🚀 Excited to share my latest Wan Animate workflow based on Kijai + Enhance!
This setup allows you to:
Animate any character from a reference video or image
Replace characters seamlessly while preserving expressions, lighting, and color tones
Generate high-quality, cinematic videos right in ComfyUI
📌 What's included in this guide [YouTube Link]:
Full step-by-step instructions to set up the workflow
Tips for model selection, text encoders, and reference images/videos
How to enhance output videos, improve FPS, and work on lower-spec machines
Practical examples showing real-time animation and character replacement
Whether you’re a content creator, animator, or AI enthusiast, this workflow revolutionizes character animation, making it faster, cleaner, and more versatile than ever.
🔥 Pro Tip: Combining Kijai with Enhance produces smoother, more realistic motion and lighting adjustments automatically.
Check it out, try it yourself, and let me know what amazing creations you come up with!
Description
workflows added
FAQ
Comments (8)
all the animate workflows give me this "Given groups=1, weight of size [5120, 36, 1, 2, 2], expected input[1, 16, 54, 104, 60] to have 36 channels, but got 16 channels instead" i know its usually a mismatch but i cant figure it out.
try forcing the video and image dimensions (Width + Height) to multiples of 16. For example instead of 648 you need to make it 640 (16*40). Let me know if it works!
@Obzy98Â negative. ive tried who knows how many different dimensions. still nothing
make sure your VAE, unet and loras are actually matched. also verify if yo uare using GGUF or safetensor
@BloodlustDream yeah ive downloaded all requirements multiple times. completely stumped. i basically can drop workflow and it runs instantly but last step on sampler just errors with channel error. just finished finished redownloading clip and vae and still the same error. continuing down the list of models but still no luck. im using safetensor
Having exactly the same issue here as Ghostguru....
@Ghostguru Maybe try updating ComfyUI and the custom nodes?
any idea how to get character loras to work w this? it seems to just ignore them in favor of the reference image, instead of using both