Upload Make_em_frontrevcowgrl_high_noise_wan2.2_14b_i2v.safetensors

This LoRA working with WAN 2.2 14B I2V will first animate the input image, then fast cut to scene with the same woman squatting the same man in reverse cowgirl

## Usage

Trigger word: frontrevcowgrl

Recommended prompt:

frontrevcowgrl, the video start with a woman and a man, the screen goes dark and then lights up again to show the same woman squatting the same man in reverse cowgirl position. the man's erect penis is in the woman's vagina. she is leaning backwards as she bounces up and down aggressively. The man's hands grips the woman's hips. camera from front, their faces are visible. Face consistant. Man's face visible behind the woman's body.

Negative prompt: I personnaly use this one :

色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走, extra head, two heads, duplicate head, floating head, extra face, deformed body, mutated, extra limbs

## Training details

- Model: WAN 2.2 14B I2V

- Steps: 1500

- Resolution: 512x512

- FPS: 16

- Duration: 5s

- Videos: 20

## Recommended settings

- Sampler: Euler

- Scheduler: Simple

- Steps: 3 high, 5 low

## Notes

- Work best with 1/1 format

- Work fine in 16/9

- 9/16 seems to elongate their bodies

- If it produce extra limbs, try to add the k3nk nsfw helper Lora at @0.5 strength https://civarchive.com/models/2206794/k3nk-4llinone-nsfw-helper-wan-22-video-lora-k3nk

But you will lose some face consistency

- I used Wan2.2_Remix_NSFW_i2v_14b_high_lighting_v2.0.safetensors (And low) who is actually my favorite checkpoint to generate these videos

- Found the images used in the showcase on CivitAI and merged them to one image using Qwen

Make_em_frontrevcowgrl_high_noise_wan2.2_14b_i2v.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f096004d1987a11ba4dea2a8e2686ae62529fad3b58b7ccb6c3cf136232f6e9e
+size 306808440