Workflow Name: Keyframe Interpolation AnimateDiff
【Workflow Introduction】
First, upload an image to the prompt reverse-engineering section. Then, use the Joy_caption_load node to reverse-engineer the prompt from the image. During this process, I uploaded seven pose images, and FILM VFI is used for in-between frame processing to achieve dynamic video guidance. Here, a 180° rotation effect is demonstrated. Next, ControlNet is used to control the image's structure, and ipadapter is employed for style transfer. Finally, the processed results are sent to animatediff and connected to the sampler to generate the video.

【Use Cases】
You can upload an image and use ipadapter for style transfer, making your image come to life. This process maintains character consistency well, ensuring that the image retains its original features while undergoing a style transformation. For professionals in the gaming and animation industries, this is undoubtedly an excellent workflow. It can provide new inspiration and methods for your creations, improve work efficiency, and make your works more stunning.

【Key Nodes】
AnimateDiff, FILM VFI

【Model Version】
SD1.5
Model Name: 3dAnimationDiffusion v10.safetensors

【LoRA Model】
None

【ControlNet Application】
HED Edge Blur Preprocessor
scribble fp16.safetensors Loader Model
Start Time: 0
End Time: 0.6
Strength: 0.5

【K Sampler】
CFG: 8
Sampling Method: dpmpp_2m
Scheduler: karras
Denoising: 1