Smooth animation with controlnet and regional prompting

[@AaronSherman]

Just want to say that this is amazing, and my new go-to example to show to people who think that AI art is just pumping a prompt into Midjourney. 😊

wow

[@xuantung191]

How to use the webui interface and extensions, the old conolnet of SD while still achieving vladmandic's optimizations?

[@gohan2091]

Unfortunately I am still at loss on how to use this. I would like you to have done a live example (seeing you type in the prompt, seeing you click the buttons, seeing you generate the image so I can see exactly what you're doing. We saw none of this. I'm lost where I put the base prompt and common prompt. I assume both within the main prompt box? How does it look? Maybe I am just stupid.

[@BagheeraTube]

Hi, I have been experimenting with this quite a bit, when I try to script it I get a lot of issues: --prompt "woman" --negative_prompt "deformed" --seed 342396165 BREAK --prompt "man" --negative_prompt "deformed" --seed 342396165 This is how I script it (my script is way bigger), had to get rid of the negative prompt and use common negative, I stacked 60-120 images in the script, and the more text the worse result. It's like the text bleeds over, and affects all the prompts, with too many negative prompts all the generated pictures will be prompted with the negative words. It's the same if I have many pictures, they all bleed together. Any ideas or thoughts?

[@BagheeraTube]

I ready someone saying that the scripting is broken so maybe that's why it doesn't work so well.

[@p_p]

thats clever. good job man

[@niallflavin4227]

That was a really interesting video, it would have been nice if you provide the commands and links in the description. Also it could be valuable to run through a whole example set by step to see it in action rather than in a narrative style! Otherwise excellent.

[@fofrai]

Great points, thank you, I've added links to the video description.

[@robertlinke2666]

so im not even able to place text beneath each other. either enter or shift enter just generates text, and if i but the breaks in on the same line it doesnt seem to work

[@jonmiles8091]

Love this. Thank you for posting and @GarryTan for RT.

[@MicahYaple]

Thank you for making this video and explaining the workflow so well, great break down! Would using a multi control net with canny as the initial + another control net using depth reduce the variation better I wonder?

[@brofessor584]

Nice video and animations! Do you plan to make a tutorial on scripting for Auto1111? Will be a nice addition. Or did you just use a prompt from text file option?

[@DavidJuk]

Great job, thx 4 sharing!

[@audiomurphs6306]

just use 19,1 18,2 17,3.... it will speed up the process as having 1,1

[@BagheeraTube]

Is there a way to script that?

[@alexgilseg]

Brilliant!

[@Repossessionn]

Very good video!

[@MOGmog]

superb！

[@d0null]

This means that that video was some hours of work right?

[@fofrai]

There was a fair bit of experimentation to find what worked. But now I have the process, other than the generations, it's pretty quick.