https://civitai.com/models/1361986/illustrious-xl-akamatsu-ken-love-hina-negima-artist-style https://civitai.com/models/1115090/illustrious-realism-enhancer https://civitai.com/models/1300857/eye-detail-lora-illustriousflux

>>1083656 >>1083907 https://files.catbox.moe/lw4iai.jpeg https://files.catbox.moe/2o4vob.jpeg

Please stop posting unspoilered pornography, this is a sfw board and as such any picture posted to this board is shown on tbe front page, this makes this particular board incredibly unsuitable for a porn thread.

>>1084013 just inpaint the eyes and details instead. parororo and gogalking eyes look really good. ekrea jan too.

>>1084033 Seems like there's a peter under the fabric.

Attempted to make another Photorealistic Vivian. I had more interesting LoRAs for Pony Realism so I still prefer it, but I was able to make the stripes work better on an Illustrious Realism checkpoint. Couldn't get a decent ribbon though so gave up on that. https://civitai.com/models/212737/goddess-of-realism I got this one since it seemed like you could maybe do moody film styles with it. Needed a fairly high CFG to get things where I want them & controlnet helps as well.

>>1084080 By contrast, Pony Realism has LoRAs for movie stills. 80s Fantasy Movie: https://civitai.com/models/789313?modelVersionId=1286824 If I use the Vivian LoRA then stripes are easier to get right. Regardless, it seems like any inpainting just makes things smoother/flatter & less realistic so I can't do much there. Would like to try Flux since it's most suited to this, but better upgrade GPU beforehand, otherwise I'll get frustrated with how slow & limited it is on my current hardware.

>>1084040 peter is on vaction.

Don't mind me, just coming through

Is there a FOSS version of something like this? https://virbo.wondershare.com/app/talkingphoto/

>>1084213 It's getting better and better

>>1084245 Wrangling WAN is the new hotness. Just imagine, for $2000 you get to stich together 5 second clips.

>>1084298 https://github.com/Wan-Video/Wan2.1 >model requires only 8.19 GB VRAM >can generate a 5-second 480P video on an RTX 4090 in about 4 minutes Sounds like a long wait for me, but at least it's possible. There's a LoRA to make Ghibli animations. Studio Ghibli Style: https://civitai.com/models/1404755 A LoRA to make DOOM gameplay footage. Doom FPS: https://civitai.com/models/1376500 Particle Effects LoRA utilizing the IMG2VID functionality. Particalize I2V Lora: https://civitai.com/models/1360349

>>1084303 >8.19 GB VRAM WAN works for AMD?

Png is back. Also a few tweaks to the Ashley Graves pic.

>>1084445 cute sip >>1084307 what's going on in this image?

>>1084445 Blue Gun was better.

Is this Loss? https://civitai.com/models/1405786?modelVersionId=1589057 | Ɩı || |_

>>1084560 lol

>>1084560 LOL

>>1084036 from your post i gathered the want for a board in which sfw stuff must be spoilered but nsfw can be either way

>>1084717 Yeah the public education system is a disgrace, illiteracy is really on the rise. I think they offer GED courses though if you are willing to try.

>>1084176 I prefer where he's taller and with funnier eyebrows

>>1084758 Nice pits. For me, it's genderswap milf tiddies.

>>1083518 >most drawing programs won't open the .webp file type Name one

>>1084560 https://civitai.com/models/1405786/il-loss?modelVersionId=1602116

>>1084560 >>1084851 I figured you might do the other panels, but they'd pretty much require regional prompting to make anything. I also thought you'd put them all into the same LoRA with different trigger tags for each panel, but I guess it's simpler to do separate LoRAs. I've been browsing LoRAs on civit for the WAN model & things are slowly advancing there. I really wanna see more character LoRAs for WAN. Like if we had a Vivian LoRA for WAN model, could generate Vivian videos, but maybe IMG2VID could also do that with images of Vivian. I'm reluctant to mess with WAN until I get a 5090, but only reasonably available ones are in Newegg combos for far over MSRP. BestBuy repeatedly screwed me out of getting an FE. After months of trying it'd be nice to just have a damn 5090 already, but I can't justify blowing hundreds more extra parts that I'd lose money trying to resell. That's money I could use on other upgrades. Stock seems to be improving though since Newegg combos are available for hours now, unlike in the past when they'd be sold in minutes no matter what overpriced bundle of shit they were trying to push.

>>1084856 I don't expect the greatest results without advanced tools, but Illustrious knows enough characters it can get decent enough sample images without them. The real question I have though is who should Miku be talking to in the third panel? I don't know Vocaloids well enough to pick one and don't want to add a guy (even if it is accurate to the original) I have a good idea for the final panel's cover image though.

>>1084858 Miku just makes me think of the President Scott of Domino's Pizza anymore.

>>1084846 Clip Studio Paint & Paint Tool SAI.

>>1084778 it's not a man anymore it sucks

https://civitai.com/models/1416146/il-better-boyshort-panties?modelVersionId=1600641 "boyshort panties" is a tag, but it doesn't work very well so I made a LoRA that does it properly. Actually got enough material of other styles for a few LoRAs of other basic pantsu types while searching for this (and one is already trained+scheduled). Also realized how the third panel of Loss is the most abstracted of the four by far. Most of the parodies come down to "two people standing" with no effort to mimic the posing, accessories, or even BG of the original and of the ones that do many only use one or two of the elements. If anyone has more parodies not in pic 5, I'd like to have them. I will try without any more if need be, but I'm confident it won't be pretty (4 is much less abstracted though).

>>1085531 Do pic 2 and pic 3 with more obvious cameltoe. Cameltoe is just good culture.

>>1085596 All you need to do is add "cameltoe" with whatever weight you want.