Minor changes to the workflow
Laufzeitprofil
Quellenbeschreibung
Minor changes to the workflow
Updated the calculation of frames to use when less than 72 frames to frame_count + 1, this fixes a bug that uses double the sampling steps because there is 1 additional frame for the reference image.
Updated the Mimic Motion Scheduler to use Align Your Steps as per this issue: https://github.com/kijai/ComfyUI-MimicMotionWrapper/issues/8 and reduced the sampling steps to 10.
Introduced Condition Zero Out for the AnimateLCM conditioning as we are using CFG 1.0 which ignores the negative prompt.
KI-generierter Kommentar
KI-generierte Erklärung auf Grundlage von Quellen- und Konfigurationsdetails. Vorschläge sind klar gekennzeichnet.
Use Tik Tok Dance Workflow (Mimic Motion + Animate Diff) to animate a character from a driving dance video and composite it over a 16-frame looping background; the listed inputs also include a prompt or setting.
Mimic Motion generates the character’s movement, AnimateDiff creates the looping background, and Grounding Dino composites the two.
For a driving video, the description recommends a static camera with the dancer doing the moving.
Vorschlag · nicht geprüft
The author reports testing the workflow on ComfyUI v0.21 and directs users to install missing custom nodes with ComfyUI Manager.
Place AnimateLCM_sd15_t2v.ckpt in models/animatediff_models and AnimateLCM_sd15_t2v_lora.safetensors in models/loras; download control_v11p_sd15_openpose.safetensors through ComfyUI Manager.
Named model files include AnimateLCM_sd15_t2v.ckpt, AnimateLCM_sd15_t2v_lora.safetensors, control_v11p_sd15_openpose_fp16.safetensors, MimicMotionMergedUnet_1-1-fp16.safetensors, and realisticVisionV60B1_v51VAE.safetensors.
Custom nodes listed: ComfyUI Impact Pack; ComfyUI's ControlNet Auxiliary Preprocessors; ComfyUI Frame Interpolation; AnimateDiff Evolved; ComfyUI-VideoHelperSuite.
Also listed: ReActor Node for ComfyUI; rgthree's ComfyUI Nodes, ComfyUI Essentials; ComfyUI-MimicMotionWrapper; segment anything; Comfyui lama remover.
Choose an SD 1.5 checkpoint to suit the video style; the author gives Realistic Vision as an example for realistic video.
Vorschlag · nicht geprüft
The description presents v3.1 as a single workflow for systems with 16GB of VRAM; v4.0 splits the workflow into five parts to limit VRAM use to 12GB.
The description warns that hair interaction may not animate properly with Open Pose input and says depth information would address this.
If insightface installation causes trouble, the description points to the ReActor node’s troubleshooting instructions.
Vorschlag · nicht geprüft
Muss das bearbeitet werden?
Melden Sie sich an, um eine Änderungsanfrage zu senden.
Quellen
1 QuelleQuellenauszüge
2 AuszügeSource context: 5070 downloads · Type Workflows · Base model SD 1.5
Remaining hardware or download-size estimates are not available yet.
Anforderungen
AnimateLCM_sd15_t2v_lora.safetensors
Nicht aufgelöstLoRA · Unknown
AnimateLCM_sd15_t2v.ckpt
Nicht aufgelöstCheckpoint · Unknown
Source context: 4803 downloads · Type Workflows · Base model SD 1.5
ControlNet · 689 MB · Hugging Face · comfyanonymous/ControlNet-v1-1_fp16_safetensors
dw-ll_ucoco_384_bs5.torchscript.pt
Nicht aufgelöstCheckpoint · Unknown
GFPGANv1.4.pth
Nicht aufgelöstCheckpoint · Unknown
inswapper_128.onnx
Nicht aufgelöstCheckpoint · Unknown
MimicMotionMergedUnet_1-1-fp16.safetensors
Nicht aufgelöstCheckpoint · Unknown
realisticVisionV60B1_v51VAE.safetensors
Nicht aufgelöstVAE · Unknown
Checkpoint · Hugging Face
Checkpoint · 208 MB · Hugging Face · hr16/yolox-onnx