来自 X · Motion Transfer

We’re almost there I took a photo of myself, found a dance video, and de

提示词

We’re almost there

I took a photo of myself, found a dance video, and described the moves in a prompt. I attached my photo, and Seedance 2.5 generated this video of me dancing

I also tried my photo and the reference video with Kling 3.0 Motion Control. It lost facial details and added lots of artifacts. Seedance gave me a much better result

I regularly test the latest video models, and results like this make me think we’re almost there

提示词采集自公开 X 帖子,仅供灵感参考。版权归原作者所有。二次创作时请署名并附上原帖链接。