Quick viewer retention tip: Show a sample of what the user experience and output is and *then* describe the installation and setup process. Someone coming across the video will not give a damn about the installation if they have no idea why they would be doing it to begin with.
Thanks for the feedback! I didn't expect any new viewers from TH-cam since I have so few videos online. It was all direct links to TH-cam from my Facebook posts. I'll have to step up my production game if these get any traction.
Fooocus also has controlnet... even better implementation. Fooocus -mrl is merged into Fooocus....already . Expansion prompt is ..gpt2 used for improvement prompts .
My memory must have failed me, when I first used Fooocus I didn't remember the controlnet feature. Good to know. Is the GPT2 prompt expansion baked into the app, or does it use an API and query an online server?
I have a silly question. I have unpacked the program. Where can I find the black background of your demonstration? I only have this white workspace. Thank you very much!
Its actually a browser setting, or is it a Windows setting? Either way look for Dark Mode. Start with your browser, go to settings and search "Dark" it also could be called Dark Reader.
The add to my own comment above. You can use Fooocus for the 1st frame. Then SVD and Comfy to create video.. about 25 frames. I made a tutorial for it on my website, but not youtube I'm afraid.
Thanks for the video, just got into the world of AI to try and improve my renderings but so far it just makes it worse! A large area of grass just turns into very CG-like mush, any reason this could be? IM using the same model as you etc.. thanks
My guess is that you're working outside the pixel limits of Stable Diffusion. SDXL has a native resolution of 1024x1024. SD1.5 is 512x512. So if you have a 5K rendering you have to either shrink it down to 1K or work in sections. Then its your denoising, Use something like .34 to .42 to simple alter your grass. You can check out my website, I have more free videos totally focused on AI for Architecture.
SURE! What are you trying to make? I'm no prompting expert, I usually start simple and add. Fooocus has lots of extra prompts it adds when you select a style.
I'm testing Fooocus, but it's melting my computer (32GB RAM, Ryzen 9 5900, Radeon 6800XT). Temperatures are off the charts, I can boil water on my GPU. There were a few times when the computer just shut down and I wondered if it would restart. Any ideas on how I can slow down Fooocus to render a little longer but not burn my computer to the ground? :)
so, i basically watched a useless video.... I obviously can't use it for architecture visualization as it just generate random stuf... thanks for the 10 minutes i will not be able to bring back.
I have over 5 hours of tutorials on AI for Architecture. Applying it to real world arch viz takes time and investment. Mostly I use Automatic1111 and a Photoshop plugin
Quick viewer retention tip: Show a sample of what the user experience and output is and *then* describe the installation and setup process. Someone coming across the video will not give a damn about the installation if they have no idea why they would be doing it to begin with.
Thanks for the feedback! I didn't expect any new viewers from TH-cam since I have so few videos online. It was all direct links to TH-cam from my Facebook posts. I'll have to step up my production game if these get any traction.
Yes
thanks for the content man, you don't want be an youtuber, but is definitely an good one, keep going :)
You made my day! Thank you for the kind words.
Fooocus also has controlnet... even better implementation.
Fooocus -mrl is merged into Fooocus....already .
Expansion prompt is ..gpt2 used for improvement prompts .
My memory must have failed me, when I first used Fooocus I didn't remember the controlnet feature. Good to know. Is the GPT2 prompt expansion baked into the app, or does it use an API and query an online server?
I have a silly question. I have unpacked the program. Where can I find the black background of your demonstration? I only have this white workspace. Thank you very much!
Its actually a browser setting, or is it a Windows setting? Either way look for Dark Mode. Start with your browser, go to settings and search "Dark" it also could be called Dark Reader.
I would love to see a showcase for creating a short video with Fooocus
You can't create videos with Fooocus. You need an extension called AnimateDiff which is available with Automatic1111 and ComfyUI
The add to my own comment above. You can use Fooocus for the 1st frame. Then SVD and Comfy to create video.. about 25 frames. I made a tutorial for it on my website, but not youtube I'm afraid.
Thanks for the video, just got into the world of AI to try and improve my renderings but so far it just makes it worse! A large area of grass just turns into very CG-like mush, any reason this could be? IM using the same model as you etc.. thanks
My guess is that you're working outside the pixel limits of Stable Diffusion. SDXL has a native resolution of 1024x1024. SD1.5 is 512x512. So if you have a 5K rendering you have to either shrink it down to 1K or work in sections. Then its your denoising, Use something like .34 to .42 to simple alter your grass. You can check out my website, I have more free videos totally focused on AI for Architecture.
@matthallettai Thanks Matt, I'll have a look. Yeah I was using a 2k downscaled from 5k. Cheers
@jonathanevans5453 if you find me on social media or my email send the image. I'm curious and like to help.
can you help me with the prompt commands ?
SURE! What are you trying to make? I'm no prompting expert, I usually start simple and add. Fooocus has lots of extra prompts it adds when you select a style.
What is your system configuration? Thank you
I have a AMD Ryzen 9 7950X with 128GB RAM and a RTX 4090 which has 24 GB VRAM.
Thank you very much!
You're welcome!
Що казати, Fooocus вже має навіть стиль 'Pa Art Ukrainian Folk Art'. І це круто!
Slava Ukraini!!
I'm testing Fooocus, but it's melting my computer (32GB RAM, Ryzen 9 5900, Radeon 6800XT). Temperatures are off the charts, I can boil water on my GPU. There were a few times when the computer just shut down and I wondered if it would restart.
Any ideas on how I can slow down Fooocus to render a little longer but not burn my computer to the ground? :)
You need a Nvidia GPU I'm afraid. You can run Stable Diffusion with an AMD GPU, but I don't know what to suggest.
@@matthallettai Thank you for response :)
@@matthallettaiThe CUDA and Tensor cores really make a difference, right?
does it work for non Nvidia card ?
Non mre variat does on its github page if u scroll down a tutorial is there for AMD cards I'm not quite sure for Intel GPU
I think it does now. Check out the github page for install instructions.
great tutors. Thanks
You are welcome!
so, i basically watched a useless video.... I obviously can't use it for architecture visualization as it just generate random stuf... thanks for the 10 minutes i will not be able to bring back.
I have over 5 hours of tutorials on AI for Architecture. Applying it to real world arch viz takes time and investment. Mostly I use Automatic1111 and a Photoshop plugin