作者@YokoAI工作流，视频号、小红书搜索：YokoAI

1、提示词最关键，一定要体现是同一个场景下的不同画面。帧数可以是2帧、4帧或更多，可以自行尝试；

2、用这种提示词模板，大部分时候出的图一致性都挺不错，但是帧的逻辑不一定都是按时间顺序，也可能是倒序；

2、由于生成的帧之间会有一定像素的间隙，因此在做切割图片时，增加了一个偏移像素值，这个要根据实际出图的间隙调整，我这里是13像素；

整体上，不是还是要多次抽卡才能抽到不多效果；

更多工作流和AI工具玩法，可以关注公众号《YokoAI工具》。加ComfyUI交流群请加V ： FBI0414，这里有一帮很好很好的人。

——————————————————————————————————————————————————

1. The key to prompt words is to reflect different images in the same scene. The frame rate can be 2 frames, 4 frames, or more, and you can try it yourself;

2. Using this prompt word template, most of the time the consistency of the generated images is quite good, but the logic of the frames may not necessarily be in chronological order and may also be in reverse order;

2. Due to the pixel gap between the generated frames, an offset pixel value was added when cutting the image, which needs to be adjusted according to the actual gap in the output image. Here, it is 13 pixels;

Overall, it is not necessary to draw cards multiple times to achieve limited results;