Magic-Me AnimateDiff Video Workflow Character Customization, train your Character!
5.0
0 reviewsDescription
The best ComfyUI demo to animate anyone you want.
Feel free to train your character based on the image of the same person.
And you can even train the embedding for your favorite characters as long as you have the images and GPU.
The ComfyUI and embedding, and training codes are available in the following Github Repo:
https://github.com/Zhen-Dong/Magic-Me
The ComfyUI is fully runnable in the github Colab link.
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (1)
- latest (2 years ago)
Node Details
Primitive Nodes (15)
MagicAlbum3DGaussianNoise (1)
Note (4)
Reroute (10)
Custom Nodes (29)
- ADE_AnimateDiffUniformContextOptions (1)
- ADE_AnimateDiffLoaderWithContext (1)
- ADE_AnimateDiffLoRALoader (1)
ComfyUI
- FreeU_V2 (1)
- VAEDecode (1)
- KSamplerAdvanced (1)
- UpscaleModelLoader (1)
- VAELoader (1)
- CLIPTextEncode (1)
- EmptyLatentImage (1)
- CheckpointLoaderSimple (1)
- ImageCASharpening+ (1)
- SEGSPreview (1)
- ImpactSimpleDetectorSEGS_for_AD (1)
- SAMLoader (1)
- SEGSPaste (1)
- SEGSDetailerForAnimateDiff (1)
- FromBasicPipe (2)
- ImpactInt (1)
- ToBasicPipe (1)
- UltralyticsDetectorProvider (1)
- BNK_InjectNoise (1)
- BNK_GetSigma (1)
- VHS_VideoCombine (3)
- BatchPromptSchedule (1)
- UltimateSDUpscale (1)
Model Details
Checkpoints (1)
realisticVision_v51.safetensors
LoRAs (1)
v2_lora_ZoomIn.ckpt