Workflow description :
The aim of this workflow is to generate video using the face from an existing photo in a simple window.
Resources you need:
đź“‚Files :
WAN2.1
Recommendation :
24 gb Vram: Q8_0
16 gb Vram: Q5_K_S
<12 gb Vram: Q4_K_S
I2V Quant Model: Wan2.1-I2V-14B-480P-gguf or Wan2.1-I2V-14B-720P-gguf
In models/diffusion_models
CLIP: umt5_xxl_fp8_e4m3fn_scaled.safetensors
in models/clip
CLIP-VISION: clip_vision_h.safetensors
in models/clip_vision
VAE: wan_2.1_vae.safetensors
in models/vae
FLUX
GGUF_Model: FLUX.1-dev-gguf
"flux1-dev-Q8_0.gguf" in ComfyUI\models\unet
GGUF_clip: t5-v1_1-xxl-encoder-gguf
"t5-v1_1-xxl-encoder-Q8_0.gguf" in \ComfyUI\models\clip
Text encoder: ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF.safetensors
"ViT-L-14-GmP-ft-TE-only-HF-format.safetensors" in \ComfyUI\models\clip
VAE: ae.safetensors
"ae" in \ComfyUI\models\vae
FLUX PuLID : pulid_flux_v0.9.0.safetensors
"pulid_flux_v0.9.0" in \ComfyUI\models\pulid
ANY upscale model (depreciated):
Realistic : RealESRGAN_x4plus.pth
Anime : RealESRGAN_x4plus_anime_6B.pth
in models/upscale_models
📦Custom Nodes :
PuLID need your python to have Insightface :
Check your python version :
for windows portable version : (the path depends on where you unzipped ComfyUI)
Download Insightface whl which corresponds here : Assets/Insightface
(Here my local python is on 310 and mobile version in 312)
Then install all prerequisites and insightface :
python.exe -m pip install --use-pep517 facexlibpython.exe -m pip install git+https://github.com/rodjjo/filterpy.git
python.exe -m pip install onnxruntime==1.19.2 onnxruntime-gpu==1.15.1 insightface-0.7.3-cp310-cp310-win_amd64.whl
Description
What's new? :
New interface,
New upscaler,
New model optimisation,
New LoRA loader.
FAQ
Comments (11)
There are so many that contribute you are definitly one of them, thank you! Is there a chanse you will create a RunPod template?
Hi there! First of all, thank you so much for your amazing contributions to the community — your workflows and resources have been incredibly helpful 🙏
I’ve recently been experimenting with Wan2.1-Fun-14B-InP and Wan2.1-Fun-14B-Control, and I noticed that there’s still very limited workflow support available for these two models, especially for the control model’s conditional inputs and trajectory-based controls.
If possible, would you consider updating or sharing a more detailed workflow for these two models? It would mean a lot to many users who are trying to get started.
Really appreciate all your efforts, and sorry to bother if this request is out of place!
I'm struggling with installing the requirements for this workflow. first i couldn't get tensor upscaling to install without errors manually or with comfyui manager but i found a reddit post from a day ago which had a solution which i followed. but now i can't get multi-gpu installed. i tried installing nightly, version before nightly, and git cloning from repo directly into custom models folder, but it all fails. i also don't have multi-gpu, i only have a single 4090, do i still need this? do i need to replace all the nodes with something else?
It says i am missing:
UnetLoaderGGUFDisTorchMultiGPU
DualCLIPLoaderGGUFMultiGPU
and going to manager it still shows this module missing i "try update" and i don't see any errors anywhere. in update or when i restart comfyui, but some reason workflow still shows it missing
Okay i am getting this error but not sure if it's related or not to multi-gpu. on startup:
C:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\timm\models\layers\__init__.py:48: FutureWarning: Importing from timm.models.layers is deprecated, please import via timm.layers
warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
I'm also getting this error and AI thinks it may be problem:
[WinError 5] Access is denied: 'E:\\ComfyUI_windows_portable\\ComfyUI\\.triton_cache\\cache\\CYECVAM3PQAFXLTXJ7SWTTMQNTO7PT5WTHGCJBFUR53G3VJB5MUQ\\cuda_utils.pyd'
i've tried turning off win defender realtime. running comfyuii in admin mode, tried deleting triton cache, i tried setting folder to full permissions, and so on. everytime i try updating multi-gpu module i get this error. I'm at a loss
you need to install also comfyui-gguf loader, it's a dependency for that node even if it's not clear at all..
@tapmoguls0o991Â Oh my goodness that worked! I spent hours on this looking up everything i could online. i used AI also and tried a million things. I scowled the multi-gpu github and issues there, reddit. Not a single source ever mentioned that comfyui-gguf is a dependency. Comfyui manager doesn't mention this or show it missing.. nothing. Thank you. I really wish comfyui was more user friendly in that it helped guide you more on what you need.
Unfortunately this workflow still isn't working. I don't have anything popping up saying i am missing modules anymore, but it generates just noise on the flux start frame. I have downloaded all the specific models from links here and versions putting them in correct folders, i reviewed the full guide. i don't know what i am missing again. What can cause the image to just be white noise? i can use flux and both wan on different workflows separately, so i am not sure what on this workflow is different to cause white noise so i can troubleshoot.
Okay flux model virtual ram from 4.0 and lowering it fixed noise problem. Now i don't know how to get wan part of workflow to work. it creates the flux start frame then workflow ends, no errors in console or anything. No clear buttons i can find for turning on the video portion of workflow or anything. so still not any success.
while troubleshooting i also found that all the brown windows for changing parameters on video like size, stepts, frames, tea cache, etc. cannot be modfied. clicking on them doesn't do anything. i tried restarting comfyui and all but they can't be modified at all. So something sadly must be broken for me on the wan/video nodes still.
ok deleting mxtoolkit module and git clone from repo fixed being unable to interact with the nodes. but still workflow stops after generating the flux first frame. This is all on a fresh install of comfyui btw, i have no idea why i am having so many issues with this workflow. i finally solve one problem then run into another problem.