<h1 id="zero-two-dance"><span style="color:rgb(230, 73, 128)">Zero</span><span style="color:rgb(255, 0, 174)">-</span><span style="color:rgb(230, 73, 128)">Two </span><span style="color:rgb(255, 10, 206)">Dance</span></h1><p>A request from one of the best Civitai members, <a target="_blank" rel="ugc" href="https://civarchive.com/user/aiartX_work">aiartx_work</a></p><p>They provided me with nice clips of the action, I processed them to my rigorous standard, captioned them and trained them. The result is another very stable and flexible LORA.</p><p>I must say this one was fun because it's not inherently sexual, but is a great consistent concept to work with. I'll definitely consider doing more like this moving forward!</p><p></p><h3 id="one.wan"><span style="color:rgb(64, 192, 87)">One.WAN</span></h3><p>Continuing to train already collected/captioned sets for WAN, which I'm starting to be won over by, which doesn't spell the best news for my HunyuanVideo work...</p><hr /><p>***November 16th, 2025: I am going to be slowing down with my AI contributions, as fun as it is and as awesome as the community is, I'm not one to charge for things (i do the buzz stuff mainly for kicks...) and so without this benefiting me in real life beyond the 'fun', I have to reassess the priority and time put into these things. I hope I can get around to fulfilling the requests a select few members have made but I'll ask for moderately extreme patience as I need to refocus my energies. Thanks for all the fish! ;)</p><hr /><p></p><p></p><h2 id="one"><span style="color:rgb(34, 139, 230)">One</span></h2><p><span style="color:rgb(134, 142, 150)">Perhaps the only version needed... I selected 10 clips, 5 nude and 5 clothed, trained for 60 frames at 24fps, blurred faces, all 9/16 with the same pose and motion. I hit all the high notes in captions and this concept didn't require teaching textures or fine details, it's more of a ControlNet style LORA in the sense that it captured the general movement to apply to any subject.</span></p><p><span style="color:rgb(134, 142, 150)">Let me know if you notice any weaknesses!</span></p><p></p><p><span style="color:rgb(64, 192, 87)">Wildcard Prompt </span>(draft)</p><pre><code>Front view of a beautiful {Russian|French|Swedish|Swiss|Latina|Austrian|Dutch|English|America|California} woman doing the zero-two dance. Her arms are up with her hands behind her head as she sways her hips from side the side.

She has {very long|long|medium length} {bright blonde|blonde|black|brown|dirty blond|bleach blonde} {straight|ponytailed|pigtailed|wavy|curly|braided|styled|tied} hair and a {slim|fit|perfect|curvy} body.

She is {naked|wearing a dress|wearing a bikini|wearing pyjamas|wearing jeans and a leather shirt}.

She is in a {night club|modern bedroom|kitchen|crowded subway train|classroom}.

Her oiled body glistens in the {bright|soft|dim} light.
realistic</code></pre><h2></h2><h2 id="training-notes"><span style="color:rgb(250, 176, 5)">Training Notes</span><br /></h2><p>I'll be away for a few weeks so this is probably the last release until the second half of August. I hope I've given you enough LORAs to explore and combine (!!) for a couple weeks.... Someone asked for a revised <a target="_blank" rel="ugc" href="https://civarchive.com/articles/13187/hunyuan-video-lora-training">training guide</a>, we'll see if I find some time to write that up while I'm away from my hardware. I did notice the that WSL/Diffusion-pipe tutorial I was linking to went down in protest of Civitai's prudelords/admoneylords came online. I didn't expect that to be there forever though, fact is <a target="_blank" rel="ugc" href="https://github.com/tdrussell/diffusion-pipe">https://github.com/tdrussell/diffusion-pipe</a> has all the info you need to get up and running; google any pain points...</p><p></p><h2 id="disclaimer"><span style="color:rgb(253, 126, 20)">Disclaimer</span></h2><p>Don't not drink responsibly.</p><p></p>

Files with SHA256: 037bb0d2e8de8af47213e345938ee44359d6aa8789be38eb23e19104ba9654c9