When using it, it is generated in two steps. First, use the first stage to generate enough frames, and then switch to the second stage for video synthesis.

使用的时候分两个步骤生成，首先使用第一阶段生成足够的帧，然后切换到第二阶段进行视频合成

Put the starting picture in the folder in advance.Simple to keep only one picture.

预先将起始图片放到文件夹内，建议只保留一张图片

欢迎在此工作流的基础上进行改进，例如prompt travel，中途改变deforum运动方向等。玩的开心 ！

Feel free to playing and working on this workflow, e.g. prompt travel, change themotion direction of the deforum halfway through, etc. Have fun !

Important！

This version supports the generation of Kolors model. The operation logic is based on process control and file management. There is no special plug-in. You can use image cropping and expansion to control the direction of lens movement. You can use multiple sets of directions and hints to try to generate more shots. Always use the last frame of image to continue to generate new frames. When you think it is enough, open the composition module to get the video.

这个版本支持可图模型的生成，操作逻辑基于流程控制和文件管理，没有使用特殊插件，可以使用图像裁切和扩展的方式进行镜头运动方向的控制，你可以使用多组方向和提示词尝试生成更多的镜头，总会使用最后一帧图像继续生成新的帧，当你觉得足够的时候，打开合成模块就可以得到视频了

灵感来自于赛博长dick老师，感谢！

inspiration from @cyberdicklang, many thanks!