An improved version based on ipiv's SDXL Lightning text2img2vid SD1.5 AnimateDiff LCM, adding localized image conversion prompt words, 2-frame Keyframe Interpolation, 4-frame/8-frame video prediction.
基于ipiv's SDXL Lightning text2img2vid SD1.5 AnimateDiff LCM工作流的改进版,增加了本地化LLM图片转Prompt,2帧关键帧插值,4/8帧关键帧视频预测,以及一个基本重绘。
Description
FAQ
Comments (7)
The ipadapter node has been recently updated. The previous node is invalid and the pull node needs to be updated again.
This workflow is so complicated!!! can you make one that uses sparsectrl section and i can just add it to my modified workflows, nm i see you grouped the sparse ctrl section thank you also i found out the scheduler you were using doesnt seem to be working very well? also leaning more into ip adapter is better as well is there a reason you have these settings. nm i think my image model had the lcm lora built in so it was over weighted
Is there any way you could record a video explaining how to complete this workflow? There are so many Load Image nodes and I'm really not sure what I should be putting in each of them. It's quite a large workflow! 😁
Look at thisAnimatediff Image2video 16/24/32Frames Attnmask Action Rgbsparsectrl SD15 - v1.0 | Stable Diffusion Workflows | Civitai, the same principle only needs load one picture