Видео

I got one question and maybe anybody can help me: Im typing in my text prompt as I wish. To try different text prompts I have to ype in the text an test and the the next. Can I use different text prompts on one picture, press a button and then wait for an hour and see the results? At the moment I am typing in one prompt press refine, see the result. Then I type in the next prompt type refine and so on. I want to type in all prompts (always used on the same single picture), press one button and then see the results later. Can anybody help me?

Thanks a lot!

When I use hires fix it stops the refiner from working. Why might this be?

Hi Res steps at 0 will use the same number of steps as you've got in your regular steps, so 20 in your case. 1 or above will specify to run the Hi Res pass that many times, so experiment, but I tend to go with about half my generation steps or lower. The more you have the longer it will take, but you'll see diminishing returns eventually.

My "Hires.fix" stopped working :( It is showing same resolution, no upscaler, "from 512*512 to 512*512. Do anyone know how to fix it? Thank you.

I know this is crazy but I’m trying to understand why and the actual application

i was thinking about ai generated maps, so you could play a completely new map each game. but this is next level stuff right here

Thanks for the video....helped a lot

Mind blown 🤯

Please use WebUI in dark mode so settings are better visible in your video. Thanks. I did not knew that I needed Trigger words for some models.

stable-diffusion-webui\embeddings\ stable-diffusion-online\models\Lora\

Do you know if it's possible to train a lora with a 8gb vram card ? (i have a 2070)

Solid. Flux image to image possible too?

thank you! been holding off on Flux on account of low vram - looks like I've no excuse now! BTW should we have the latest comfyUI installed?

As a user of both, I'd say pick what you want, you still going to have to put your brain in gear when either of them break. That's when you really learn ! But auto then comfyui I'd say was a good way for me. At moment I'm going back to Auto to see what I missed while running comfyui for so long.

To me it seems like one of those things that if you need it you need it. Otherwise just stick to ComfyUI. Most people only scratch the surface of A11111 as it is

Amazing work !! did you continue your experiment ?

Brilliant series, I'm quite versed but even this helped me pick up little things. Much appreciated!

How did you get that generation history panel at the bottom? I've been trying to get it to show since I started using Comfy a few weeks ago, but haven't been able to. I've seen people saying there should be an x bottom right that toggles it on or off, but I don't see one, no matter the browser used, including Brave. Some also said set the page zoom to 80% as it's probably hidden, but still nothing. Is it a custom plugin?

Finally 🤤

I was wondering why. Thank you

Its nothing like generative fill

Shit I still don't get it...

Thanks for the Video. Im wondering is it possible to "Bulk" highres fix? Like If I have 10 pictures I wanna highres fix with the same stats? Right now im manually reloading them via png info -> texttoimage and highres fix

Is it better than Photoshop?

So Automatic1111. Got it.

what do I do if there are multiple different faces? Should I just send it to inpainting and do one face at a time?

Hm, this looks more like a simple color overlay, not a real change of lighting conditions 🤔

Fantastic video. Having a summary of pros and cons int the end for each model would be wonderful tho. Anyways outstanding work

Thank you

thank you for the tutorials 👍

Thank you sir ❤

Very good!

Very helpful!

then how do i download sd 1.5 ?

somehow my results are unsatisfying with stable diffusion automatic 1111 when trying to generate locally on my pc with my old gpu gtx 1080 ti 11gb, cant even get anything close to just even looking good or anything comparable, if i have to be honest automatic 1111 local generation cant even compare to the simple and basic generation of leonardo ai, so with that said am i missing anything? people claim they were able to generate stunning art with automatic 1111 but when i tried it, its soooo childishly unsatisfy like the results are worst than a 5 year old drawing, i've tried about 300x generative images now and i can just conclude that it wont take me anywhere without a better guide to a more pleasing result

Im trying sd1.5 cyberealisticv50 and pictures are always goofy needing editing at best. I think others dont have this problem SDXL seems so much better maybe its because i dont use embedding?

clear explanation, next textual inversion?,

You will be the VaatiVidya for SD....

thanks, clear explanation

Thank u so much

But what I dont understand most video tutorial said comfy UI is easier... what the hell

did not work

thank you very much for this video

omg the ctrl arrow keys. didn't know that. tyvm!

7:38 - not really. You could save and load different presets ;-)

Great tutorials man. Appreciate you 👏

I like the hi res steps because you can turn on an extra prompt box that will put in those prompts during those steps, it is an easy way to change faces. 👍

Question: With Hires enabled, 4 batch for generation selected, how will I build on an image I like and let Hi Res do the rest? If I apply highres, each time I will get different results and miss a previously better generated one.

I just wanted to say thank you for your simple explanation. This video is extremely helpful for someone who doesn't know anything about stable diffusion and it's just getting into it. Thanks for making this video!