Animatediff-v3 workflow with IPadapter and controlnets
5.0
0 reviews
Description
This workflow is my latest in the series of animatediff experiments in pursuit of realism.
All you need to have is a video of a single subject with actions like walking or dancing.. Upload the video and let Animatediff do its thing.
For consistency, you may prepare an image with the subject in action and run it through IPadapter. I have tweaked the IPAdapter settings for vertical videos.
Model suggestion: Realistic_vision v5.1 or the latest EpicRealism model for best results.
Discussion
(No comments yet)
Loading...
Resources (1)
Reviews
No reviews yet
Versions (1)
- latest (2 years ago)
Node Details
Primitive Nodes (110)
IPAdapterApply (2)
Image scale to side (2)
Integer (2)
PrimitiveNode (1)
Reroute (103)
Custom Nodes (76)
- ADE_EmptyLatentImageLarge (1)
- ADE_AnimateDiffUniformContextOptions (1)
- ADE_AnimateDiffLoaderWithContext (1)
- CM_IntToBool (2)
- CM_BoolToInt (2)
- CM_IntBinaryOperation (3)
- CM_IntBinaryCondition (1)
ComfyUI
- VAEDecode (2)
- CLIPTextEncode (4)
- InvertMask (1)
- SolidMask (3)
- MaskComposite (1)
- MaskToImage (2)
- LoraLoaderModelOnly (1)
- SaveImage (3)
- CheckpointLoaderSimple (1)
- PreviewImage (6)
- ImageCompositeMasked (1)
- CLIPVisionLoader (1)
- ControlNetApplyAdvanced (4)
- FreeU_V2 (1)
- EmptyImage (1)
- ImageScaleBy (2)
- CLIPSetLastLayer (1)
- KSampler (2)
- LatentUpscaleBy (1)
- LoadImage (1)
- ImageCrop+ (3)
- MaskFromColor+ (1)
- MaskBlur+ (1)
- FaceDetailer (1)
- SAMLoader (1)
- UltralyticsDetectorProvider (1)
- ImageListToImageBatch (1)
- ImpactImageBatchToImageList (1)
- LineArtPreprocessor (1)
- DWPreprocessor (1)
- OneFormer-COCO-SemSegPreprocessor (1)
- IPAdapterModelLoader (1)
- PrepImageForClipVision (2)
- ControlNetLoaderAdvanced (4)
- If ANY execute A else B (3)
- DebugPrint (1)
- VHS_VideoCombine (1)
- VHS_LoadVideo (1)
Model Details
Checkpoints (1)
realistic_vision_v5.1.safetensors
LoRAs (1)
v3_sd15_adapter.ckpt