Great job on the video! Super helpful information! I hate to see Fooocus just sit when it's such a great UI, but I understand where mashb1t is coming from and needs to move on. Hopefully someone will pick it back up and continue to update it. That being said I am also looking forward to new videos featuring Flux and SwarmUI. Your teaching style is what the SwarmUI community is in desperate need of! I'm sure the mcmonkey (the SwarmUI dev) will great appreciate good quality tutorial videos such as yours! Keep up the good work buddy!
As for the base model you can use whatever one you prefer. Of course certain ones will give you different results depending on what they are trained for. For example the Cheyenne model is good for comics and other artistic styles but does not really do photographic type images. Chinook is trained to do cinematic type content. So really it is a matter of the results you are looking for.
You can use inpainting to do that. I have a few other text videos that can help with that depending on what you are attempting to do. This was one of my previous ones that might help. th-cam.com/video/BbKeDEQ7uik/w-d-xo.html
Will you also try the SimpleSDXL2 from metercai which supports flux? That would be awesome! Support Flux, SD3m, Kolors, and HyDiT models on the familiar Fooocus UI
I was hoping the Fooocus guy would spruce up the inpainting experience. It especially needs a hotkey to hide those pesky icons that always get in the way. And an 'Unpaint' brush, so you don't have to repeatedly CTRL Z through your brush strokes to correct an error you made early on in your painting. I actually put a Github ticket in for it, and then you just said that Fooocus has no developer. Which let's face it, means it's dead. Although I mostly learned how to use Fooocus by myself in 2 days by trial and error (after not using AI for ages because of the horrendous Easy Diffusion), your channel is great for learning all the stuff I wouldn't even think of thinking is possible - like this text stuff. The only thing is, you so much as sneeze and Fooocus is downloading 6 gigs here, and 6 gigs there, and 300meg here, and 600 meg there - for all these models and Loras that take over my SSD ! I wish you could make a video showing what the intended use, or best use of/for each model that is in Fooocus' drop down menu upon install, is for. This is best used for Anime, this is best used for 3D art etc etc. Then we would be able to decide which 6 gig model is worth downloading for our needs. Apologies for the essay 🙂
I do cover some of the options in other videos but would suggest if you want to know about any models I suggest looking them up on CivitAI.com for more info about them. Obviously the anime preset is best for anime and the realistic is meant more for photo type images although I normally use the default which uses JuggernautXL which is more of a general use model which works very well. I do have videos on Pony and Playground that goes into more details on those. SAI just uses the original SDXL which is usually not that great. If you want more detail on each preset there is a preset folder in the Fooocus folder and in there files for each preset and if you look in those json files you should see the name of the model it uses and you can look those up for more info. I actually use other models not included that I get on CivitAI that way I can choose exactly what I want.
I have not jumped back to Forge yet since I had already started using Swarm. But I do plan on checking it out soon. But will be doing stuff on Flux soon. But it changes so rapidly right now and I can't get a video out quick enough right now without it being outdated.
Welcome back! Very happy to see your developments again!
Good to know you are back! Great video, thanks!
Your videos are so damn interesting and detailed, educational.
Thank you Kleebz!
Great job on the video! Super helpful information! I hate to see Fooocus just sit when it's such a great UI, but I understand where mashb1t is coming from and needs to move on. Hopefully someone will pick it back up and continue to update it. That being said I am also looking forward to new videos featuring Flux and SwarmUI. Your teaching style is what the SwarmUI community is in desperate need of! I'm sure the mcmonkey (the SwarmUI dev) will great appreciate good quality tutorial videos such as yours! Keep up the good work buddy!
Thank you! And thanks for the help in getting myself going with Swarm and Flux. Can't wait until I have more free time to really get into it.
Good to see you back.
Thanks! Good to be back.
LoRAs used in video are linked in the description.
it is clear when you change the Lora's model, but can you explain please , when you change the Base model SDXl ? and Why you changed ?
As for the base model you can use whatever one you prefer. Of course certain ones will give you different results depending on what they are trained for. For example the Cheyenne model is good for comics and other artistic styles but does not really do photographic type images. Chinook is trained to do cinematic type content. So really it is a matter of the results you are looking for.
@@KLEEBZTECH thank you for the detailed reply, I appreciate 🙏
Great info as always! Is there a way or method to add text on to an existing image? Maybe creating the text without a background?
You can use inpainting to do that. I have a few other text videos that can help with that depending on what you are attempting to do. This was one of my previous ones that might help. th-cam.com/video/BbKeDEQ7uik/w-d-xo.html
Will you also try the SimpleSDXL2 from metercai which supports flux? That would be awesome!
Support Flux, SD3m, Kolors, and HyDiT models on the familiar Fooocus UI
Nice.I try it today. Does it work with "flux schnell" also?
@@metanulski Yes, SimpleSDXL does work with flux1-schnell and flux1-dev.
I was hoping the Fooocus guy would spruce up the inpainting experience. It especially needs a hotkey to hide those pesky icons that always get in the way. And an 'Unpaint' brush, so you don't have to repeatedly CTRL Z through your brush strokes to correct an error you made early on in your painting.
I actually put a Github ticket in for it, and then you just said that Fooocus has no developer. Which let's face it, means it's dead.
Although I mostly learned how to use Fooocus by myself in 2 days by trial and error (after not using AI for ages because of the horrendous Easy Diffusion), your channel is great for learning all the stuff I wouldn't even think of thinking is possible - like this text stuff.
The only thing is, you so much as sneeze and Fooocus is downloading 6 gigs here, and 6 gigs there, and 300meg here, and 600 meg there - for all these models and Loras that take over my SSD ! I wish you could make a video showing what the intended use, or best use of/for each model that is in Fooocus' drop down menu upon install, is for. This is best used for Anime, this is best used for 3D art etc etc. Then we would be able to decide which 6 gig model is worth downloading for our needs.
Apologies for the essay 🙂
I do cover some of the options in other videos but would suggest if you want to know about any models I suggest looking them up on CivitAI.com for more info about them. Obviously the anime preset is best for anime and the realistic is meant more for photo type images although I normally use the default which uses JuggernautXL which is more of a general use model which works very well. I do have videos on Pony and Playground that goes into more details on those. SAI just uses the original SDXL which is usually not that great. If you want more detail on each preset there is a preset folder in the Fooocus folder and in there files for each preset and if you look in those json files you should see the name of the model it uses and you can look those up for more info. I actually use other models not included that I get on CivitAI that way I can choose exactly what I want.
Does this work on unsupported text like mandarin or Arabic etc...
Not something I have tried but without having the text in the prompt it is much more difficult to generate with controlnet.
Hi could you some tutorials about Flux on Forge webui. Thanks
I have not jumped back to Forge yet since I had already started using Swarm. But I do plan on checking it out soon. But will be doing stuff on Flux soon. But it changes so rapidly right now and I can't get a video out quick enough right now without it being outdated.
SwaamUI is amazing, i import all my Fooocus styles it works fine.@@KLEEBZTECH
@@KLEEBZTECH Great Choice I hope to see your SwarmUI tutorials. I am using SwarmUI with Fooocus styles. works fine
A video that isn't about flux very bold also great lettering have got a chance of using the flux fork RuinedFooocus yet. Thanks
lol. Thanks. I am not a fan of RuinedFooocus. I have tried using it a few times and just does not work for me.