This is my simple animatediff workflow with some cnets and so on. Purpose is to generate a low res animation first as a preview and then a higher resolution version. Mainly to create short loops.
The only upscaling is from the highresfix script, you can obviously add a separate upscaling to the workflow yourself.
Follow pink notes in the workflow for some instructions.
THINGS YOU WANT TO GET:
v3_sd15_mm.ckpt - motion module
v3_sd15_adapter.ckpt - adapter to use as a lora
control_v11p_sd15_openpose - cnet
control_v11f1p_sd15_depth - cnet
control_v11p_sd15_softedge - cnet
control_v11p_sd15_canny - cnet
(if you have different cnet models you can just use those)
Use manager in comfyui to install missing nodes.
Description
FAQ
Comments (7)
I get blurry results. Any advice?
well i can't really tell just from that what's going on, but if you got all nodes and everything installed correctly it should work, also did you enable the detailer from the bypasser switch? maybe also reselect the model/lora/vae to make sure they are all correctly selected.
where is your futa model？？？it's lost
If you mean YeFaMIX it's still there with both versions, if you mean a lora of somekind i've never released a futa lora, except the bulge one.
best Comfyui workflow i ever stumbled upon, nice work Astraali !
Thank you!
v3_sd15_mm.ckpt & v3_sd15_adapter.ckpt, why are they pickle tensors and not safetensors. I know you didn't make them but I'm just asking.