Home
Creators
Search
Recent
Missing
Random
Posts
DMs
Tags
Random
Importer
Import
FAQ
Account
Register
Login
Favorites
Keys
Logout
Telegram
(For SOTAI Squad) Wan 2.2 ComfyUI Tutorial: 5x Faster Rendering on Low VRAM (Patreon)
Downloads
Missing 1 full-res photo, 1 file, as this post has not yet been imported.
Content
Video: https://youtu.be/WU9rr04_D4Y
Learn how to use WAN 2.2 in ComfyUI with low VRAM and faster rendering methods - Optimize workflow
Step 1: Update ComfyUI and ComfyUI front-end package
pip install comfyui_frontend_package >=1.23.4
Step 2: Prepare Wan2.2 models
Wan2.2 model: https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/tree/main/split_files
(Original model, just for experiments)
GGUF wan2.2 model: https://huggingface.co/QuantStack/Wan2.2-T2V-A14B-GGUF
(Download Wan2.2-T2V-A14B-HighNoise-Q5_0.gguf model and put it in "ComfyUI/models/unet")
Note: If you run the attached workflow and hit an out-of-memory error, try experimenting with smaller-sized gguf versions.
Step 3: Prepare LoRA
Download and put in "ComfyUI/models/loras"
Step4: Install GGUF custom node
Install "ComfyUI-GGUF" custom node, detail in https://youtu.be/sVzQosJZWk8?t=623
Please check the workflow in the attachments.
Files
Previews only
