|
Who sang it better? https://m.youtube.com/watch?v=mTq8cZSfztE https://m.youtube.com/watch?v=W3x9d0EIkYU
|
#
¿
Jul 18, 2023 20:10
|
|
|
|
| # ¿ May 13, 2024 21:30 |
|
|
Just started messing with comfyui   Get in my swamp.
|
|
#
¿
Mar 4, 2024 23:02
|
|
|
      e:     e2: I was trying SDXL with a null prompt, and it tends mostly towards birds and flowers in an oil painty style but every once in a while I get an oddball like this fellow:  snorch fucked around with this message at 07:10 on Mar 19, 2024 |
|
#
¿
Mar 19, 2024 02:21
|
|
|
I'd watch any one of those. Also PaleFigure and swagman what are y'all using to get those images? They seem unusually crisp and coherent, or is it just the Art of the Prompt at work? I get decent results with SDXL but I'd say it only ever reaches about 80% of that fidelity. 
|
|
#
¿
Mar 19, 2024 18:21
|
|
|
I think I found a groove here with Juggernaut-XL. My favorite thing to do is feed it simple prompts and see where it goes.   e:  Way too many images to post here so I'll post this gallery where I've named them according to their prompt: https://postimg.cc/gallery/Nw1NkN2 snorch fucked around with this message at 08:09 on Mar 23, 2024 |
|
#
¿
Mar 23, 2024 07:26
|
|
|
It has been a bountiful day.                    
snorch fucked around with this message at 02:52 on Mar 29, 2024 |
|
#
¿
Mar 29, 2024 02:46
|
|
|
Oh neat, it's very high-res what model is that? I've been playing with OpenPose, everything here uses this same pose as input:            
|
|
#
¿
Apr 6, 2024 05:10
|
|
|
https://i.imgur.com/Cp0yM8T.mp4 Eat it Sora.
|
|
#
¿
Apr 6, 2024 07:16
|
|
|
XYZAB posted:stuff Yeah to me it's a wide mysterious world but all the youtube tutorials have been super helpful in wrapping my head around how people actually work with these things. Some more experiments today, this time it's mostly running JuggernautXL in two stages, through two KSamplers that have been fed separate prompts. It's good for doing variations on an overarching theme, you can freeze the seed on the first stage and just run 10 steps or so of the finisher stage. Simple prompts seem to still work best, I've gotten so much mileage out of the primer "party jesus" alone: weed party jesus/budgie smugglers  dog/cat  skyrim/photography  Bigger gallery here. This is spooky and wet and moody and I love it. snorch fucked around with this message at 02:38 on Apr 7, 2024 |
|
#
¿
Apr 7, 2024 02:22
|
|
|
As a musician, can I just say, LOL. Nashville's never hiring anybody real ever again. I've been messing with a new workflow, using IP-Adapter with a fixed image in the composition input, swapping in random images in the style input, and providing a simple prompt like "biker keg party". I'm really in love with the control it gives over the output, it's like a lora you don't have to train. For today's dump, some poster art.              
snorch fucked around with this message at 23:56 on Apr 10, 2024 |
|
#
¿
Apr 10, 2024 23:01
|
|
|
Funny, of all things I've never been able to get SD to paint a skateboard that isn't misshapen.
|
|
#
¿
Apr 11, 2024 03:15
|
|
|
Hadlock posted:Serious "creepy doctor who" vibes, here It wouldn't feel out of place in Vancouver's Stanley Park, reminds me of https://en.m.wikipedia.org/wiki/A-maze-ing_Laughter e: also of this fellow I found on a back trail at a park: 
snorch fucked around with this message at 02:05 on Apr 13, 2024 |
|
#
¿
Apr 13, 2024 01:25
|
|
|
If y’all want to DIY it there’s a lora on civit called "ps2 style", not sure if that's what the site is using but seems to give similar results. e: huggingface got hugged to death, might be why the web app is currently out of order. snorch fucked around with this message at 21:04 on Apr 23, 2024 |
|
#
¿
Apr 23, 2024 18:26
|
|
|
Longpig Bard posted:Finally got AnimatedDiff set up, this is pretty fun Very cool! I’ve been trying to set it up but can’t get the SDXL motion model to produce anything but gobbledygook. Might try it again today with 1.5 since I really want to mess with video.
|
|
#
¿
Apr 24, 2024 17:01
|
|
|
So I felt inspired and decided to teach myself to train LoRAs. It was a terrible pain to get going, involving some hacky things that felt very wrong, including trying to prevent the tools from running from two separate Python installs, and forcing specific versions of pytorch and transformers. In the end though I think I've got a pretty streamlined workflow within ComfyUI. I'd like to do a little more futzing and see if I can get it to work with SDXL, but for now SD 1.5 is fine for training my Screenshotinizer:
snorch fucked around with this message at 06:31 on Apr 25, 2024 |
|
#
¿
Apr 25, 2024 06:28
|
|
|
Today in HL2 mods that don't exist:        
|
|
#
¿
Apr 25, 2024 21:54
|
|
|
Today's topic was textual embeddings. I found this blingy one and combined with IP-adapter it's producing a range of fancy cocktails that I'd drink the gently caress out of:           Sir,
|
|
#
¿
Apr 26, 2024 17:02
|
|
|
|
| # ¿ May 13, 2024 21:30 |
|
|
Finally got some video working so we can have all the weirdness in motion. Not sure if these will embed but here's the gallery just in case.       
|
|
#
¿
May 10, 2024 02:34
|
|