4step>

CN:

1.使用SD/MJ生成一些具有连续性的静态帧/分镜

2.使用TOONCRAFTER去做图与图直接的动画插值然后合并

3.合并后的视频预处理得到不错的黑白线稿动画

4.使用IPADAPTER+SD15CN:SPARSECTRL scribbile+scribble+AD V3运动模型进行渲染

EN:

1.use MJ/SD IMAGE ON(Get some coherent image)

2.use dynamicrafter-tooncrafter > to get image+image=animation

3.ANIMATION PROCESS TO GET LINEART ANIMATION

4.USE IPADAPTER+SD15CN:sparsectrl scribble + scribble cn + AD V3 motion model render

更新记录update record:

当前版本V4-current wf version:V4

1.修正了image resize节点更新后的设置,且需要更新tooncrafter节点 -fix the node"image resize" setting after updated, then need to update node "dynamicrafter" by kijai

2.修正了英文版注释-fix the EN exegesis,  

3.增加了2.1clip对tooncrafter的支持和clipvison等加速2倍生成,且显存降低质量提升,10s内仅需10GB显存内就可以生成质量不错的16帧插值动画-ADD 2.1clip with tooncrafter and Accelerate the generation effect by two times, reduce the GPU VRAM memory and improve the quality.(now can generate 16frames in 10s with 10GB vram to me)

4.补充了部分区域的使用说明和注释及vram需求,修正ipadapter的权重并添加注释-Added usage instructions and comments for some areas, as well as VRAM requirements. Corrected the weights of the IP adapter and added comments.

5.增加了新的演示-ADD A NEW EXAMPLE SHOWCASE

6.WORKFLOW默认下载为中文版,英文注释请移步discord社区下载-The default download for WORKFLOW is in Chinese version. For English comments, please visit the Discord community to download.

贡献:DREX KIJAI

待开发:除LINE之外的OPEN POSE等其他类型插值AD渲染工作流