Text slider to travel in time.What you must know : I didn't use WAN for video generation, but only for image generation, by the way, the High-noise lora was not tested as I only use the low-noise checkpoint with some "turbo" loras(*2) (let me know in the comments if it works well). Made with <a target="_blank" rel="ugc" href="https://github.com/ostris/ai-toolkit">Ostris AI-Toolkit</a> on a "low-end" hardware(*1) just because everybody seems to say it's not possible :).I kept 75 steps as it appears to me the best results (I've tested every 25 steps from 25 steps to 300 steps). Recommended maximum strength is -7.5 / +7.5, but you can test further, it will depend on other loras and/or prompt.At high positive strength, the persons could become like puppets, and at high negative strength, they could become dwarfs.Edit : I've added some images testing some concepts, and I'm not totally satisfied by some of them (vehicles in particular, but landscape too, and others...). I'm working on a v2. Comments appreciated.__________________________________________________________________________________________________________________________(*1) : Trained on a laptop with 32Gb of RAM and a mobile RTX4060 with 8Gb of VRAM.(*2) FYI : To make T2I I use this specific combination of loras (and all example images were made using this) : <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1872e83a-2f36-4475-9c8d-4855672ee3d5/width=525/1872e83a-2f36-4475-9c8d-4855672ee3d5.jpeg" />These are loras for WAN2.1, but I found them working better (at least some time ago when I made my tests).

Files with SHA256: cb44bffc19c067a70f0130152522babfad246f68a4b0ed524b0ca0ecfa6ede84