>>53646 I did not make this bro, it was from deviantart https://www.deviantart.com/maniacalcharles/art/Secret-Agent-Baby-1-1218799931 >>53653 need to wait for a few days to be able to generate more videos in whisk for free

>>53663 ah had one more kek

>>53296 I have my own workflow but it is nothing fancy. I can recommend getting a prompt helper that just adds randomized keywords and that knows anime model prompting. Other than that, I just gen at normal resolutions and then upscale, being careful to not change the prompt text with the randomizer when upscaling. You also want to read the wiki entries for prompt keywords on danbooru, especially concerning style and special postures or circumstances. It can make the pictures more generic, but if you have a scene in mind, it helps getting some intuition about tags. Illustrious models work best for me at the moment, the diaper lora is my own but probably worse than what is out there. I don't share loras or workflows because they often contain too much machine specific info, and scrubbing it carefully is a PIA. In other news, adding fancy text to a picture takes a lot of time. First I wanted to imitate an anime advertisement by creating a fake title logo with fancy bubbly kanji, but I couldn't find an AI model that spit out the kanji I entered. Settled for this, could be an advertisement for a bed wetting school afternoon school club.

When testing age reversal drugs, you better come prepared.

mamabelle on tumblr has some pretty good ones

Some more experiments with WAN2.2. Generating a video takes 1-2 minutes, cutting it up properly so it loops is comparatively hard. I might just loop the video within the file so the frames smoothing algorithm can work its magic at the seams.

Now with sound. It isn't easy to sync the sound properly, I might have to find a lip syncing tool that just edits the mouth. Feedback appreciated.

>>53690 What do you use for audio? Is it an existing one just slapped onto the video? Tried vibevoice?

>>53693 I made the crinkling sound myself while watching the video, the girl voice is from a free audio clip (see filename). So far I haven't tried any voice syncing AI models.

Last one for today. I am normally not a content creator so this took forever. Please watch with sound.

>>53695 >I am normally not a content creator You are now

I don't like hyper but this one had such a funny face I had to keep it.

>>53643 These are really great! Interested in generating any more? What AI are you using and what images/prompts are you putting in?

>>53718 Glad you like them anon. Really not much "workflow" at all though, and no base images. It's all just basic prompting in Whisk that someone more skilled that me could probably improve on. And sure I'm always open to new ideas too if they happen to align with my schizo fetish autism (pics related). Example prompt below from the previous if you care: Facebook photo depicts an adult woman at an exotic alien desert resort being cradled in the massive arms of a towering giant green-skinned well-endowed maternal alien female figure with face off-screen at top. The woman has blonde hair in high ponytail with bangs and is nursing. She is wearing a juvenile colorful printed toddler-esque swimsuit with solid neon color swim diaper with button snaps on sides, and pool floaties on arms. The giant alien figure holding the woman is wearing a nursing sundress, and is nursing the woman. Exotic alien desert plants and structures are visible in the background, as well as a pool. A sign in the foreground reads "All humans must wear swim diapers, NO EXCEPTIONS". >>53741 So long gay bowser!

>>53741 This is really fucking funny to me. I'd save it and use it as a reaction video if it wouldn't expose me as a diaperfag.

>>53749 Glad you enjoyed it. You should see the one that dident work out and turned out... Wrong. You know what fuck. Here are some of the failures.>>53749

One turned out well. One did not xD

>>53749 You might be able to use this one. Its not diaper related

Have you ever tried MotionMuse to animate images? It seems some of the prompts are free for a while if you register. https://motionmuse.ai/r/pnpuiczi

Qwen megamax lora first pass went really well! Where the hell do we even upload loras anymore

>>53844 torrents and ddl sites

>>53844 SeaArt seems to allow it still. plenty of diaper and ABDL loras on there and searching still brings up results

Chroma lora test success.

https://gofile.io/d/9e7b4a25-1750-49e5-a732-b512757c0940 Here's the diaper package lora for Chroma. It should also recognize diaper wipe boxes and folded diapers. Trained on Chroma Base but should 99% work on the other sidegrades.

aldniki_qwen_edit_reality_transform_RH test turned out pretty good.

>>53882 Why tf has gofile files privated by default. Smh. Should work now. >>53884 Man I wish qwen was a porn model.

>>53882 Added a chroma lora for booties and mitens.

>>53852 >torrent people have said this before, can you give an example of one or two torrent sites where you have even seen something listed?

>>53968 Haven't looked yet, but it shouldn't be an issue since it's just a file like any other. Something like nyaa but for ai could also work and for ddl the shit game pirate sites do with the multiple link to various ddl sites.

>>53882 Retrained the booties lora on higher res. https://huggingface.co/clover-supply/Chroma-loras/tree/main Also started a test HF folder.

nano bananab9792b