Hi, this is my try to create an LTX2.3 Lora.

It’s Image To Video, I didn’t test it as Text To Video.

The initial image should be when a woman is lying on her back doing the side splits (look my Side splits start Lora).

Trigger word: swinglegs, but should work without it

Sample phrases:

swingupdownlegs, a nude woman is lying with her legs spread wide, she swings her legs up and down

swingupdownlegs, a nude woman is sitting with her legs spread wide, she swings her legs up and down with her arms

Check out attached videos for detailed prompts.

PS:

From what I see the Wan2.2 gives better results for Loras with complex motion. But the speech generation with LTX model gives some fun, also more long video duration is a plus.