WAN 2.2 I2V GGUF — My 8GB Daily Workflow | Upscale + RIFE - v2.0 Showcase

v2.0 of my daily WAN 2.2 I2V workflow. Same 2-pass pipeline (euler + heun) optimized for Q3_K_M on 8GB VRAM, now with LoRA support and quality improvements.

This workflow is also available as a 1-Click Pack — same pipeline but with ComfyUI, 7 custom nodes, and all models pre-installed. No manual downloads, no node hunting, no configuration. Just extract and start generating. Available in my Ko-fi shop → KoFi

What's new:

• 8 LoRA slots (4 HIGH + 4 LOW) using rgthree Lora Loader Stack — select from dropdown, set "None" to skip

• FastUnsharpSharpen added before upscale for sharper details

• Auto-adapts to any image orientation — no more forced cropping

• Complete negative prompt and NAG conditioning for better quality

• Resolution corrected to 480×832

LoRA strength guide:
1-3 LoRAs → strength 1.0
4-5 LoRAs → strength 0.7
6+ LoRAs → strength 0.6

New custom node required: rgthree-comfy (install via ComfyUI Manager → Install Missing Nodes)

Full technical details on why these settings work are in the v1.0 description.