick that as my diffusion method and use it in the screen to select a checkpoint? Or is default fine
Great video. I use comfy ui and is there a way to use my ComfyUI checkpoint folder instead of the Fooocus one. I don't want to do a copy of the checkpoints. It's big files. Thanks for your help
Thank you. I asked this Question on Github I have got the best answer. Great Video. I have one more question how do you usually upscale your final image. Are you changing the model to non turbo SDXL and choose generation to Quality. What do you recommend.
I've been using Topaz Gigapixel AI to upscale my images. It's not free but I feel it's worth the cost because I use it so often.
@@MindRenders I use Gigapixel and Nik Collection *all the time* within Photoshop. Great plugins ^_^
Thanks for the good video
One thing I would like at some point is being able to use any as base model being forced to use some sort of SDXL is a bit of a downside as like many other models and what not they have pro's and cons
Good point. I was playing around today using Lucidia Playground V2 as the base and OpenDallev1.1 as the refiner and had some great results.
@@MindRenders You can sort of "force" it via setting it to swap at 0.1 though it for sure doesn't like doing that in most cases, now this is something I've not figured out in my last hour of looking :D is there some "update" script for Fooocus?
So what causes the images to have that burnt look or like someone added a layer a put it on overlay with some noise to it? I have the setting you used and they are not coming out quite as clear as yours and I cant figure out what is doing it. I do have a fairly good computer with an older rtx2070 super and some i9 CPU.
GOODJOB
2 questions please:
1. Dreamshape xl turbo works with LORA and if it does which one yoi recommend.
2. If i copy paste your promote and lets say ur useing LORA and i dont have it what will it do?
Thnx
3. Which top 3 turbo u recommend for high quality 1024x1024 that can perfectley work with Lora if possible
Good questions and thank you!
1. I only use the locon_perfecteyes_v1_from_v1_64_32 (1).safetensors so far
2. I only do testing without loras. So I only share my raw parameters unless the settings are baked into Fooocus.
Any attempt to change debug mode list are met with a white "forbidden" icon. This applies to all of the list such as guidance, sampler, scheduler etc. The debug mode is checked but I still am unable to change anything. Any ideas how to disable the disabler????
I haven't come across this before. Have you asked over on the issue tracker here? github.com/lllyasviel/Fooocus/issues You can also search and see if the issue is already resolved there.
As a new AI user I'm still trying to understand why my images are blurry. I've been creating simple forest scenes (leaves, trees, plants etc) and much of them are... blurry. Nothing I do seems to sharpen them. Will "DreamshaperXL" help solve my problem, or, is this intended just for creating faces? Thanks.
HOW DO YOU DO IT THROUGH GOOGLE COLLAB , IF FOOCUS ISNT ON YOUR LOCAL PC
Having connection error for fooocus, is there a way to fix that ?
I have a 1660ti
Maybe there is a command for the fix ?
Share the error on the fooocus GitHub issue tracker. 👍
@@MindRenders thanks
Also wanted to ask the directory for the *IPAdapter model* comfy ui
Thanks bro!
Awesome work mate! Sub!
I want yse turbo on google colab plz make quick video