Part of my Wan 2.2 experiments, similar the previous 'Tonuge Job' but has many more options leveraging the wan new motion capabilities of the model, still not perfect but getting better...

I2V works best with already positioned initial frames and combined with "Self Forcing" (Wan21_I2V_14B_lightx2v_cfg_step_distill_lora_rank64_fixed) lora at 1.5 str on both low and high noise; 3-5 steps usually enough when using with it on low noise version. I tried new 2.2 version of that lora but it seem to kill motion in I2V...

As usual I just picked some generated images randomly to animate from related models, kudos to creators.

No trigger word used directly but 'penis worshipping' is common in all. I recommend focusing on the action describing the subject with the concept related words. Some nsfw words included in the train set such as: cum, blowjob etc.

Here’s an example prompt I used in some of these generations—feel free to alter it heavily:

A woman is engaged in enthusiastic penis worshipping, holding the erect penis and moving it close to her face. She licks the shaft, with close-up camera angles focusing on her face and the act.

A woman performs penis worshipping, holding a large, erect penis with both hands and bringing it towards her face. She is licking the shaft, smiling. She continues to kiss and suck the penis, then rubs it on her face. Camera is focused from a low, upward angle.

The low noise version can work standalone like 2.1 but for best results I recommend using it combined with high noise version. The high noise version is lower rank, and usually one step before low noise pass is enough.

The low noise version should work on Wan 2.1 workflows too but didn't test it...

I use total steps of 4, 1 for high noise 3 for low noise part multistep res beta samplers.

This is just early access and experimental version, use it at your own risk xD

Note: If you want to support my experiments, feel free to message me. Testing different settings and learning new training recipes for wan 2.2 seems to be expensive for my taste, I am open for runpod gift vouchers for that purpose... (Begging part)

As usual:

Please use the model responsibly, what you create is totally your responsibility. No real person faces were used in the training.